Why do people think stud WRs are more valuable... (1 Viewer)

[heavy-set1]

then why do you keep taking the bait?

I also don't find your argument to be all that clear or convincing. It appears you are basically arguing that the VBD way is the only possible way to determine a player's value. Yet, you also state that "mock drafts are meaningless for the simple fact that we dont know how the players will perform". If you don't know how players will perform, then what is the point of making projections, which are later used to determine the player values that you keep referring to? :wacko:

I'm sure there are many factors (both subjective and objective) that could influence player 'value'. Performance relative to others at their position in terms of year-end statistics are simply one individual factor among the many. Mock drafts, when done well, are clearly a useful way of understanding the 'value' of players to many other fantasy owners. I believe you lose some credibility by calling them "meaningless".

i dont do projections. i use several years of past results to get an average of how each position will score. avt tells you historcally how positions perform for a given scoring system. this combined with your starting roster requirements will produce a baseline. compare the scoring differences of the different positions to determine what the best value is at any point in a draft. mocks mean nothing to me casue they change all the time. but, if i have a player ranked wr9 and i move him up to wr6 due to subjective and objective factors, i still know , historically, how wr 6 will do, and can assign him a value.

thats what value is. i guess i am saying that avt/vbd is the only way to define value. you can alter your baselines, but ultimately value is determined by the scoring differences between different positions. maybe it was clear

value isnt, in a 2wrl, wr1 scores a greater %age of a teams points than a 3wrl, so stud wrs are more valuable in a 2wrl.

value isnt solely looking at the wr scoring and not comparing it to other positions

you are right, i shouldnt take the bait, but i like helping people.

hope it helps.

 

[aaronr28]

i dont do projections. i use several years of past results to get an average of how each position will score. avt tells you historcally how positions perform for a given scoring system. this combined with your starting roster requirements will produce a baseline. compare the scoring differences of the different positions to determine what the best value is at any point in a draft. thats what value is. i guess i am saying that avt/vbd is the only way to define value. you can alter your baselines, but ultimately value is determined by the scoring differences between different positions.value isnt, in a 2wrl, wr1 scores a greater %age of a teams points than a 3wrl, so stud wrs are more valuable in a 2wrl. value isnt solely looking at the wr scoring and not comparing it to other positionsyou are right, i shouldnt take the bait, but i like helping people.hope it helps.

While AVT and VBD are definitely useful models, I'm not certain that they are the be-all and end-all of determining player value. IMHO, they are simply one method of determining how valuable a player is. Even if they are the best and most objective method of those currently available, that does not mean that other methods should not be considered as potentially useful and informative.I also don't think the % of team total is a convincing argument for a player's value. Any player will represent a larger % of team total points when the starting requirements shrink, and vice versa.Perhaps BnB hasn't been all that persuasive in proving his case to many of you...but, I certainly don't think he has any kind of hidden agenda (as you are implying), and I'm not yet convinced that AVT/VBD strategies are as flawless as some of you seem to believe.

 

[heavy-set1]

While AVT and VBD are definitely useful models, I'm not certain that they are the be-all and end-all of determining player value.

exactly.its simply a useful guide. i just happen to find it the most useful.

 

[aaronr28]

exactly.its simply a useful guide. i just happen to find it the most useful.

OK, I would probably agree.But your (and others) arguments against BnB so far strike me as similar to saying "I'm right because I use AVT/VBD, and you're wrong because you're not using them" often without really explaining WHY one is better than the other.I just think it's interesting to look deeper sometimes. Blindly relying on the numbers spit out by AVT/VBD models does not seem an ideal way to determine player value. I see nothing wrong with attempting to evaluate player value in a unique way.

 

[heavy-set1]

exactly.its simply a useful guide. i just happen to find it the most useful.

OK, I would probably agree.But your (and others) arguments against BnB so far strike me as similar to saying "I'm right because I use AVT/VBD, and you're wrong because you're not using them" often without really explaining WHY one is better than the other.I just think it's interesting to look deeper sometimes. Blindly relying on the numbers spit out by AVT/VBD models does not seem an ideal way to determine player value. I see nothing wrong with attempting to evaluate player value in a unique way.

the info iam using has already happened, so its not wrong. i think i have clearly documented its usefullness, is it perfect? of course not, but the #s are the #s, and unbiased conclusions are drawn from them.

 

[rabidfireweasel]

Let me ask you one question. Based upon the mocks above (and they are very similiar to the mocks and expert rankings I've seen around here) and the avg. points tied to each spot, where would you draft WR1 in a S2L and a S3L (standard scoring)? Also, based upon your scoring system, where does WR1 go in both leagues?

Bass-I have been in a car for 9 hours today and my brain is fuzzy- so I may make a comment or two that I completely recant on later. However, it is IMO, the smartest people that can changes their minds and admit later when they are wrong, so I will say what I think now.First, I think you have done a tremendous amount of work in an attempt to investigate a situation where you are going to get verbally hammered. Then you put it out into the world. Whether I agree with your results or not, I want to thank you for your efforts.I have some issues with your assumptions and some issues with your conclusions. Here are some problems I have (and unfortunately for some points I do not have a suggestion for a better solution)Problems with process:1) First you are synthesisizing two very different things : AVT (which is not VBD) and a mock draft and asserting that your results prove something in the world.I think this is a stretch. I think this can lead people to try to think about things differently, but ceertainly this offers no proof. I believe that the mock drafter who does not follow AVT, drafts a player (sometimes) on what he projects him to do the next year, not what the 5th running back taken has historically done in that situation. Furthermore, mock drafts are not a good indicator of how people will actually draft. Many people check out, allowing the computer to pick for them. Others that are more serious, try out draft strategies, such as Stud WR by team 8 in1/2/3, that they may or may not actually use in a real draft. I think a more accurate gauge would be to compile data from actual drafts. 2) You data set makes it impossible for the first team to lose. WHy? Because over the last three years the #1 back has dramatically outscored other players (THis is true VBD rearing it's head.) THe team that drafts the WR1 reallly isn't doing badly. I am uncertain, due to how you calculate the math, if their really is any difference betweenteams 3 and 9. IN any case, you need to elimate team 1 from discussion, because no one can catch him. Everyone is playing for second place.Also, let's try to take some of your theory and apply it in a real world way. If I was drafting at 1.8 and I knew if I drafted a runningback that he would definitely finish 8th in points, and I knew at the turn I would take the place who would score best there, I probably would not draft Harrison or Owens at 1.8. I would draft the player with the highest VBD based on definite numbers, and I don't think I could lose. But the reality is, after the first 7 running backs are gone, for real people who are drafting by projecting into the future, they may take Harrison because he is a much surer bet than any running back chosen there. You have elimantaed any risk for the person drafting a rb at 1.8. Reality shows that there is a lot of risk there. Unlike most every other player, including Warner, Faulk, Edge, Gonzo, Martin, George, Green, etc, Harrison and Owens have been locks at their position. Your presumption, where you give AVT points, takes away the huge advantage in drafting Harrison- in that he is always close to the top in points. One the other hand, it was Priest who put up Faulk like numbers last year, but I didn't see anyone taking Priest at 1.1.3)No decimal points of numbers. I realize that you have already done a tremendous amount of work, but by rounding nubers and having some teams fall several places because of 1 point when you have not carried out the math to at least one decimal place is bad science. 4) You have a te required in one league and not in another. Apples to oranges again. Either require a TE in both or don't. A scientific experiment is wher you only change one variable. You have changed two. Now this infoo may not change anything, but it poor scientific method. THis makes it impossible to have meaningful conversation about how these drafts relate to each other.SOme problems with your conclusions. (I will act as if team one does not exist, because not team can compensate for the @1rb numbers if drafting is AVT based)1)You pick and choose results to talk about. You never mention that team 11 that drafted a stud wr is round 2 did very well in each league. You don't mention that the teams that did the worst in the start 3 league have the worst receivers. You don't mention that the team 4 in the start three league does very well and has excellent wr's- #6,14 and 17.This tends to support the idea that wr's are important in a start 3 league-not reject it.More tomorrow

 

[BassNBrew]

then why do you keep taking the bait? 

I also don't find your argument to be all that clear or convincing. It appears you are basically arguing that the VBD way is the only possible way to determine a player's value. Yet, you also state that "mock drafts are meaningless for the simple fact that we dont know how the players will perform". If you don't know how players will perform, then what is the point of making projections, which are later used to determine the player values that you keep referring to?  :wacko:

I'm sure there are many factors (both subjective and objective) that could influence player 'value'. Performance relative to others at their position in terms of year-end statistics are simply one individual factor among the many. Mock drafts, when done well, are clearly a useful way of understanding the 'value' of players to many other fantasy owners. I believe you lose some credibility by calling them "meaningless".

i dont do projections. i use several years of past results to get an average of how each position will score. avt tells you historcally how positions perform for a given scoring system. this combined with your starting roster requirements will produce a baseline. compare the scoring differences of the different positions to determine what the best value is at any point in a draft. mocks mean nothing to me casue they change all the time. but, if i have a player ranked wr9 and i move him up to wr6 due to subjective and objective factors, i still know , historically, how wr 6 will do, and can assign him a value.

thats what value is. i guess i am saying that avt/vbd is the only way to define value. you can alter your baselines, but ultimately value is determined by the scoring differences between different positions. maybe it was clear

value isnt, in a 2wrl, wr1 scores a greater %age of a teams points than a 3wrl, so stud wrs are more valuable in a 2wrl.

value isnt solely looking at the wr scoring and not comparing it to other positions

you are right, i shouldnt take the bait, but i like helping people.

hope it helps.

If you read my early post, you will see that the mocks spit out virtually a perfect AVT model for the RB/WR postion when using the data for the last three seasons. That had me astounded. I'm not talking close in most instances, I'm talking exact. Occasionally the point transitions varied by a spot or two, but most often it was dead on.example: r29=9, w21=9, w22=9, w23=9, w24=9 r30=8, r31=8.

Here's the thing about AVT. You can construct a perfect draft. Assuming that everyone drafts to form (in theory of course), some owners are destined to lose regardless of who they select because of their assignment within the serpent. Now I'll look at it further, but I don't see a case between these two models where taking WR1 earlier is a benefit.

 

[BassNBrew]

Let me ask you one question.  Based upon the mocks above (and they are very similiar to the mocks and expert rankings I've seen around here) and the avg. points tied to each spot, where would you draft WR1 in a S2L and a S3L (standard scoring)?  Also, based upon your scoring system, where does WR1 go in both leagues?

Bass-I have been in a car for 9 hours today and my brain is fuzzy- so I may make a comment or two that I completely recant on later. However, it is IMO, the smartest people that can changes their minds and admit later when they are wrong, so I will say what I think now.First, I think you have done a tremendous amount of work in an attempt to investigate a situation where you are going to get verbally hammered. Then you put it out into the world. Whether I agree with your results or not, I want to thank you for your efforts.I have some issues with your assumptions and some issues with your conclusions. Here are some problems I have (and unfortunately for some points I do not have a suggestion for a better solution)Problems with process:1) First you are synthesisizing two very different things : AVT (which is not VBD) and a mock draft and asserting that your results prove something in the world.I think this is a stretch. I think this can lead people to try to think about things differently, but ceertainly this offers no proof. I believe that the mock drafter who does not follow AVT, drafts a player (sometimes) on what he projects him to do the next year, not what the 5th running back taken has historically done in that situation. Furthermore, mock drafts are not a good indicator of how people will actually draft. Many people check out, allowing the computer to pick for them. Others that are more serious, try out draft strategies, such as Stud WR by team 8 in1/2/3, that they may or may not actually use in a real draft. I think a more accurate gauge would be to compile data from actual drafts. 2) You data set makes it impossible for the first team to lose. WHy? Because over the last three years the #1 back has dramatically outscored other players (THis is true VBD rearing it's head.) THe team that drafts the WR1 reallly isn't doing badly. I am uncertain, due to how you calculate the math, if their really is any difference betweenteams 3 and 9. IN any case, you need to elimate team 1 from discussion, because no one can catch him. Everyone is playing for second place.Also, let's try to take some of your theory and apply it in a real world way. If I was drafting at 1.8 and I knew if I drafted a runningback that he would definitely finish 8th in points, and I knew at the turn I would take the place who would score best there, I probably would not draft Harrison or Owens at 1.8. I would draft the player with the highest VBD based on definite numbers, and I don't think I could lose. But the reality is, after the first 7 running backs are gone, for real people who are drafting by projecting into the future, they may take Harrison because he is a much surer bet than any running back chosen there. You have elimantaed any risk for the person drafting a rb at 1.8. Reality shows that there is a lot of risk there. Unlike most every other player, including Warner, Faulk, Edge, Gonzo, Martin, George, Green, etc, Harrison and Owens have been locks at their position. Your presumption, where you give AVT points, takes away the huge advantage in drafting Harrison- in that he is always close to the top in points. One the other hand, it was Priest who put up Faulk like numbers last year, but I didn't see anyone taking Priest at 1.1.3)No decimal points of numbers. I realize that you have already done a tremendous amount of work, but by rounding nubers and having some teams fall several places because of 1 point when you have not carried out the math to at least one decimal place is bad science. 4) You have a te required in one league and not in another. Apples to oranges again. Either require a TE in both or don't. A scientific experiment is wher you only change one variable. You have changed two. Now this infoo may not change anything, but it poor scientific method. THis makes it impossible to have meaningful conversation about how these drafts relate to each other.SOme problems with your conclusions. (I will act as if team one does not exist, because not team can compensate for the @1rb numbers if drafting is AVT based)1)You pick and choose results to talk about. You never mention that team 11 that drafted a stud wr is round 2 did very well in each league. You don't mention that the teams that did the worst in the start 3 league have the worst receivers. You don't mention that the team 4 in the start three league does very well and has excellent wr's- #6,14 and 17.This tends to support the idea that wr's are important in a start 3 league-not reject it.More tomorrow

Great points for hip shots as you infer...I will address what I can from the hip and dig into some of the complicated matters tomorrow.First of all thanks for the comments. Even if someone flat out convinces me that I'm barking up the wrong tree, I have learned a lot.Item 1) The mocks were a compilation of over 100 mocks in total. Yes some crazy theories may have been applied in some drafts, however the end result matchs much of what I've seen in shark drafts. The average mocks come from serious mocks only, so computer control isn't much of an issue. Also (for whatever reasons), the mocks almost lineup exactly with AVT.Item 2) I agree with everything you say. If you draft RB1 (and he plays during the playoffs...Holmes, cough, cough) and you have a respectable draft every else, you win. The first spot has the best opportunity to draft RB1. Everyone elses chance to do that are slightly diminished at each sucessive position. I also agree that Harrison's "slam dunk" factor needs to be considered. Does an owner see RB2 still left on the board and how confident are they about it in slot 8? But does this really change from a S2L v. a S3L. Personally I don't think so. I think that a "slam dunk" is more important when your only dealing with 2 slots v. 3 slots.Item 3) You're correct about the decimal places. I couldn't justify the extend calculations from a time standpoint. I felt that if we can agree that the numbers are close (within a point or so), that it would be enough to demonstrate that's there's no significant statistical advantage to taking WR1 earlier because you start more WR.Item 4) I knew the TE required v. no TE wasn't the perfect scientific example. However it was the only example that I had that captured numerous mocks. I felt the trade off of getting numerous mocks averaged outweighed the downfall. It was a method suggested by MT or someone else here that I greatly respect. If I could get there another way, I would.Regarding my conclusions, I haven't looked at every angle yet. Personally I wanted you guys to look at it , kick it around, and spit out some thoughts like you did. Furthermore, I still haven't optimize the teams. I will do that and see if I can improve them. I guess my bottom line is that I don't see anything that jumps out saying that WR1 should be drafted earlier in a start 3 league. After I manipulate some of the teams, I may find something. What I was expecting based upon the baseline counter arguements was a clear indication that WR1 should be drafted 3-5 spots earlier in the S3L. I don't see that.Now off on another tangent. Let's assume that I totally wrong and that standard VBD with the worst starter baseline is the end all. If the rest of the league doesn't draft accordingly, then what good is the baseline. I'll point to the FanEx draft and Dodds comments (I realize the differences in the leagues). He stated that even though WR offered more value based upon VBD at his draft position, the scarcity of the RBs forced him to draft RBs. In reality, based upon AVT, there were plenty of RBs that would have been available late, he just wasn't confident in his ability to find them (like most of us). Thanks for your comments. I've taken away a lot from most every post here.

 

[Maurile Tremblay]

I believe that the mock drafter who does not follow AVT, drafts a player (sometimes) on what he projects him to do the next year, not what the 5th running back taken has historically done in that situation.

Lots of good points, RBF. A small note that is relevant to BnB's methodology: AVT doesn't know how the 5th RB taken has historically done. It only knows how the fifth RB in the final standings, based on hindsight, has historically done.BnB, I don't really have any comments to add because I'm still unsure what conclusions you are drawing, and how you are drawing them. To show whether Harrison is more valuable in a 2-WR versus a 3-WR league, I think you would have to look at how Team 5 would have fared by selecting Harrison versus how Team 5 would have fared by selecting somebody else -- in both a 2-WR league and a 3-WR league -- and holding everything else constant to the extent possible. The mock drafts don't allow you to do that.

I can't think of a good way to empirically test your theory. I think thought experiments will have to substitute for actual experiments on this issue, and the best thought experiment that's been proposed so far (IMO) is GregR's. Imagine an optimal draft in a 2-WR league in which each of the 12 drafters always takes the best guy available given his team needs, etc. You're likely to see something like 42 WRs get drafted over 16 rounds. Now suppose you expand the draft to 18 rounds and have each team start 4 WRs instead of two. There will be at least 24 extra WRs drafted. You will agree, won't you, that WR #42 should be taken earlier in this draft than he was in the 2-WR-league draft? He may have gone in the 15th or 16th round in the first draft, but now he will probably go in the 13th or so (at the latest), since 24 more WRs still have to go after him and there are only 18 total rounds.

For the same reason, the #41 WR will also go earlier in the second draft. And the #40 WR.

Everyone is drafting optimally, so these aren't mistakes -- the WRs should go earlier in the second draft. All of them, from #1 to #42 should go earlier (or, at worst, in the same place). They all get bumped up.

This is consistent with static-baseline VBD, and it is consistent with dynamic-baseline VBD. It is consistent with GregR's quite reasonable thought experiment. It is consistent with how people actually draft. It is consistent with the very simple, two-team, two-position draft I proposed on page 2 of this thread.

Your position is inconsistent with all of those things.

And if that doesn't convince you, here's one more argument, a reductio absurdum: if Harrison becomes more valuable when you reduce the number of starting WRs in your league, he ought to be most valuable of all in zero-WR leagues.

 

[heavy-set1]

If you read my early post, you will see that the mocks spit out virtually a perfect AVT model for the RB/WR postion when using the data for the last three seasons.  That had me astounded.  I'm not talking close in most instances, I'm talking exact.  I don't see a case between these two models where taking WR1 earlier is a benefit.

AVT is based on actual stats that have already happened. historically, you can see what the best value is based upon a particular draft slot. this isnt about ficticious rosters and how teams draft and what you can end up with. this is about value. and value is the scoring difference between players of the same position and then comparing that scoring difference with the scoring difference of players from other positionsfrom my earlier example, wr1, 168 value, 4th overall draft position. (3wrl)wr1, 118 value, 12th overall draft position, (2wrl)3wrl, 10 of the top 20 values are wrs2wrl, 2 of the top 20 values are wrs.dont see how thats close or exact.

 

[aaronr28]

the info iam using has already happened, so its not wrong. i think i have clearly documented its usefullness, is it perfect? of course not, but the #s are the #s, and unbiased conclusions are drawn from them.

but, how do you determine who is your RB1, RB2, etc. each year? That is a very subjective process and based on predictions. Even if you don't make actual statistical predictions, you are inherently predicting which players will do better than others, and that makes it very easy for bias to enter into your drafting decisions. Doesn't it?

 

[aaronr28]

AVT is based on actual stats that have already happened. historically, you can see what the best value is based upon a particular draft slot. this isnt about ficticious rosters and how teams draft and what you can end up with. this is about value. and value is the scoring difference between players of the same position and then comparing that scoring difference with the scoring difference of players from other positionsfrom my earlier example, wr1, 168 value, 4th overall draft position. (3wrl)wr1, 118 value, 12th overall draft position, (2wrl)3wrl, 10 of the top 20 values are wrs2wrl, 2 of the top 20 values are wrs.dont see how thats close or exact.

I believe what BnB is questioning is whether that is a sound strategy. For example, even though your model suggests that Marvin Harrison (WR1) is more valuable and should be drafted 8 points higher in a S3L than in a S2L, what is the impact on the team? Taking him that high would mean you are passing up on a top 4 RB. If WRs are going to go 10 out of the top 20 in your S3L, then I'd be happy to take 2 of the top 10 RBs in the first 2 rounds and fill my WR spots later.I see a possible problem here: You are basing your argument on draft 'value', while BnB is basing his argument on post-draft 'value'. By doing an ideal mock draft and evaluating AVT numbers, he is trying to evaluate whether a team in a S2L or a S3L benefits the most from taking a stud WR. The actual draft positions that you are talking about in your value argument are almost irrelevant to his.

 

[heavy-set1]

but, how do you determine who is your RB1, RB2, etc. each year? That is a very subjective process and based on predictions. Even if you don't make actual statistical predictions, you are inherently predicting which players will do better than others, and that makes it very easy for bias to enter into your drafting decisions. Doesn't it?

no, i pick who i think will finish where. if i am wrong, it doesnt change the fact that historically, rb1, rb2, rb3 .... scores X amount of points. i can still compare the scoring differences between positions to determine the best value on the board.

 

[Greg R]

Alright, I've now crunched the numbers for for a S2L and a S3L.  Here's whst I've noticed.1. Despite what people claim, WR are not drafted significantly earlier in a S2L v. a S3L.  Basically the drafts mirror each other until such time that a the rest of the roster has been filled and some RB/QB back-ups are in place.  Then the 3rd WR is drafted.  Those that draft a 4th WR gain little advantage because the drop off to the next tier of WR is neglible and lost to the drop off at the other positions.  It doesn't matter what the baselines say if people are ignoring them.

I don't agree with the conclusions drawn. It's good work putting it all together, just the conclusions part is the problem. What should we be looking for here to back up BnB's position? WRs from 1-24 should be drafted in the same spot or later when a 3rd WR is added.What should we be looking for here to back up VBD? WRs 1-24 should be drafted in the same spot or earlier when a 3rd WR is added, and those at the bottom of the curve ( near WR24) should see more of a change than WRs at the top (near WR1). Below is BnB's data, showing where each WR from 1-24 was drafted in the 2 WR league, in the 3 WR league, and the change. If the change is negative, it means the WR increased in value and moved up to a higher draft spot. If positive, he lost value and moved down.

Player     2 WR        3 WR     ChangeWR1          10             9             -1WR2          11            11              0WR3          17            16            -1WR4          26            25            -1WR5          27            27              0WR6          28            28              0WR7          29            31              2WR8          32            33              1WR9          34            34              0WR10        36            37              1WR11        37            38              1WR12        43            39            -4WR13        44            43            -1WR14        47            45            -2WR15        49            47            -2WR16        50            49            -1WR17        51            52              1WR18        54            53            -1WR19        59            55            -4WR20        63            56            -7WR21        64            58            -6WR22        66            59            -7WR23        68            60            -8WR24        69            61            -8-----------------------------------------Overall difference:    -48Average difference:    -2

The results seem pretty indisputable that in BnB's data, WRs are going earlier in a 3 WR league. We also see the other trend that you'd expect if drafters are seeing player value the same as VBD. There is an obvious trend that players near the bottom of the position gain more value from adding another WR than players at the top gain.BnB's experiment shows that the real world is behaving exactly how the thought experiment I proposed said it should be.(Edit to make the numbers line up for easier viewing)

 

[Maurile Tremblay]

no, i pick who i think will finish where. if i am wrong, it doesnt change the fact that historically, rb1, rb2, rb3 .... scores X amount of points.

AVT is still a form of projections. It generally over-projects the top prospects at each position, but if it does this by the same amount (percentage-wise) across all positions, there is no harm in that. There are reasons to suspect, however, that some positions (namely, the less predictable ones) are over-projected at the top by a greater percentage than others. Also, it doesn't know how many stud TEs there will be this year. If the data for your AVT was gathered during the Gonzalez years, it will think there is only one stud TE this year, not three.

 

[Greg R]

AVT is still a form of projections. It generally over-projects the top prospects at each position, but if it does this by the same amount (percentage-wise) across all positions, there is no harm in that. There are reasons to suspect, however, that some positions (namely, the less predictable ones) are over-projected at the top by a greater percentage than others. Also, it doesn't know how many stud TEs there will be this year. If the data for your AVT was gathered during the Gonzalez years, it will think there is only one stud TE this year, not three.

That's what I found when I used the same set of real results for input into VBD and AVT. The players at the top of each position were poorly ranked vs other positions, and TE1 was glaringly bad, off by nearly a round from where he should have been.Once you got down into the linear part of each positional curve, you couldn't tell a meaningful difference between AVT and VBD using real season results.

 

[heavy-set1]

AVT is still a form of projections. It generally over-projects the top prospects at each position, but if it does this by the same amount (percentage-wise) across all positions, there is no harm in that. There are reasons to suspect, however, that some positions (namely, the less predictable ones) are over-projected at the top by a greater percentage than others. Also, it doesn't know how many stud TEs there will be this year. If the data for your AVT was gathered during the Gonzalez years, it will think there is only one stud TE this year, not three.

agreed. in a nutshell its a useful guide, not fact. as for tes, i dont bother with them AVT/vbd wise. once you have the core of your team and the next several players at qb/wr/rb become interchangeable, i look at te. its more of a "feel" thing for me. once you see the needs of the players around you and how the draft is unfolding, you just know when its time for a top te. hope that makes sense.

 

[Greg R]

its more of a "feel" thing for me. once you see the needs of the players around you and how the draft is unfolding, you just know when its time for a top te. hope that makes sense.

Yep, makes perfect sense. You subconsciously go through the same mathematical evaluation that VBD folks do consciously.

 

[aaronr28]

no, i pick who i think will finish where.

that is inherently a subjective (and possibly biased) prediction. Just because RB3 and WR1 scored a certain number of points over the past 3 seasons doesn't indicate that your 3rd ranked RB and 1st ranked WR will also score those points.

 

[aaronr28]

I don't agree with the conclusions drawn. It's good work putting it all together, just the conclusions part is the problem. What should we be looking for here to back up BnB's position? WRs from 1-24 should be drafted in the same spot or later when a 3rd WR is added.

What should we be looking for here to back up VBD? WRs 1-24 should be drafted in the same spot or earlier when a 3rd WR is added, and those at the bottom of the curve ( near WR24) should see more of a change than WRs at the top (near WR1).

Below is BnB's data, showing where each WR from 1-24 was drafted in the 2 WR league, in the 3 WR league, and the change. If the change is negative, it means the WR increased in value and moved up to a higher draft spot. If positive, he lost value and moved down.

(Data removed)

The results seem pretty indisputable that in BnB's data, WRs are going earlier in a 3 WR league.

We also see the other trend that you'd expect if drafters are seeing player value the same as VBD. There is an obvious trend that players near the bottom of the position gain more value from adding another WR than players at the top gain.

BnB's experiment shows that the real world is behaving exactly how the thought experiment I proposed said it should be.

(Edit to make the numbers line up for easier viewing)

BnB can correct me if I am wrong, but I still think you guys are missing the main point of his argument. I think what he is asking is something like this:Will drafting a stud WR help your team more in a S2L or a S3L?

There are plenty of reasons and facts to back up the fact that WRs go higher in a S3L (and nobody is disputing them here), but are the TOP 3 WRs more or less valuable? If you ignore value based on draft position for a moment and even ignore what happens at other positions to a certain extent (although I know a lot of you have a big problem with doing that), in which type of league would you MOST WANT to have a stud WR, a S2L or a S3L?

If I argue this: I think WRs in general are more valuable in a S3L, but I believe having a stud WR is more valuable in a S2L. What are the main reasons for why I would be mistaken? Is it possible that I could be correct? If so, what type of information would you accept as persuasive evidence of my argument?

 

[Greg R]

Will drafting a stud WR help your team more in a S2L or a S3L?

A stud WR in a 3 WR league helps your team more than he helps a 2 WR team that drafted him with the same overall draft pick.A stud WR taken at his correctly valued spot in a 3 WR league is just about as exactly helpful as the same player taken at the correctly valued spot for him in a 2 WR league.

The "correctly valued" spot is an earlier draft pick in a 3 WR league than it is a 2 WR league.

in which type of league would you MOST WANT to have a stud WR, a S2L or a S3L?

It's a matter of where you get him for each league. If we're saying we're going to take him with the same overall draft pick, you want him more in the S3L. If you get him at the correctly valued spot in both the S3L and the S2L (which are different spots) it doesn't matter.

If I argue this: I think WRs in general are more valuable in a S3L, but I believe having a stud WR is more valuable in a S2L. What are the main reasons for why I would be mistaken?  Is it possible that I could be correct? If so, what type of information would you accept as persuasive evidence of my argument?

You're not putting any context on the situation. The value of taking him at any spot is affected by the stud WRs value, the value of the other WRs, and the value of players at other positions you could have taken in his place. It sounds silly to say, but an amount of value is equal to a similar amount of value. The problem is that you're not defining what either quantity is. Saying you have a stud WR is no description of his value. Saying you have a stud WR in a S2L or a S3L is not a description of value.

Saying you get a stud WR with pick #5 in a S2L says you got a bad value. Saying you got him at pick #15 in a S2L says you got a good value (assuming realistic curves for RB and QB). Just having a stud WR without some frame of reference is meaningless in terms of discussing his value.

 

[aaronr28]

A stud WR in a 3 WR league helps your team more than he helps a 2 WR team that drafted him with the same overall draft pick.

A stud WR taken at his correctly valued spot in a 3 WR league is just about as exactly helpful as the same player taken at the correctly valued spot for him in a 2 WR league.

The "correctly valued" spot is an earlier draft pick in a 3 WR league than it is a 2 WR league.

It's a matter of where you get him for each league. If we're saying we're going to take him with the same overall draft pick, you want him more in the S3L. If you get him at the correctly valued spot in both the S3L and the S2L (which are different spots) it doesn't matter.

You're not putting any context on the situation. The value of taking him at any spot is affected by the stud WRs value, the value of the other WRs, and the value of players at other positions you could have taken in his place.

It sounds silly to say, but an amount of value is equal to a similar amount of value. The problem is that you're not defining what either quantity is. Saying you have a stud WR is no description of his value. Saying you have a stud WR in a S2L or a S3L is not a description of value.

Saying you get a stud WR with pick #5 in a S2L says you got a bad value. Saying you got him at pick #15 in a S2L says you got a good value (assuming realistic curves for RB and QB). Just having a stud WR without some frame of reference is meaningless in terms of discussing his value.

I asked that you ignore draft position as the primary way of determining value. I don't understand why that is so impossible for people to do here. I would also argue that there is no such thing as a "correctly valued" spot. That implies a level of certainty that is just not present in fantasy football.Take this example: You are in a dynasty league and all the players are on somebody's roster. Say you are interested in acquiring a Stud WR via trade. Would you give up more in trade to acquire Marvin Harrison in a S2L or in a S3L?

Or, if you owned Marvin Harrison, and someone inquired about his availability. Would you ask for more in a S2L or in a S3L?

I know there are many other variables like strength at other positions, etc. that would have to be factored in if the situation was real, but they are somewhat irrelevant to the basic questions being asked here.

 

[Greg R]

I asked that you ignore draft position as the primary way of determining value. I don't understand why that is so impossible for people to do here. I would also argue that there is no such thing as a "correctly valued" spot. That implies a level of certainty that is just not present in fantasy football.Take this example: You are in a dynasty league and all the players are on somebody's roster. Say you are interested in acquiring a Stud WR via trade. Would you give up more in trade to acquire Marvin Harrison in a S2L or in a S3L?Or, if you owned Marvin Harrison, and someone inquired about his availability. Would you ask for more in a S2L or in a S3L?I know there are many other variables like strength at other positions, etc. that would have to be factored in if the situation was real, but they are somewhat irrelevant to the basic questions being asked here.

what's the difference? Draft position = who you pass over to take him. Trade = who you give up to take him.It's the same factor in his value. You can't ignore it without asking for it in other terms. It's the same in both cases. You give up a higher draft pick in a S3L for him (i.e. pass over better players). You give up better players for him in a S3L trade (i.e. give up players taken with higher draft picks).

 

[Maurile Tremblay]

I asked that you ignore draft position as the primary way of determining value. I don't understand why that is so impossible for people to do here.

Instead of asking "Where should he be drafted?" to determine value, we could ask "How much would you pay for him in an auction?" Either way, the answer should be the same: Marvin Harrison is worth relatively more in a S3 league than an S2 league. (For instance, if you are indifferent between Marvin Harrison and RB9 in an S2 league, you may be indifferent between Harrison and RB7 or RB8 in an S3 league. "Indifferent" meaning willing to pay the same price for.)

 

[aaronr28]

alright, we'll just have to agree to disagree because this is starting to fee like I understand all of the points you guys are raising, and I am not disputing them. I just think they are not the answers to the particular questions I am asking.Personally, I would draft Harrison (and maybe Owens, but not Moss) higher in a S2L than in a S3L, but I'm not sure I can provide the appropriate evidence to support the wiseness of that decision. I don't think I would take him over the top 8 RBs (LT2, RW, CP, MF, SA, DM, TH, PH) in either league, but I would definitely take him at 9th overall in S2L. In a S3L, however, I'd probably take a couple other RBs ahead of Harrison so the earliest I would consider him would be 11th overall. If other owners felt the opposite, and took Harrison and other WRs higher in a S3L, then I'd be happy to suck up all the RB value and trade for a stud WR later.

 

[Greg R]

...but I'm not sure I can provide the appropriate evidence to support the wiseness of that decision.

Here is the evidence of why not to. 3 drafts. One an S3L by VBD. Another 2 that are S2L. VBD says WR1's value goes down, and one of the two S2L drafts will be by VBD. BnB/aaronr28 say his value goes up (or stays equal), so one of the two S2L drafts will keep his value equal to that of the S3L league. Then we compare points and see if he was more valuable or not.

Standard scoring system (1/10, 1/20, 4/6 TDs, -2 miscues) from Inside the Stats. Converted to FPG for players who played in at least 8 games. Ranked by position. Did VBD sheets for 2 and 3 WR leagues for the data set. Did 3 drafts, where all teams used VBD values to decide who to take (except for later when the team takes WR1 early). 1/2/3/1 and 1/2/2/1 leagues. No backups.

The first draft was the 1/2/3/1 league. WR1 was the 3rd best player overall.

  1       2          3          4          5          6         7          8         9         10       11      12

RB1     RB2     WR1     QB1     RB3     RB4     WR2     RB5     WR3     RB6     RB7     QB2

TE1     WR8    QB4       RB11   WR7    WR6    WR5     WR4    RB10    QB3    RB9     RB8

RB12   WR9    WR10   TE2     WR11   WR12   TE3     WR13  RB13    RB14   WR14   WR15

QB7     QB6    RB18    WR19  RB17    WR18    RB16    RB15    TE4     WR17    QB5     WR16

WR20   TE5     RB19   WR21    WR22   RB20    RB21    WR23    QB8     WR24    WR25   TE6

WR30   WR29   WR28   RB23   TE8     QB11    QB10    QB9     WR27    WR26    TE7     RB22

WR31   RB24    TE9     WR32    QB12    TE10    WR33    TE11    WR34    TE12    WR35   WR36

84.9     84.5        84.1     83.9     83.1     83.6     83.8     83.2     82.0     82.1     82.6     81.9

Note that team 3 got WR1, and took him with the 3rd pick, after RB2. So his individual value is "more than anyone but RB1 and RB2" whether you're talking draft or trade.VBD for a S2L says he should go 5th overall. It suggests that RB1-4 and QB1 are more valuable. You're saying that WR1 is more valuable here than he was in the last draft.

One draft will take him 5th overall. The other will assume he's the same value or better as in the S3L by taking him at the same spot.

Due to how much work it takes to get things to line up when you can't use tabs, I'm just going to post Team 3 for both drafts. Total points for team 3 are in bold.

Taking him 3rd overall (i.e. BnB)

WR1

WR5

RB13

RB20

QB9

TE12

Total points by team:  76.9   76.4   74.7   75.5   75.0   74.9   74.4   74.8   74.3   74.2   73.9 74.2

Taking him 5th overall (VBD)

QB1

WR5

RB13

RB19

WR17

TE12

Total points by team:  76.9   76.5   75.4   75.0   74.9   74.7   74.4   74.8   74.3   74.2   73.9   74.2

So the team 3 that drafted WR1 as being worth less in a S2L did better, by .7 FP a week. Not only that, but most of the points lost by team 3 were gained by the team immediately following... because a more valuable player (QB1) was passed up for a less valuable one (WR1).WR1 is worth less to your team in an S2L than in an S3L.

 

[aaronr28]

WR1 is worth less to your team in an S2L than in an S3L.

once again, I almost always play in 1pt/reception leagues...I'm assuming this would change the numbers you reported slightly.Also, it looks like every team who takes WR1 in your 3 drafts does not take a RB in the 1st 2 rounds. To me, that strategy is doomed to fail. So, the benefit of taking WR1 is very difficult to perceive because that owner probably ends up with the weakest RBs in the league.Lastly, how can QB1 be the 4th most valuable player in a league that only starts 1 QB? According to the logic being used to refute BnB's claim, the fewer starting spots for a given position the LESS valuable any player at that position will be.I guess the point is that it is much easier to predict who the WR1 will be in an upcoming season than it is to predict the player who will be RB9, RB10, RB11, etc. That certainty becomes more "valuable" to me (regardless of how other owners feel) in a S2L than in a S3L. If picking at 1.09, would you rather take a player who is 90% certain to be the WR1 in a given year, or a player who is 50% certain to be RB9? Would your preference change based on whether the league was a S2L or a S3L?

 

[Maurile Tremblay]

Lastly, how can QB1 be the 4th most valuable player in a league that only starts 1 QB?

He would have been the 1st-most valuable player in a league that starts 3 QBs.

I guess the point is that it is much easier to predict who the WR1 will be in an upcoming season than it is to predict the player who will be RB9, RB10, RB11, etc. That certainty becomes more "valuable" to me (regardless of how other owners feel) in a S2L than in a S3L.

Why?

 

[Greg R]

once again, I almost always play in 1pt/reception leagues...I'm assuming this would change the numbers you reported slightly.Also, it looks like every team who takes WR1 in your 3 drafts does not take a RB in the 1st 2 rounds. To me, that strategy is doomed to fail. So, the benefit of taking WR1 is very difficult to perceive because that owner probably ends up with the weakest RBs in the league.Lastly, how can QB1 be the 4th most valuable player in a league that only starts 1 QB? According to the logic being used to refute BnB's claim, the fewer starting spots for a given position the LESS valuable any player at that position will be.I guess the point is that it is much easier to predict who the WR1 will be in an upcoming season than it is to predict the player who will be RB9, RB10, RB11, etc. That certainty becomes more "valuable" to me (regardless of how other owners feel) in a S2L than in a S3L. If picking at 1.09, would you rather take a player who is 90% certain to be the WR1 in a given year, or a player who is 50% certain to be RB9? Would your preference change based on whether the league was a S2L or a S3L?

1 pt per reception is going to change the spread, but it won't change that WR1 is more valuable with 3 WRs. And yes, moving someone up in your perceived value because of less risk of not reaching your numbers is fine, though that isn't the reason that's been given for drafting them earlier. I think there is less risk in value, that's a good point.

 

[Greg R]

Why?

Oh, I missed the part about it would be more valuable in an S2L than in an S3L. I'm curious as to why as well.

 

[aaronr28]

If I can only start 2 WRs on a given week, I want them to be 2 very consistent performers. If I can start 3, I'm more willing to take chances on less consistent players in hopes that 2 of the 3 will go off on a given week.In a pt/rec league, Marvin Harrison (WR1) scored 384 points last year with an average of 24 points per week, and a St. Dev. of 9.08. Thus, in the majority of weeks he could be relied on to score between 15 and 33 points.My rosters are almost always deep in talent, and I'm confident in my abilities to find quality throughout a draft. But, as a result, I almost always have trouble deciding who to start on a weekly basis. In a S3L, I don't have to be as precise in picking the right starting lineup, thus having a stud WR is less important to me. In a S2L, the quality of your bench will be quite a bit better and I think it is much more difficult to decide who to start on a weekly basis. Thus, knowing I could rely on Marvin Harrison to basically be the WR1 on my team for the entire season, without getting hurt or suspended, would be very valuable.I guess it's just a personal preference thing. I tried backing it up with statistics, but I don't think they were very useful or informative so I removed most of them so as not to confuse the issue anymore.

 

[Greg R]

I guess it's just a personal preference thing. I tried backing it up with statistics, but I don't think they were very useful or informative so I removed most of them so as not to confuse the issue anymore.

I don't think it's that the statistics aren't meaningful or informative. I think they indicate that that line of reasoning isn't how things are really working.And that's ok. That's easily the better half of the value of FBGs. Constantly giving us new things to think about, and hopefully making us evaluate our beliefs. Beliefs about players, about strategies, whatever. There are many beliefs I've held that have turned out to be false. RBs are more consistent than WRs (false), players do better on turf (false), consistent in weekly scoring in the past probably means consistent weekly scoring next year (false). The list goes on and on.I hate to see anyone ignore the stats when they have a clear cut indication of something other than their original belief. I'd personally rather participate in discussions like this where my own misconceptions get pointed out with evidence. That's how I'm going to improve, by getting rid of my false beliefs.I don't mean anything personal, and hope saying this wasn't offensive. I hope it wasn't taken that way at all. I just see an opportunity for someone to improve and I hate seeing it missed. And I suck at staying quiet about things like that.

 

[aaronr28]

I don't think it's that the statistics aren't meaningful or informative. I think they indicate that that line of reasoning isn't how things are really working.And that's ok. That's easily the better half of the value of FBGs. Constantly giving us new things to think about, and hopefully making us evaluate our beliefs. Beliefs about players, about strategies, whatever. There are many beliefs I've held that have turned out to be false. RBs are more consistent than WRs (false), players do better on turf (false), consistent in weekly scoring in the past probably means consistent weekly scoring next year (false). The list goes on and on.I hate to see anyone ignore the stats when they have a clear cut indication of something other than their original belief. I'd personally rather participate in discussions like this where my own misconceptions get pointed out with evidence. That's how I'm going to improve, by getting rid of my false beliefs.I don't mean anything personal, and hope saying this wasn't offensive. I hope it wasn't taken that way at all. I just see an opportunity for someone to improve and I hate seeing it missed. And I suck at staying quiet about things like that.

hey, this thread has been informative for me as well. Honestly, I don't feel that strongly about the feeling that WR1 is more valuable in a S2L than a S3L. As I said, I just think it makes picking a starting lineup easier. But, it could very well also be the case that ignoring a RB in round 1 will eventually make your RB starting decisions tougher because you will be choosing among weaker players without an established stud.Basically, I just like to argue and discuss issues like this for fun. But, I was running out of different ways to get at the same issue. Mostly, I was playing devil's advocate here. As I've tried to say all along, I understand the arguments that have been raised in support of WRs being more valuable in S3Ls. They all make sense, and I'm sure I will draft accordingly. But, somehow, they still just aren't that persuasive to me.Perhaps the most persuasive statistic would be that teams who had Marvin Harrison last year had a better record in S3Ls than they did in S2Ls. To me, that would be the true definition of "value." If somebody could show me that, then I would not question the fact that Stud WRs are more valuable in S3Ls ever again.I'm not sure if that type of data exists anywhere, but it would be a very interesting database to compile. I think Yahoo used to do something similar where they listed the players who were owned by the most teams in the top 10% or something like that. IT would be cool to be able to compare information like that between different types of leagues to see how the "value" of players change based on scoring system and starting requirements.

 

[Greg R]

I'm not sure if that type of data exists anywhere, but it would be a very interesting database to compile. I think Yahoo used to do something similar where they listed the players who were owned by the most teams in the top 10% or something like that. IT would be cool to be able to compare information like that between different types of leagues to see how the "value" of players change based on scoring system and starting requirements.

A month ago I spent a few hours looking into how to data mine myfantasyleague.com. Use CURL to download rosters for all of their leagues to create such a database. I saw enough problems in doing so that I skipped it and went onto something else to work on. It would indeed be easier at a site like yahoo that has little customization available between leagues. It was all the different league rules that made me not try with MFL.My best guess is that the results wouldn't show a whole lot for stud WRs. The best look at different leagues that I saw happened a year or two ago on the old board. The thing that seemed to drive success of teams was whether you managed to draft or otherwise acquire players who vastly outperformed their draft position. Players like Portis, Henry, Price, Ward, etc in 2002. It didn't look at losing teams, and I'm guessing the biggest thing they have in common is not hitting many breakout players while at the same time having picks who really underperformed for the high rounds they were picked in.

 

[BassNBrew]

If I can only start 2 WRs on a given week, I want them to be 2 very consistent performers. If I can start 3, I'm more willing to take chances on less consistent players in hopes that 2 of the 3 will go off on a given week.In a pt/rec league, Marvin Harrison (WR1) scored 384 points last year with an average of 24 points per week, and a St. Dev. of 9.08. Thus, in the majority of weeks he could be relied on to score between 15 and 33 points.My rosters are almost always deep in talent, and I'm confident in my abilities to find quality throughout a draft. But, as a result, I almost always have trouble deciding who to start on a weekly basis. In a S3L, I don't have to be as precise in picking the right starting lineup, thus having a stud WR is less important to me. In a S2L, the quality of your bench will be quite a bit better and I think it is much more difficult to decide who to start on a weekly basis. Thus, knowing I could rely on Marvin Harrison to basically be the WR1 on my team for the entire season, without getting hurt or suspended, would be very valuable.I guess it's just a personal preference thing. I tried backing it up with statistics, but I don't think they were very useful or informative so I removed most of them so as not to confuse the issue anymore.

That's exactly what I've been trying to say and here are the numbers again to back that up.S1L: Harrison owner's average = 16 pt, every other owner 9-15 pt.S2L: Harrison owner's average = 26 pt, every other owner 16-29 pt.S3L: Harrison owner's average = 34 pt, every other owner 22-42 pt.In all cases, the average point difference between the Harrison owner and the other owners stays at 5 points. In terms of percentage of the other owner's average points to the Harrison owner's average points, you get the following.S1L: 11/16 = 69%S2L: 21/26 = 81%S3L: 29/34 = 85%Again, I'm talking about stud value, not the group as a whole.

 

[BassNBrew]

Alright, I've now crunched the numbers for for a S2L and a S3L.  Here's whst I've noticed.1. Despite what people claim, WR are not drafted significantly earlier in a S2L v. a S3L.  Basically the drafts mirror each other until such time that a the rest of the roster has been filled and some RB/QB back-ups are in place.  Then the 3rd WR is drafted.  Those that draft a 4th WR gain little advantage because the drop off to the next tier of WR is neglible and lost to the drop off at the other positions.  It doesn't matter what the baselines say if people are ignoring them.

I don't agree with the conclusions drawn. It's good work putting it all together, just the conclusions part is the problem. What should we be looking for here to back up BnB's position? WRs from 1-24 should be drafted in the same spot or later when a 3rd WR is added.What should we be looking for here to back up VBD? WRs 1-24 should be drafted in the same spot or earlier when a 3rd WR is added, and those at the bottom of the curve ( near WR24) should see more of a change than WRs at the top (near WR1). Below is BnB's data, showing where each WR from 1-24 was drafted in the 2 WR league, in the 3 WR league, and the change. If the change is negative, it means the WR increased in value and moved up to a higher draft spot. If positive, he lost value and moved down.

Player     2 WR        3 WR     ChangeWR1          10             9             -1WR2          11            11              0WR3          17            16            -1WR4          26            25            -1WR5          27            27              0WR6          28            28              0WR7          29            31              2WR8          32            33              1WR9          34            34              0WR10        36            37              1WR11        37            38              1WR12        43            39            -4WR13        44            43            -1WR14        47            45            -2WR15        49            47            -2WR16        50            49            -1WR17        51            52              1WR18        54            53            -1WR19        59            55            -4WR20        63            56            -7WR21        64            58            -6WR22        66            59            -7WR23        68            60            -8WR24        69            61            -8-----------------------------------------Overall difference:    -48Average difference:    -2

The results seem pretty indisputable that in BnB's data, WRs are going earlier in a 3 WR league. We also see the other trend that you'd expect if drafters are seeing player value the same as VBD. There is an obvious trend that players near the bottom of the position gain more value from adding another WR than players at the top gain.BnB's experiment shows that the real world is behaving exactly how the thought experiment I proposed said it should be.(Edit to make the numbers line up for easier viewing)

w1 = 16w2 = 15w3 = 14w4 = 13w5-7 = 12w8-12 = 11w13-18 = 10w19-24 = 9w25-33 = 8I never was claiming the group as a whole was less valuable...just the studs. If you make the cut-off at the 18th spot it's pretty much a wash. If people both believed that they were more valuable and drafted accordingly, then you would see a significant difference right away. The point in time you're seeing the difference is the exact spot in the draft where those drafting in a S2L say, "what the heck, there's a ton of supply left that is all about the same and my opponents have already drafted their starters, I can wait a while longer and go a different direction". Those in the S3L are saying, "I've drafted my RBs and now have got an opportunity to get my 3WR before some teams get a second."What's great about this game is that there's always numerous ways to intrepret the stats.

 

[BassNBrew]

Here is the evidence of why not to. 3 drafts. One an S3L by VBD. Another 2 that are S2L. VBD says WR1's value goes down, and one of the two S2L drafts will be by VBD. BnB/aaronr28 say his value goes up (or stays equal), so one of the two S2L drafts will keep his value equal to that of the S3L league.

Then we compare points and see if he was more valuable or not.

Standard scoring system (1/10, 1/20, 4/6 TDs, -2 miscues) from Inside the Stats. Converted to FPG for players who played in at least 8 games. Ranked by position. Did VBD sheets for 2 and 3 WR leagues for the data set. Did 3 drafts, where all teams used VBD values to decide who to take (except for later when the team takes WR1 early). 1/2/3/1 and 1/2/2/1 leagues. No backups.

The first draft was the 1/2/3/1 league. WR1 was the 3rd best player overall.

  1       2          3          4          5          6         7          8         9         10       11      12

RB1     RB2     WR1     QB1     RB3     RB4     WR2     RB5     WR3     RB6     RB7     QB2

TE1     WR8    QB4       RB11   WR7    WR6    WR5     WR4    RB10    QB3    RB9     RB8

RB12   WR9    WR10   TE2     WR11   WR12   TE3     WR13  RB13    RB14   WR14   WR15

QB7     QB6    RB18    WR19  RB17    WR18    RB16    RB15    TE4     WR17    QB5     WR16

WR20   TE5     RB19   WR21    WR22   RB20    RB21    WR23    QB8     WR24    WR25   TE6

WR30   WR29   WR28   RB23   TE8     QB11    QB10    QB9     WR27    WR26    TE7     RB22

WR31   RB24    TE9     WR32    QB12    TE10    WR33    TE11    WR34    TE12    WR35   WR36

84.9     84.5        84.1     83.9     83.1     83.6     83.8     83.2     82.0     82.1     82.6     81.9

Note that team 3 got WR1, and took him with the 3rd pick, after RB2. So his individual value is "more than anyone but RB1 and RB2" whether you're talking draft or trade.VBD for a S2L says he should go 5th overall. It suggests that RB1-4 and QB1 are more valuable. You're saying that WR1 is more valuable here than he was in the last draft.

One draft will take him 5th overall. The other will assume he's the same value or better as in the S3L by taking him at the same spot.

Due to how much work it takes to get things to line up when you can't use tabs, I'm just going to post Team 3 for both drafts. Total points for team 3 are in bold.

Taking him 3rd overall (i.e. BnB)

WR1

WR5

RB13

RB20

QB9

TE12

Total points by team:  76.9   76.4   74.7   75.5   75.0   74.9   74.4   74.8   74.3   74.2   73.9  74.2

Taking him 5th overall (VBD)

QB1

WR5

RB13

RB19

WR17

TE12

Total points by team:  76.9   76.5   75.4   75.0   74.9   74.7   74.4   74.8   74.3   74.2   73.9   74.2

So the team 3 that drafted WR1 as being worth less in a S2L did better, by .7 FP a week. Not only that, but most of the points lost by team 3 were gained by the team immediately following... because a more valuable player (QB1) was passed up for a less valuable one (WR1).WR1 is worth less to your team in an S2L than in an S3L.

Good work Greg and I'll agree with you in principal, but still dispute this in practice. Maybe I should admend my assertion to say that VBD is flawed and leads us to draft stud WR earlier in a S3L when in reality they shouldn't be.My biggest issue with your experiment is that it doesn't even come close to representing a real draft. Where are the back-up RBs? 12 TEs taken by the end of round 7?

Here's where I need some help. Let's say that we have a VBD top 100 for a S3L that matches your rankings. Knowing how people draft (RB heavy), if you drafted truly by VBD, you would always end up with 3 WR, 1TE and 1 QB as your first five picks. What would that leave you with in terms of RBs? Would you draft this way?

I think the answer for most would be no. Well now we've discounted VBD, the same VBD and baselines that everyone quoted as gospel when arguing against my assertion. Obviously this whole issue is not as black and white as some have made it out to be.

 

[Sack-Religious]

Here's where I need some help. Let's say that we have a VBD top 100 for a S3L that matches your rankings. Knowing how people draft (RB heavy), if you drafted truly by VBD, you would always end up with 3 WR, 1TE and 1 QB as your first five picks. What would that leave you with in terms of RBs? Would you draft this way?

I strongly disagree with your statement.VBD is a tool for determining how YOU value a player. It does not tell you WHEN to draft a player.If I valued Moe Williams as the number one back, I would not draft him with the number one pick. I would wait until I thought he would not last another round.The value of the stud receiver is not completely determined through the points that they score. A clear value of the stud receiver is that it gives you the OPTION of waiting on WR2, and WR3. That is a luxury during a draft. Once you start drafting entirely for need instead of value during a draft, you are screwed.

 

[BassNBrew]

I strongly disagree with your statement.VBD is a tool for determining how YOU value a player. It does not tell you WHEN to draft a player.If I valued Moe Williams as the number one back, I would not draft him with the number one pick. I would wait until I thought he would not last another round.The value of the stud receiver is not completely determined through the points that they score. A clear value of the stud receiver is that it gives you the OPTION of waiting on WR2, and WR3. That is a luxury during a draft. Once you start drafting entirely for need instead of value during a draft, you are screwed.

Thank you Ken. You probably just made one of the best arguments in my defense to date.

 

[rabidfireweasel]

I strongly disagree with your statement.VBD is a tool for determining how YOU value a player.  It does not tell you WHEN to draft a player.If I valued Moe Williams as the number one back, I would not draft him with the number one pick.  I would wait until I thought he would not last another round.The value of the stud receiver is not completely determined through the points that they score.  A clear value of the stud receiver is that it gives you the OPTION of waiting on WR2, and WR3.  That is a luxury during a draft.  Once you start drafting entirely for need instead of value during a draft, you are screwed.

Thank you Ken. You probably just made one of the best arguments in my defense to date.

You will need to elaborate on this.I don't see how he makes anyone's case here.

 

[Sack-Religious]

To add to my earlier post, you need to be flexible while using VBD. You need to understand that when the curve for a position is flat, you are better off drafting from a different position.In a recent draft, I took Marvin Harrison in the second round in a RB heavy draft after taking a RB at # 6 overall. My draft went RB, WR, QB, WR, QB, RB, WR, WR, WR.I ended up with my # 1, 8, 17, 19 and 22 ranked WRs. So as you can see, my WR did not suffer because I took Harrison. My RBs however did suffer, but I did get my #1 and 8 QBs. I should be able to deal a QB and WR for a better than average RB. Teams with overall balance very rarely win in FF. The teams that win are the teams that go heavy at 1 or 2 positions and end up trading to shore up their weakness.Well, this is more about draft strategy than about S2L vs. S3L. Good drafting strategy/ability is more important than the difference between S2L and S3L.p.s. My # 5 WR is Rod Gardner and none of them had less than 71 rec., 1000 yds in 1 pt rec., 1pt/10yds league.

 

[Greg R]

My biggest issue with your experiment is that it doesn't even come close to representing a real draft. Where are the back-up RBs? 12 TEs taken by the end of round 7?

For the purposes of what we're looking at, it does come close to representing a real draft. In the beginning I said it's a 7 player draft. I said there are no backup RBs. Those are simplifying assumptions that were listed.If you don't think they make the result valid, tell me why. Specifically what are the cause and effect of those assumptions that changes the part we want to look at?I didn't make those assumptions lightly. If I added the backup RBs, the effect would be that RBs would gain in value compared to no backups. However, those values would be the same in all 3 drafts. The WRs will still move up, which is the behavior we're looking for (or not wanting to see in your case). It doesn't matter if WR1 was mostly equivalent to RB3 and then moved past him when we add more players, or if we use backup RBs and he was equivalent to RB6 and moved past him when we add more. The stud WR still gains in value. The team still suffers if you take him too early. It doesn't change the factor we're interested in. Similarly, I don't expect a shortened draft (7 players, which is why TEs go in the 7th round) to change the behavior of the WRs that we're looking for. WRs are still going to move up past the players around them. The team who takes a WR when there is a player at another position with more value is still going to end up giving the team behind him (that takes advantage of that bargain player) some of what could have been his points.So how do those assumptions not make it a realistic draft?

 

[Maurile Tremblay]

All WRs from 1-24 are more valuable in a 3 WR league than in a 2 WR league. Again, simple thought experiment that shows it must be so.Picture in your mind a 1 QB, 2 RB, 2 WR draft. No backups for simplicity. Draft ends at round 5. Assume everyone drafts wisely, so that each WR has a more valuable player above him and a less valuable player below him.Now what happens if you increase it to 3 WRs and add a 6th round? If WRs 1-24 keep the exact same value, they can't move in relation to the RBs and QBs. If they lose value they should slide further in the draft. If they gain value they should move up in the draft.Now, you have 12 new WRs (24-36) to be drafted, and a new round. Those 12 WRs must be picked after WR 24 since they are obviously less valuable than him. Presumably you have WR24 being one of the last picks of round 5. So if the WRs 1-24 don't gain any value, this means a proper draft for a 1 QB, 2 RB, 3 WR league should have close to 12 WRs being taken in the last round if you think WR 1-24 hold their value constant... and if they slide, you're going to be moving up RBs and QBs past them and more than the last 12 picks will be WRs.It should be obvious to anyone in FF that this isn't right. The last RBs and QBs should slip from the end of round 5 into the end of round 6, and other QBs and RBs ahead of them will also shift down slightly. The net effect is that all WRs move up in value compared to other positions.

Bump.

 

[Marc Levin]

GB GregR

 

[AB in DC]

Some QBs and RBs would move down, and some WRs up, but not all of them. I think only those toward the end of GregR's draft would be affected. The thought experiment doesn't necessarly prove that the stud WRs (aka the WRs picked in rounds one and two) move up in value.

 

[GregR_2]

Some QBs and RBs would move down, and some WRs up, but not all of them. I think only those toward the end of GregR's draft would be affected. The thought experiment doesn't necessarly prove that the stud WRs (aka the WRs picked in rounds one and two) move up in value.

When you say "move down" or move "up", if you're saying not all players will necessarily move into a different draft spot, that is of course true. Heck, you could design a pool of players where no one changes spots by going to a 3 WR league. The size of the value gain and the size of the gap from one player to the next could be created to be whatever you want.But, all WRs gain in value, whether or not it is enough to pass the RB/QB ahead of them. The thought experiment was designed just to show that from our own experience, we can see that WRs have to increase in value, not decrease.

If you want to address how much each WR gains... they all gain the same amount of value because of going from 2 to 3 WRs, and that value gain is equal to the drop from WR24 to WR36 for a 12 team league.