Stud Rb Theory Is A Sham. (1 Viewer)

[Unlucky]

Seriously - draft strategy doesn't matter at all compared to accuracy of rankings/projections.The difference in X-values is probably 1 to 3 PPG when you are deciding which position to draft. The error in projections is AT LEAST 1 to 3 PPG for most of your players. I really need to crunch some numbers and write an article. But the bottom line is: draft players that perform well. If you still don't believe me, consider these simple examples:I drafted Randy Moss in the first round several times last year. I would have been much better off drafting Harrison. Is my draft strategy at fault? No. Harrison's value was worthy of a top 8 pick. I just picked the wrong WR.Many people drafted A. Thomas, C. Martin, and E. George ahead of McAllister last season. I had McAllister ranked ahead of them. Were they at fault for drafting a RB? No. Were they at fault for having worse rankings than me? Yes.I could go on and on, but I'm fairly confident that errors in projections far outweigh drafting errors.

 

[goblue]

I see a problem with not going stud RB...and it's evident in my FBAL dynasty league team.First pick in the draft...went RB (naturally...Ricky Williams)Second round (2:12)...went QB (Dante Culpepper)Third round (3:01)...went WR (Hines Ward)By the time my pick got to me in the fourth round, I was down to RBBC RBs. Best option at the time was Betts

Then go elsewhere Bebop.Get the best TE, K, WRs, Def....ya can't "settle" that early, it is NOT your best option.

I would never suggest waiting on a RB to grab the best K or DEF.

 

[Bri]

Seriously - draft strategy doesn't matter at all compared to accuracy of rankings/projections.

What about going with your gut because "ya just know" someone will play far better than expected?Ranking that player is a bit arbitrary no?

 

[Bri]

I see a problem with not going stud RB...and it's evident in my FBAL dynasty league team.First pick in the draft...went RB (naturally...Ricky Williams)Second round (2:12)...went QB (Dante Culpepper)Third round (3:01)...went WR (Hines Ward)By the time my pick got to me in the fourth round, I was down to RBBC RBs.  Best option at the time was Betts

Then go elsewhere Bebop.Get the best TE, K, WRs, Def....ya can't "settle" that early, it is NOT your best option.

I would never suggest waiting on a RB to grab the best K or DEF.

Depends on the scoring of course. Defenses seem to be getting more points in alot of leagues. Maybe it's the typical FF folks way of working toward adding IDP? Anyhow, best TEAM wins not best RB.I would wait too, probably, but not really my point

 

[LadyBeBop]

Then go elsewhere Bebop.Get the best TE, K, WRs, Def....ya can't "settle" that early, it is NOT your best option.

At the time, the best option was Betts. Keep in mind, this was pre-Trung, and Betts was still number one. The quality of WRs were such that I could still get a good one at 6:12. Defenses? Well, I ended up with Tampa Bay at 7:01.

 

[Bottomfeeder Sports]

A bunch of random thoughts...First going back to old Bill James argument that there was much more strategy with the DH in the American League then in the National League because by defintion strategy is what you do to differentiate yourself from the competition then the Stud RB Theory is a defense, not a strategy.Along the same lines Stud RB Theory is "by the book drafting." A rule that if you follow will keep you from losing a lot, but won't stop you from being beat. "By the book" coaching seldom wins championships.A reach is a reach! Really drafting any player earlier then his predicted value dictates by definition weakens your team. Taking a weak RB in the 2nd Round because everyone else has 2 is a reach.VBD's basic weakness (Bryant - Principles of VBD (6/7/01)) is this little caveat

It's ranking the players by their value, or where they deserve to be drafted. This is not necessarily the order in which you should draft them.

The above is a function of the Last Starter baseline. I like to think of Last Starter as the value of a player relative to the entire league. The problem is this is not the same as the value of the player to my team, nor does it account for the player's cost. The cost being the position where you must draft the player to obtain him.There are several better drafting static baselines. The simplest of which is simply Joe's "Top Secret" baseline formulas in the VBD app. I don't think those go far enough to reflect and I have two that I like better. The one I use and no one else likes so we will skip it. The other is the Average Backup. That is the average between the last starter and last player drafted. Since this is pretty much into the straight line slopes you can simply use ([last starter] plus [last drafted]/2) and be close enough. We really don't have appropriate tools for Dynamic or Sliding Baselines at this time. While this would be the way to go to maximize your team, the reality is if you have a minute to be making a pick you don't have time to analyze a dozen teams previous n picks, making a half dozen calculations, etc. to make the pick.On the same lines, while you need to be flexible and able to change directions at any second during the draft that flexibility needs to be a function of your draft preparation. It is unlikely that a spur of the moment decision on a spur of the moment thought during the heat of the draft is going to be wiser then the decision made during the hours, days, and months leading into draft day.In the leagues I've been in the teams that do the best are the teams that score below average number of points the fewest times. While this may seem very obvious I don't think it really is. Basically this is saying you want a team that shows up every week. To have this means you need balance in all the positions. You need a solid core. You still need the players with upside to go off at the right times, you just can't have a team full of sleepers with potential.I'll quit now...

 

[bagger]

369 posts in a day! Wow! He certainly has a lot of free time on his hands

As far as the Stud RB Theory goes, I used to be a huge backer of the system and used it for many years. However that was before it became so popular and I find in this day and age it is less effective than it once was. If everyone selects a RB in the first two rounds, it really benefits the guys at the top of the draft (slots 1 thru 5) because they are getting the true "Stud" running backs while the rest of the owners are playing catch-up to them.

VBD is definitely the way to go but there are still some guidelines I follow...

1) – Running backs are tweaked upwards in my rankings to reflect the scarcity of legitimate starting running backs.

2) – Quarterbacks are tweaked downwards in my rankings because of the depth at the position. It is much easier to get a solid starting fantasy quarterback in round five of the draft than to try and find a starting running back in that stretch.

3) -I leave defenses and kickers out of my rankings and wait towards the end of the draft to select them. Both are very tough to project year to year and I like having QB, RB, WR and TE filled with starters and at least 1 backup before looking at drafting a defense. Kickers I always take in the bottom 20 % of the draft. That is the only place a kicker truly has value.

4) – If I do go quarterback in rounds 1 or 2, I will ignore the position altogether until at least round 8. There is no point burning another pick before the 8th round on a quarterback.

5) – If I go through the first two rounds and do not pick a running back (i.e. – WR/WR or WR/QB) then it is imperative to get at least 3 RB’s in the next four rounds before all of the potential backs are gone.

So to summarize, STUD RB is not the ultimate draft strategy it once was but running backs are still the most important aspect of your fantasy roster simply because there are less of them to go around.

obviously VBD does not tell you when to draft but best value when you're drafting. when you "tweak" your players are you using a player's ADP or a dynamic baseline based on positions taken in the next round to do this? or do you just use gut?

 

[bagger]

We really don't have appropriate tools for Dynamic or Sliding Baselines at this time. While this would be the way to go to maximize your team, the reality is if you have a minute to be making a pick you don't have time to analyze a dozen teams previous n picks, making a half dozen calculations, etc. to make the pick.

i disagree with this statement. if you have access to a laptop during the draft this is not an issue as you should be tracking this during other owners' picks.

 

[Greg R]

What about going with your gut because "ya just know" someone will play far better than expected?Ranking that player is a bit arbitrary no?

That's kind of impossible. I don't mean that you can't "know that someone will play" well. But if you expect them to play that well, that is your expectation. If they match your expectation, then how is their ranking invalid? Unless you're using someone else's rankings, in which case there are bigger problems with that draft strategy than what were talking about now.

Seriously - draft strategy doesn't matter at all compared to accuracy of rankings/projections.

I believe you're absolutely right about the magnitude of the errors. During the AVT debates I did some studies that took FBG projections for 2001, true results for 2001, and ran them through both AVT and VBD. Comparing results, it was pretty clear that the errors in projections was much greater than the difference between AVT and VBD (with the possible exception of the top of each positional curve in AVT).At some point though, I think there is a usefulness for draft strategy though. As you (or was it MT?) said earlier, many new FFers overvalue QBs. I doubt there are often huge differences in quality from one set of projections to another in terms of accuracy, not when we're talking about 12 different sets in a given league. So I don't think going through the steps is a bad thing. But it would be good to keep in mind the possible error in the projections. That's one thing I've been advocating here for awhile, a user-friendly method that will allow you to have range of projections along with some measure of your certainty. A method that can display this in a manner that can help you make your decisions.

 

[Maulers]

Fascinating.Reading all this just makes me want to and then some more.Honestly. Now I need some Tylenol.

 

[Anarchy99]

I see a problem with not going stud RB...and it's evident in my FBAL dynasty league team.First pick in the draft...went RB (naturally...Ricky Williams)Second round (2:12)...went QB (Dante Culpepper)Third round (3:01)...went WR (Hines Ward)By the time my pick got to me in the fourth round, I was down to RBBC RBs. Best option at the time was Betts (this was before the Trung signing ), and I immediately followed that up with Duckett. (And this is a league with a flex RB/WR position. So chances are good that I'll be starting 4 WRs...unless Candidate, Dunn and/or Edge go down).Point is...you need two RBs early, else you'll be starting the wrong end of a RBBC.

Being in Lady BeBop's league, I can report that RBs went EXTREMELY fast, especially since teams can play 3 RB each week.No matter what system/approach/strategy you want to label it, here is my two cents. RBs to me are a lot like playing musical chairs. Like Lady BeBop mentioned, you want someplace to sit when the music stops, because all the chairs will be gone pretty quickly. There are not that many quality RB, and there are not that many starting NFL RB to begin with. After 3 rounds, there is a high percentage of dead weight at RB to chose from, and a league of sharks will not normally let any good ones slip very far.I was in a league last year where someone picked up TO, Moss, and McNabb in the first 3 rounds and got PUMMELED each week with simply AWFUL RB performance. I think he wound up with Pittman and Allen as his RB. He finally traded one one of his WR but by then it was too late.As others have suggested, it is always easier to command a trade if you have a stable of RB than it is being on the shortend trying to obtain one.In terms of STUD RB, the secret is having a top 5 pick and getting a true stud but being able to select from the remaining RB talent at the end of round 2 and picking out the one RB that that "slipped through the cracks" that no one else picked that also could be near a top 5 back. But not many of us have crystal balls to predict with.

 

[bagger]

I see a problem with not going stud RB...and it's evident in my FBAL dynasty league team.First pick in the draft...went RB (naturally...Ricky Williams)Second round (2:12)...went QB (Dante Culpepper)Third round (3:01)...went WR (Hines Ward)By the time my pick got to me in the fourth round, I was down to RBBC RBs. Best option at the time was Betts (this was before the Trung signing ), and I immediately followed that up with Duckett. (And this is a league with a flex RB/WR position. So chances are good that I'll be starting 4 WRs...unless Candidate, Dunn and/or Edge go down).Point is...you need two RBs early, else you'll be starting the wrong end of a RBBC.

was ward really the best value play? was tiki barber not available?

 

[TylerDurden]

Seriously - draft strategy doesn't matter at all compared to accuracy of rankings/projections.The difference in X-values is probably 1 to 3 PPG when you are deciding which position to draft. The error in projections is AT LEAST 1 to 3 PPG for most of your players. I really need to crunch some numbers and write an article. But the bottom line is: draft players that perform well. If you still don't believe me, consider these simple examples:I drafted Randy Moss in the first round several times last year. I would have been much better off drafting Harrison. Is my draft strategy at fault? No. Harrison's value was worthy of a top 8 pick. I just picked the wrong WR.Many people drafted A. Thomas, C. Martin, and E. George ahead of McAllister last season. I had McAllister ranked ahead of them. Were they at fault for drafting a RB? No. Were they at fault for having worse rankings than me? Yes.I could go on and on, but I'm fairly confident that errors in projections far outweigh drafting errors.

This is not so much an argument about the Stud RB theory as much as it is the pointlessness of projections.Seriously, does ANYBODY really believe their own projections are anything better than an educated guess? All it does is quantify your own qualitative feelings, which makes them anything but scientific. Projected stats are only one factor to consider when drafting a player. You also have to consider the expected variance of those stats (what are the odds this player will totally bust or, more happily break out?) and what kind of value this player might provide as trade bait. And once you start to factor in those "tweaks," you're no longer doing "pure" VBD based on X-values. You're using your head instead of relying on a silly set of projections.

 

[Marvelous]

I am agreeing whole heartedly with Smoo. I think my league is overvalueing rbs and qb (thankfully) in the early rounds. We have a 10 team league and start 1 qb, 2 rb, 3wr/te, 1k, 1d. I picked 8th. Here are the results of the draft and how deep into the pool the player was:1. Owens (2nd wr)2. Harrison (3rd wr)3. Taylor (13th rb)4. Davis (16th rb)5. Kevin Johnson (18th rb)6. Tiki (20th rb)7. Henry (22nd rb)8. Bennett (26th rb)9. Vick (12th qb)The point here is that the best value will be going against the trend of the league. I followed the VBD application fairly closely for valuing players and then used judgement and other mock drafts to figure where other people may draft them. I was able to grab a lot of rb when everyone else was drafting all their wrs. Also note, VBD is all based on projections. With 5 decent rbs like this, if one finishes in the top 10 and another in the top 15, you will be doing ok.

 

[TylerDurden]

Being in Lady BeBop's league, I can report that RBs went EXTREMELY fast, especially since teams can play 3 RB each week.No matter what system/approach/strategy you want to label it, here is my two cents. RBs to me are a lot like playing musical chairs. Like Lady BeBop mentioned, you want someplace to sit when the music stops, because all the chairs will be gone pretty quickly. There are not that many quality RB, and there are not that many starting NFL RB to begin with. After 3 rounds, there is a high percentage of dead weight at RB to chose from, and a league of sharks will not normally let any good ones slip very far.

This is true. In good leagues, RBs go fast and hard. Unless you have a top-5 pick or so, you are going to "overpay" for a RB no matter when you draft one. For example, is Corey Dillon (say) really worth a late 1st round pick over Harrison or Owens? VBD says no. But if you don't overpay for Dillon now, you're going to be stuck overpaying for James Stewart or Trung Canidate later on. Knowing that you're going to have to overpay for your RBs anyway, why not overpay for somebody who you know will actually be a quality FF starter, as opposed to someone who figures to be a liability?I have literally never been in a league that I can recall where the wiseguy who took Moss/Owens at 1.12 and 2.01 ended up having a good season. That owner almost invariably ends up with crap at RB which dooms him over the course of the season.

 

[TylerDurden]

I am agreeing whole heartedly with Smoo. I think my league is overvalueing rbs and qb (thankfully) in the early rounds. We have a 10 team league and start 1 qb, 2 rb, 3wr/te, 1k, 1d. I picked 8th. Here are the results of the draft and how deep into the pool the player was:1. Owens (2nd wr)2. Harrison (3rd wr)3. Taylor (13th rb)4. Davis (16th rb)5. Kevin Johnson (18th rb)6. Tiki (20th rb)7. Henry (22nd rb)8. Bennett (26th rb)9. Vick (12th qb)The point here is that the best value will be going against the trend of the league. I followed the VBD application fairly closely for valuing players and then used judgement and other mock drafts to figure where other people may draft them. I was able to grab a lot of rb when everyone else was drafting all their wrs. Also note, VBD is all based on projections. With 5 decent rbs like this, if one finishes in the top 10 and another in the top 15, you will be doing ok.

Notice that you benefited tremendously by having brain-dead owners in your league. Taylor at 3.08, Tiki in the 6th (!!), Henry in the 7th (!!!) and Bennett in the 8th (!!) were a function not of your brilliance, but your colleagues' ineptitude. I'm sure anybody in the shark pool could roll over your league using any semi-intelligent draft strategy.

 

[blenderfrog]

If Stud RB theory = drafting two RBs that finish in the top 5 to 7 (for RBs), then it works.

So, then, if all people are trying to draft two RB's from the top 5 or 7, then, by YOUR (2nd) DEFINITION, nobody is going to be using Stud RB. Since all seven of those RBs would likely be gone before the first round ends. I think your first definition is the only one that's logically consistent and therefore will be the one I use, and therefore my conclusion that it's silly holds.

That's where the team owner's skill comes in there Smooser. I mix a little Stud RB, VDB, and gut feelings in. In my oldest league (consisting of grudge holding co-workers) last year, I applied the value-based-stud-RB theory:I picked Ahman Green at the end of the 1st round and Harrison at the top of the 2nd. Granted, Green didn't do as well as I would have liked. Culpepper slid to the 3rd and I couldn't pass him up (we get great scores for rushing QB's). I then picked up Fred Taylor in the 4th and Travis Henry in the 5th. I felt Green was a top 5, and between Taylor and Henry I could get a RB in the 5 - 10 range.

I would have preferred to grab a more guaranteed stud RB in the 2nd round, but I didn't think the available picks warranted an early 2nd pick. Stud RB theory doesn't require owners to pick RB's in the first 2 rounds. I could have taken Ant Thomas in the 2nd, but I preferred Henry and gambled that I could get him later.

 

[Bottomfeeder Sports]

We really don't have appropriate tools for Dynamic or Sliding Baselines at this time. While this would be the way to go to maximize your team, the reality is if you have a minute to be making a pick you don't have time to analyze a dozen teams previous n picks, making a half dozen calculations, etc. to make the pick.

i disagree with this statement. if you have access to a laptop during the draft this is not an issue as you should be tracking this during other owners' picks.

I have and use a laptop. I have a Spreadsheet which is a superset of what can be downloaded which dynamically keeps track of the status of every player and each team. Every year I try to implement dynamic capabilities but they breakdown under the time constraints of keeping up with a real draft. Maybe I visualize the wrong model as to how this should work. Maybe my mind is such that I over complicate the problem. Is there a downloadable tool where this is accomplished? I try to download them all - at least the samples - and adopt (steal) ideas from them. I have yet to see this implemented - at least in a free/sample tool! I still make my picks based on educated guesses as to whom will still be available at the next pick and try to cascade the implications of taking one position over another, but I don't believe that the tools available help much other then player tracking. Maybe the tools give me more time to make more calculations, but enough time to make all the calculations I would want to make. And equally important doesn't make many of the calculations for me.What are the mechanism you use to accomplish this? Is it just calculations in your head? In this case I'd love to be wrong.

 

[Unlucky]

At some point though, I think there is a usefulness for draft strategy though. As you (or was it MT?) said earlier, many new FFers overvalue QBs. I doubt there are often huge differences in quality from one set of projections to another in terms of accuracy, not when we're talking about 12 different sets in a given league. So I don't think going through the steps is a bad thing. But it would be good to keep in mind the possible error in the projections.

I agree that you have to draft for value. You can use VBD, dynamic VBD, or your gut. (which is VBD using years of experience instead of numbers) You could even have a preset position list such as: RB, RB, WR, RB, WR, QB, WR, TE which would probably produce a good team as long as your rankings are better than everyone else's.I'll say it once again, but a bit differently: Don't waste your time worrying about draft strategy. Instead, spend time researching players to improve your draft board. It's also important to study a few mock drafts and such to see where players are being drafted. Most drafts are quite similar, so you need to have an idea of how late you can get the players you want.I think it's OK to miss out on a bunch of stud players as long as the players you rank higher than everyone else (and therefore draft) actually meet your expectations.

 

[Greg R]

This is not so much an argument about the Stud RB theory as much as it is the pointlessness of projections.Seriously, does ANYBODY really believe their own projections are anything better than an educated guess? All it does is quantify your own qualitative feelings, which makes them anything but scientific. Projected stats are only one factor to consider when drafting a player. You also have to consider the expected variance of those stats (what are the odds this player will totally bust or, more happily break out?) and what kind of value this player might provide as trade bait. And once you start to factor in those "tweaks," you're no longer doing "pure" VBD based on X-values. You're using your head instead of relying on a silly set of projections.

I guess we have a differing view of what VBD is. To me it is a tool that allows comparison between players at different positions, by normalizing them based on a set of baselines.It doesn't tell me who is the best value. It helps me figure out the supply side of supply and demand. It isn't any more scientific than thinking, "Tiki vs Holt" and deciding which I think is a better value by pulling an answer out of the air. But it is more systematic. It allows me to put a lot of thought into how each player will do, and capture all of that thought in a projection. So when I compare Tiki to Holt, the final result reflects everything I thought of when I did my projections... things that you won't necessarily come back up with going with your gut.And you're right about the variance part. I've been saying for awhile I think this is the next major improvement waiting to happen. But I don't agree when you say "And once you start to factor in those "tweaks," you're no longer doing "pure" VBD based on X-values." There is no such thing as pure VBD. It's there to give you the X-values. After that what you do with it is up to you. (Edit: I suppose some might say that taking those X values and drafting straight down the resulting cheatsheet is "pure VBD", but that's a drafting strategy that doesn't get the most out of the tool.) I would argue that using a dynamic baseline approach with the results, along the lines of what MT said, is the best way to do it at present. But I don't see any reason to think that we can't add our feelings of uncertainty in our projections and still use VBD as a tool.

 

[Thread7]

I argue that the most predictable position in fantasy football is the running back. That is what makes them so valuable -> not that they score many more fantasy points than a Qb or WR. It is just that there are some guys you can count on getting 1 TD and 100 yds per game. Only two or three WR's in the league consistently get 100 yards. And even Terrell Owens often has a 200 yd / 4 TD game followed by a 60 yd / 0 TD game. Ricky Williams could have the same thing but the wide variation is much less frequent.

 

[Unlucky]

Seriously - draft strategy doesn't matter at all compared to accuracy of rankings/projections.The difference in X-values is probably 1 to 3 PPG when you are deciding which position to draft. The error in projections is AT LEAST 1 to 3 PPG for most of your players. I really need to crunch some numbers and write an article. But the bottom line is: draft players that perform well. If you still don't believe me, consider these simple examples:I drafted Randy Moss in the first round several times last year. I would have been much better off drafting Harrison. Is my draft strategy at fault? No. Harrison's value was worthy of a top 8 pick. I just picked the wrong WR.Many people drafted A. Thomas, C. Martin, and E. George ahead of McAllister last season. I had McAllister ranked ahead of them. Were they at fault for drafting a RB? No. Were they at fault for having worse rankings than me? Yes.I could go on and on, but I'm fairly confident that errors in projections far outweigh drafting errors.

This is not so much an argument about the Stud RB theory as much as it is the pointlessness of projections.Seriously, does ANYBODY really believe their own projections are anything better than an educated guess? All it does is quantify your own qualitative feelings, which makes them anything but scientific. Projected stats are only one factor to consider when drafting a player. You also have to consider the expected variance of those stats (what are the odds this player will totally bust or, more happily break out?) and what kind of value this player might provide as trade bait. And once you start to factor in those "tweaks," you're no longer doing "pure" VBD based on X-values. You're using your head instead of relying on a silly set of projections.

I am saying the exact opposite of you. I do believe that projections are VERY important. I do believe that my projections ARE better than the average person's rankings. While my projections are an educated guess, they are a better guess than other people's. Projecting all of the stats is how I form my draft board. I am using far more information than someone who just makes a list. This is where I feel I have a great advantage.I include the expected variance in my projections. I project for the "most likely" stats. If a player has more upside than downside, than his numbers will be higher. How important is "trade bait" in rankings? I don't really think it's very important. I want the best performers. I'm pretty sure the best performer is usually the best trade bait.

 

[TylerDurden]

And you're right about the variance part. I've been saying for awhile I think this is the next major improvement waiting to happen. But I don't agree when you say "And once you start to factor in those "tweaks," you're no longer doing "pure" VBD based on X-values." There is no such thing as pure VBD. It's there to give you the X-values. After that what you do with it is up to you. I would argue that using a dynamic baseline approach with the results, along the lines of what MT said, is the best way to do it at present. But I don't see any reason to think that we can't add our feelings of uncertainty in our projections and still use VBD as a tool.

I can more or less agree with that. VBD is a somewhat useful tool, but its foolish to follow it blindly.=I disagree that the variance thing is "the next major improvement waiting to happen." This is always going to be subjective, and I can't imagine how you would quantify that in any meaningful way. Sometimes going with your "gut" is the best. The human mind is still better than any computer when it comes to the fuzzy qualitative decisionmaking that you need to win at FF.

 

[Greg R]

Hey Unlucky, speaking of the accuracy of projections, are you going to do a project compilation 2002?

 

[Anarchy99]

Was Ward really the best value play at 3.01? Was Tiki Barber not available?

Tiki was long gone. Here were the first 2 rounds of picks:1.01 Ricky Williams, MIA1.02 Ladanian Tomlinson, SD1.03 Clinton Portis, DEN1.04 Priest Holmes, KC 1.05 Deuce McAllister, NO1.06 Michael Vick, ATL1.07 Shaun Alexander, SEA1.08 Marshall Faulk, STL1.09 Travis Henry, BUF1.10 Ahman Green, GB1.11 Jamal Lewis, BAL1.12 Michael Bennett, MIN2.01 Corey Dillon, CIN2.02 Randy Moss, MIN2.03 Tiki Barber, NYG2.04 William Green, CLE2.05 Donovan McNabb, PHI2.06 Terrell Owens, SF2.07 Edgerrin James, IND2.08 Peyton Manning, IND2.09 Marvin Harrison, IND2.10 Fred Taylor, JAX2.11 Curtis Martin, NYJ2.12 Daunte Culpepper, MIN

 

[bagger]

At some point though, I think there is a usefulness for draft strategy though. As you (or was it MT?) said earlier, many new FFers overvalue QBs. I doubt there are often huge differences in quality from one set of projections to another in terms of accuracy, not when we're talking about 12 different sets in a given league. So I don't think going through the steps is a bad thing. But it would be good to keep in mind the possible error in the projections.

I agree that you have to draft for value. You can use VBD, dynamic VBD, or your gut. (which is VBD using years of experience instead of numbers) You could even have a preset position list such as: RB, RB, WR, RB, WR, QB, WR, TE which would probably produce a good team as long as your rankings are better than everyone else's.I'll say it once again, but a bit differently: Don't waste your time worrying about draft strategy. Instead, spend time researching players to improve your draft board. It's also important to study a few mock drafts and such to see where players are being drafted. Most drafts are quite similar, so you need to have an idea of how late you can get the players you want.I think it's OK to miss out on a bunch of stud players as long as the players you rank higher than everyone else (and therefore draft) actually meet your expectations.

this is true.

 

[bagger]

Was Ward really the best value play at 3.01? Was Tiki Barber not available?

Tiki was long gone. Here were the first 2 rounds of picks:1.01 Ricky Williams, MIA1.02 Ladanian Tomlinson, SD1.03 Clinton Portis, DEN1.04 Priest Holmes, KC 1.05 Deuce McAllister, NO1.06 Michael Vick, ATL1.07 Shaun Alexander, SEA1.08 Marshall Faulk, STL1.09 Travis Henry, BUF1.10 Ahman Green, GB1.11 Jamal Lewis, BAL1.12 Michael Bennett, MIN2.01 Corey Dillon, CIN2.02 Randy Moss, MIN2.03 Tiki Barber, NYG2.04 William Green, CLE2.05 Donovan McNabb, PHI2.06 Terrell Owens, SF2.07 Edgerrin James, IND2.08 Peyton Manning, IND2.09 Marvin Harrison, IND2.10 Fred Taylor, JAX2.11 Curtis Martin, NYJ2.12 Daunte Culpepper, MIN

was this a redraft league? because whoever drafted portis at 3rd overall and vick 6th overall were crazy. worked out great for them (and i have them both in my deep keeper league) but the risk/reward on those two players at those draft slots is pretty hard to justify.but that's a whole different topic of conversation.

 

[TylerDurden]

I am saying the exact opposite of you. I do believe that projections are VERY important. I do believe that my projections ARE better than the average person's rankings. While my projections are an educated guess, they are a better guess than other people's. Projecting all of the stats is how I form my draft board. I am using far more information than someone who just makes a list. This is where I feel I have a great advantage.

Color me skeptical on this point. Everybody thinks they're above-average poker players, too.

I include the expected variance in my projections. I project for the "most likely" stats. If a player has more upside than downside, than his numbers will be higher.

How do you conflate a mean and variance into one number? And more importantly, why do you think it's more accurate than gut-based drafting (GBD)?

How important is "trade bait" in rankings? I don't really think it's very important. I want the best performers. I'm pretty sure the best performer is usually the best trade bait.

Not too important, admittedly. It matters if you draft early in the preseason and want to swing trades, but once the season is underway, it doesn't matter much otherwise.

 

[Greg R]

I disagree that the variance thing is "the next major improvement waiting to happen."  This is always going to be subjective, and I can't imagine how you would quantify that in any meaningful way.  Sometimes going with your "gut" is the best.  The human mind is still better than any computer when it comes to the fuzzy qualitative decisionmaking that you need to win at FF.

I don't think that what I'm talking about and what you are talking about is any different. In the end, the human mind is making the decision. I'm suggesting we prepare as much information in advance that will be useful in that decision making, and then present it in a way that lets you view it, digest it, and make your decision quickly.Some examples of varying degrees of this:If I just made a list of rankings and didn't take the time to go look up stats on each player of fantasy significance, how many rushes their OC normally has in a season, etc... I would have really underranked Michael Bennett in 2002. Because I took the time to do projections, and look up all of that information, I came a lot closer to predicting his true 2002 performance. My brain still made the decision, I just did it with more information in front of me. (Edit to add: I think this is an example of Unlucky's point. I convinced my buddy Jeremy that Bennett would do better than either of us had originally thought, because when I showed him the stats from 2001, they indicated he was better than either of us had thought, using our gut. If I was drafting against Jeremy, my projection would have been better than his for Bennett, because I did what Unlucky is saying.)Extending that to a hypothetical example with the variances... my projections which are my "most likely" number had Deuce as about the #12 RB. If I had created upper and lower bounds, the upside would have been much better than the downside, as I felt he had the potential to be a top 5 back, but I didn't think he'd underperform by much unless he was injured.Now let's say I record that variance while I do my projections and I'm thinking about Deuce. If I believe Deuce's realistic upper bound is that high, it will also play a role in Joe Horn's upper and lower bounds, on Aaron Brooks upper and lower bounds, etc. If I think about all of these collectively as a team, I can put more information into those numbers. Information, from human brain decisions, that I was able to take time to reflect on and look at from all sides.When I'm in the middle of a draft with 90 seconds before I pick and my top 2 guys I was going to take gone so I have to rethink quickly whether to take Horn or someone else... I'm probably not going to make as informed of a decision if I just wing it all from the top of my head. I'm not going to think about what I decided the distribution of runs and passes would be if Deuce was a 1500 yard rusher, vs an 1100 yard rusher. If I work this out ahead of time, I've given myself more information to make my final decision from.

 

[bagger]

And you're right about the variance part. I've been saying for awhile I think this is the next major improvement waiting to happen. But I don't agree when you say "And once you start to factor in those "tweaks," you're no longer doing "pure" VBD based on X-values." There is no such thing as pure VBD. It's there to give you the X-values. After that what you do with it is up to you. I would argue that using a dynamic baseline approach with the results, along the lines of what MT said, is the best way to do it at present. But I don't see any reason to think that we can't add our feelings of uncertainty in our projections and still use VBD as a tool.

I can more or less agree with that. VBD is a somewhat useful tool, but its foolish to follow it blindly.=I disagree that the variance thing is "the next major improvement waiting to happen." This is always going to be subjective, and I can't imagine how you would quantify that in any meaningful way. Sometimes going with your "gut" is the best. The human mind is still better than any computer when it comes to the fuzzy qualitative decisionmaking that you need to win at FF.

your misunderstanding is that if you utilize vbd properly, it's impossible to "follow it blindly" because it does not tell you who to draft when. this is the biggest misconception with the proper use of vbd.vbd tells you value across positions and the corresponding dropoff between draft picks (if you use it dynamically). it is then a player's average draft position that should be used as a cross-reference to indicate who you should draft.

highest vbd # does not = draft that player.

 

[TylerDurden]

I don't think that what I'm talking about and what you are talking about is any different. In the end, the human mind is making the decision. I'm suggesting we prepare as much information in advance that will be useful in that decision making, and then present it in a way that lets you view it, digest it, and make your decision quickly.Some examples of varying degrees of this:

I can agree with all of this. I do projections myself mainly for the reasons you described. Doing projections forces you to gather info on all players and their situations. After putting together an initial list based on those projections, I basically throw the projections away and move players around subjectively based on upside, expected variance, etc. So it sounds like we more or less agree.

 

[bagger]

Color me skeptical on this point. Everybody thinks they're above-average poker players, too.

so you choose to use other people's projections rather than your own. this cannot give you more insight. even if your projections are slightly off (which they will be) just the process of going through player by player and team by team is a learning process that is invaluable.***edit***

nevermind...just saw your above post.

 

[Anarchy99]

was this a redraft league? because whoever drafted portis at 3rd overall and vick 6th overall were crazy. worked out great for them (and i have them both in my deep keeper league) but the risk/reward on those two players at those draft slots is pretty hard to justify.but that's a whole different topic of conversation.

Newly formed dynasty league drafted AFTER the Bucs won the SB.

 

[Unlucky]

Hey Unlucky, speaking of the accuracy of projections, are you going to do a project compilation 2002?

I got too busy at the deadline last year and thus didn't compile any projections. I'll try again this season.

 

[Unlucky]

I am saying the exact opposite of you.  I do believe that projections are VERY important.  I do believe that my projections ARE better than the average person's rankings.  While my projections are an educated guess, they are a better guess than other people's.  Projecting all of the stats is how I form my draft board.  I am using far more information than someone who just makes a list.  This is where I feel I have a great advantage.

Color me skeptical on this point. Everybody thinks they're above-average poker players, too.

I think my track record over the past two seasons has proven to me that my projections are an important part of winning. I've won more than my fair share in 23 leagues over the past two years. (side note: all leagues were for money and at least fairly competitive - no freebies)

 

[TylerDurden]

Hey Unlucky, speaking of the accuracy of projections, are you going to do a project compilation 2002?

I got too busy at the deadline last year and thus didn't compile any projections. I'll try again this season.

I checked out the 2001 projection compilation on your website, and I notice that the difference in the median PPG error from the worst set of projections to the best set of projections is only about 1 PPG. In other words, the BEST set of projections was only 1 PPG better than the WORST set of projections.This seems to argue strongly against the idea that "accurate" projections can make a big difference in your FF success in good leagues.

 

[Greg R]

I can agree with all of this. I do projections myself mainly for the reasons you described. Doing projections forces you to gather info on all players and their situations. After putting together an initial list based on those projections, I basically throw the projections away and move players around subjectively based on upside, expected variance, etc. So it sounds like we more or less agree.

Yeah, I think you're right, we more or less agree. I think that the info you throw away can be useful. But then, right now I feel like I'm throwing that info away too, because I'm not truly capturing the variance in anything but my brain. That's the part I think can be the next improvement in draft strategy.I just wish I knew the best way to do it. If I did, I'd probably have an app out there for people to download by now.

 

[Greg R]

I argue that the most predictable position in fantasy football is the running back. That is what makes them so valuable -> not that they score many more fantasy points than a Qb or WR. It is just that there are some guys you can count on getting 1 TD and 100 yds per game. Only two or three WR's in the league consistently get 100 yards. And even Terrell Owens often has a 200 yd / 4 TD game followed by a 60 yd / 0 TD game. Ricky Williams could have the same thing but the wide variation is much less frequent.

Meant to comment on this earlier.Someone (Drinen?) did a study that looked at that. He found the opposite of what you said. RBs are no more consistent than the other skill positions. The reason they seem to be more valuable was that the top ones score more than the lesser ones, compared to QB and WR.

 

[Marvelous]

I am agreeing whole heartedly with Smoo. I think my league is overvalueing rbs and qb (thankfully) in the early rounds. We have a 10 team league and start 1 qb, 2 rb, 3wr/te, 1k, 1d. I picked 8th. Here are the results of the draft and how deep into the pool the player was:1. Owens (2nd wr)2. Harrison (3rd wr)3. Taylor (13th rb)4. Davis (16th rb)5. Kevin Johnson (18th rb)6. Tiki (20th rb)7. Henry (22nd rb)8. Bennett (26th rb)9. Vick (12th qb)The point here is that the best value will be going against the trend of the league. I followed the VBD application fairly closely for valuing players and then used judgement and other mock drafts to figure where other people may draft them. I was able to grab a lot of rb when everyone else was drafting all their wrs. Also note, VBD is all based on projections. With 5 decent rbs like this, if one finishes in the top 10 and another in the top 15, you will be doing ok.

Notice that you benefited tremendously by having brain-dead owners in your league. Taylor at 3.08, Tiki in the 6th (!!), Henry in the 7th (!!!) and Bennett in the 8th (!!) were a function not of your brilliance, but your colleagues' ineptitude. I'm sure anybody in the shark pool could roll over your league using any semi-intelligent draft strategy.

I am not sure they were that braindead. After getting 2 starting rbs, most people were trying to get decent wrs. When they were drafting wrs, I was drafting rbs. After the top 15 or so rbs, there was not much demand for them. Also, when starting 3 wrs as opposed to 2, there is a lot more value on wr.

 

[Unlucky]

Hey Unlucky, speaking of the accuracy of projections, are you going to do a project compilation 2002?

I got too busy at the deadline last year and thus didn't compile any projections. I'll try again this season.

I checked out the 2001 projection compilation on your website, and I notice that the difference in the median PPG error from the worst set of projections to the best set of projections is only about 1 PPG. In other words, the BEST set of projections was only 1 PPG better than the WORST set of projections.This seems to argue strongly against the idea that "accurate" projections can make a big difference in your FF success in good leagues.

I see your point - on the whole, there isn't much difference in the projections. However, I'm not sure that my method of measuring the error is all that accurate. Consider one set that overprojects Player A by 1 PPG and another that underprojects Player A by 1 PPG. They both have the same error, but their projections are quite different. Here's my bottom line: by making your own projections, certain players will grade out higher than the consensus. These are players that you will likely draft because you have them rated higher than most others. These are the players that you need to pan out since you'll be drafting them.

 

[Bottomfeeder Sports]

I argue that the most predictable position in fantasy football is the running back. That is what makes them so valuable -> not that they score many more fantasy points than a Qb or WR. It is just that there are some guys you can count on getting 1 TD and 100 yds per game. Only two or three WR's in the league consistently get 100 yards. And even Terrell Owens often has a 200 yd / 4 TD game followed by a 60 yd / 0 TD game. Ricky Williams could have the same thing but the wide variation is much less frequent.

Meant to comment on this earlier.Someone (Drinen?) did a study that looked at that. He found the opposite of what you said. RBs are no more consistent than the other skill positions. The reason they seem to be more valuable was that the top ones score more than the lesser ones, compared to QB and WR.

That would be Shick - Standard Deviation (6/29/01) specifically Question 6 right above conclusions

So who’s more consistent? The WR’s have a lower standard deviation, but there’s a catch. The RB’s score more points than the WR’s do. If we talk about the average top 10 RB and top 10 WR, we need to consider how many points they are scoring to see an appropriate SPREAD of these points. Remember that the standard deviation is giving us a range above and below a player’s average. These are values that we expect to see about 68 percent of the time. Check out this graphic.Neither position is way more consistent than the other is. The graphic above displays that the range for RB’s and WR’s are almost the same, but every now and then your top flight WR will score 5 points while your RB at least got you 10 points. It happens. It doesn’t mean that RB’s are more consistent. It means they score more points!

 

[bagger]

was this a redraft league? because whoever drafted portis at 3rd overall and vick 6th overall were crazy. worked out great for them (and i have them both in my deep keeper league) but the risk/reward on those two players at those draft slots is pretty hard to justify.but that's a whole different topic of conversation.

Newly formed dynasty league drafted AFTER the Bucs won the SB.

makes sense. gotcha.

 

[Marvelous]

Was Ward really the best value play at 3.01? Was Tiki Barber not available?

Tiki was long gone. Here were the first 2 rounds of picks:1.01 Ricky Williams, MIA1.02 Ladanian Tomlinson, SD1.03 Clinton Portis, DEN1.04 Priest Holmes, KC 1.05 Deuce McAllister, NO1.06 Michael Vick, ATL1.07 Shaun Alexander, SEA1.08 Marshall Faulk, STL1.09 Travis Henry, BUF1.10 Ahman Green, GB1.11 Jamal Lewis, BAL1.12 Michael Bennett, MIN2.01 Corey Dillon, CIN2.02 Randy Moss, MIN2.03 Tiki Barber, NYG2.04 William Green, CLE2.05 Donovan McNabb, PHI2.06 Terrell Owens, SF2.07 Edgerrin James, IND2.08 Peyton Manning, IND2.09 Marvin Harrison, IND2.10 Fred Taylor, JAX2.11 Curtis Martin, NYJ2.12 Daunte Culpepper, MIN

Looking over this draft, everyone valued rbs very high. Look at the person drafting 12th. He got Michael Bennett and Corey Dillon, the 11th and 12 picked rbs of the draft. Suppose he picked 2 wr instead (my choice would be Owens and Harrison). He can probably put a decent backfield together from the following: Garner, George, Zereoue, Davis, Emmitt, Thomas, Dunn, Stewart, Barlow. There are rbs out there. If he get 3 of these, I am sure one of them will outproduce Bennett and Dillon. I don't know which one, but if I have enough rbs, it will happen.By the way, if these league only starts 2 wr, I probably would not go wr/wr. But if you need 36 starting wr, having 2 of the top 3 is awesome.

 

[Smoo]

This thread has gone off in many interesting directions, so I apologize if my brief backtrack here interrupts any conversational flow.

A lot of the defense put up by Stud RB proponents here seems to include caveats that they would jump on a better player if they "slipped through the cracks" or if the opportunity arose. I just want to reiterate that this is completely consistent and in agreement with my original point. People who are doing that are not using Stud RB as they claim to be, they are using VBD, whether they know it or not. Using VBD doesn't mean you have to have Bryant's spreadsheet in front of you. Any time you take what you perceive as the best value pick available, you are using some form of rudimentary VBD.

My criticism was of Stud RB in its purest form, which is to take RBs and RBs exclusively for the first two or three rounds, regardless of other options. And yes, there are people out there who do that.

Most of the Stud RB proponents I've seen commentary from above are not (I would assert) pure Stud RB proponents and are actually following the principles of the value system.

 

[bagger]

This thread has gone off in many interesting directions, so I apologize if my brief backtrack here interrupts any conversational flow.

A lot of the defense put up by Stud RB proponents here seems to include caveats that they would jump on a better player if they "slipped through the cracks" or if the opportunity arose. I just want to reiterate that this is completely consistent and in agreement with my original point. People who are doing that are not using Stud RB as they claim to be, they are using VBD, whether they know it or not. Using VBD doesn't mean you have to have Bryant's spreadsheet in front of you. Any time you take what you perceive as the best value pick available, you are using some form of rudimentary VBD.

My criticism was of Stud RB in its purest form, which is to take RBs and RBs exclusively for the first two or three rounds, regardless of other options. And yes, there are people out there who do that.

Most of the Stud RB proponents I've seen commentary from above are not (I would assert) pure Stud RB proponents and are actually following the principles of the value system.

i drafted using stud rb in its purest form drafting only RBs the entire draft.you only do that once.

 

[Smoo]

i drafted using stud rb in its purest form drafting only RBs the entire draft.you only do that once.

:rotflmao:

 

[Bri]

I argue that the most predictable position in fantasy football is the running back. That is what makes them so valuable -> not that they score many more fantasy points than a Qb or WR. It is just that there are some guys you can count on getting 1 TD and 100 yds per game. Only two or three WR's in the league consistently get 100 yards. And even Terrell Owens often has a 200 yd / 4 TD game followed by a 60 yd / 0 TD game. Ricky Williams could have the same thing but the wide variation is much less frequent.

I think QB is predictable in a manner similar to RBs. It's just that there always seems to be plenty so there's little worry in landing one

 

[Marvelous]

This thread has gone off in many interesting directions, so I apologize if my brief backtrack here interrupts any conversational flow.

A lot of the defense put up by Stud RB proponents here seems to include caveats that they would jump on a better player if they "slipped through the cracks" or if the opportunity arose. I just want to reiterate that this is completely consistent and in agreement with my original point. People who are doing that are not using Stud RB as they claim to be, they are using VBD, whether they know it or not. Using VBD doesn't mean you have to have Bryant's spreadsheet in front of you. Any time you take what you perceive as the best value pick available, you are using some form of rudimentary VBD.

My criticism was of Stud RB in its purest form, which is to take RBs and RBs exclusively for the first two or three rounds, regardless of other options. And yes, there are people out there who do that.

Most of the Stud RB proponents I've seen commentary from above are not (I would assert) pure Stud RB proponents and are actually following the principles of the value system.

When I first started in fantasy football many years ago, I was a big stud rb drafter. Of course, this was before many people knew about it, so it worked really well and I always got two good rbs no matter where my draft position was.However, over time, others have been placing more emphasis on rb. I would still try to get rbs with my first two picks, but it seemed like I was just getting mediocre teams like everyone else (maybe worse). On a side note, even though draft position is randomly drawn, I never get a high pick.

Anyway, last year I decided I was going to get top wr if I middle/late draft pick. I got 8 out of 10 teams. I ended up with Owens and Harrison (Moss was taken in front of me). It worked out great. I clearly has the best team, although I lost the championship because of Philly D romping Dallas. Owens not playing also hurt.

Anyway, if everyone else is doing stud rb, you will end up with a team like everyone elses. And if you pick near the end of the draft, your team will probably be worse. In every draft, some players slip. My goal is to get those players. In years past, I would hesitate drafting a bargain rb in round 5 since I would already have 3 rbs and only 1 wr.

 

[Bottomfeeder Sports]

....In every draft, some players slip. My goal is to get those players. ...

That is the key to drafting. Picking the players that are still untouched at the top of your cheatsheet surrounded by a sea of highlighted players picked by everyone else. Skipping over names to get your favorite players, or more recognizable name, or a hot sleeper is reaching. Reaching is giving up value!

 

[Pops]

I posted this in the other 3 page topic on Stud RB's but I think in my league at least it proves that there is a big drop off from the first tier or RB's to the second (at least more than QB or WR). So If I can get 2 RB out of this top 12, I am for sure going to take it. Now I did draft last year using VBD not just on the theory of taking 2 RB's. I ended up taking 3 RB's in a row followed by a WR and then my 4th RB. I couldn't pass up Martin, McAlister, Barber & Henry with my first 5 picks in a 12 team league."This is an interesting thread because I have always been the guy to take RB, RB, RB and more RB. Last season I took 3 RB, a WR and a 4th Rb in the first 5 rounds. I lost the superbowl to another owner who had 3 strong RB (he was lucky enough to take Clinton Portis and it paid off).So I started to look at scoring and stats in my league. here is what I came up with.A few of our guys have been complaining that the QB is irrelevant, so I did a quick little analysis of our league.Point scoring is like this.QB:1 pt per 25 yards3 for 300 yard game4 pt per passing TD6 pt per rushing TD-2 per INTRB:1 pt per 10 yards3 for 100 game6 pt per rushing TD9 pt per receiving/passing TD1/2 pt per receptionWR/TE:1 pt per 10 yards3 for 100 game6 pt per receiving TD9 pt per rushing/passing TD1/2 pt per receptionStarting lineup is like this.1QB, 2RB, 2WR, 1TE, 1K, 1D 1 Flex at RB or WR. (We vote every year on the flex and every one loves the options.)So I took this scoring and did an average of each of these positions based on the final 2002 NFL stats. Here is what I found. With the above scoring, the Top 12 QB's averaged more points per game than any one player at another position. The 2nd tier of QB's averaged almost as much as the top RB. The second and third tiers of RB's and WR's were almost identical.QB 2002 Averages Top 12 16.58Top 24 13.69Top 32 11.53RB 2002 AveragesTop 12 14.09Top 24 9.62Top 36 7.48Top 48 4.13WR 2002 AveragesTop 12 11.64Top 24 9.16Top 36 7.76Top 48 6.78Top 60 6.17TE 2002 AveragesTop 12 5.52Top 24 3.43"