Dual Baseline Draft Value (1 Viewer)

[waj707]

This below describes is the method that I use for determining draft value, assuming that you have developed or are using projected fantasy points for every player. I think that it does a better job than straight VBD in calculating value because it takes into account the changes in value caused by league roster sizes. I am assuming no flex and no multiple position elgiblity.

1.) The goal of the draft is to maximize the value of the starters at every position AND maximize the value of the roster as a whole.

Comment: Value of the roster as a whole refers to value of backups/bench players. Scarcity in available players and point production drive some backups to be worth more than starters. This explains why PK are drafted at the end of draft even though they are starters.

2.) The number of candidates for any draft is the number of total draft picks.

3.) Any player that is a candidate to be drafted has a positive value.

Comment: This is one thing I always hated about straight VBD, that non-starter have negative VBD values...this seems wrong. Any player that should get drafted should have positve draft value.

4.) Any player that is not a candidate to be drafted has a value of 0.

Comment: Worthless is worthless...no sense rubbing it in by saying how much so.

5.) If a position is deemed important enough to warrant drafting bench players, the overall value for the players at that position increases in relation to the other positions. The more bench players warranted for a particular position, the addiitonal value that position gains in respect to the others.

6.) The baseline for measuring a player's starter value in relation to others at that position is the first non-starter at that position. If a player's forecasted point total is equal or less than the first non-starter, then that player's starting value is zero.

Comment: The baseline here is staightforward based on league rules.

7.) The baseline for measuring a player's roster value in relation to others at that position is the first non-rostered player for that position. If a player's forecasted point total is equal or less than the first non-rostered player, then that player's roster value is zero.

Comment: The baseline here is now determined by whom you determine is a candidate to draft. This is subjective and is determined by looking at the forecasted point totals and seeing where the points for a particular postion flat-line (i.e. a great deal of players with nearly the same value), or at least become linear. Note that you can have a position where the the number of draft candidates are less than the number of starters (i.e. you can decide that only 4 kickers are worth drafting so early in the pre-season). Note that in the end the number of draft candidates for all the positions has to equal the number of draft picks in the draft.

8.) The combined value for a player is the average of a player's starting value and a player's roster value.

Comment: If you are trying find balance between starter value and roster value, it makes sense to average the two values.

Note that it the FGB VBD excel spreadsheet makes this easy to calculate...you just insert columns to calcuate starter value, roster value and overall value.

Please let me know what you think. Note that I realize that this does not take into multiple position elgiblity/flex, ADP, or any dynamic value changes. I've got some things figured out for these items that I don't want to share yet. Thanks!

 

[GregR_2]

I totally understand where you're going. But I think there are a lot of problems with trying to use a VBD approach to backups. For example:

1) The value of the backup depends on how many games you may start him. Since sometimes you may start a backup based on matchups, this means his value is dependent on the quality of the starter you already have. The better your starter the less likely you would play a backup over him. Ie. if you have AJ, Roddy White, and Greg Jennings, your WR4 has a lot less value to you than he does if you start Steve Johnson, Dez Bryant, and Santana Moss.

2) For RB25, his projected fantasy points probably does reasonably express his value. However for someone like Derrick Ward, his value he should be drafted by is more likely what he will score if Arian Foster is injured, not the "most likely" value you projected for him that is closer to what you expect as he plays as a backup. So some players projected points works, but others you really would need a number that expresses their upside.

3) Continuing with the thought of #2... if a backup is an NFL backup, his value is also tied to if you own the player ahead of him. Knowing that he'll be available on the same weeks your starter won't if injured makes him more valuable to you than he is to some other team.

4) The thing you said in #5 about more backup slots means the position is more valuable isn't always so. This is because of #1. If you have the first group of WRs you might only carry 2 more WRs because you don't expect to need them. If you had the second group, you might carry 4 more WRs hoping one will break out, and hoping to play matchups.

I run into this in my dynasty contract league that has player salaries, because to get at salaries I need a number to reflect the starter's values, but I also need a number that scales properly to come up with the backup's auction value.

I just don't see a way to do that in a program and have it reflect how I feel about the players. Definitely not with just a set of projections that only capture your "most likely" for the player. I think for an auction, you're better off deciding how much money should go to backups total, come up with the starter prices with that taken into account, and then go and assign prices to the backups manually to reflect how you feel about each of those players. A lot of the time, it's breakout potential that dictates the price. Once you've done that you might find you allocated too much to backups or not enough, and then can recalculate the starters with an adjusted amount for backups and do it again so it looks better.

And such a backup auction price is probably about the most accurate reflection of value you could do. So I might suggest you look at it like that instead of trying to use a projection to generate the backup values.

That said, if you're in a small league with a small bench where all the bench players are NFL starters and so their projection really is the value you'd draft by, then your approach might have more merit, though there will still be issues like the backups that are actually a part-time starter in your mind as opposed to truly a backup.

 

[Lab]

The dual baseline (averaged, if I understand you correctly) is fine. I suspect it's a big improvement over a single static baseline. Even better might be three baselines averaged, since that's more likely to be applicable in more situations. But I think where all this thinking should be leading you is toward a shifting baseline that changes depending on which part of the draft you are in. The best baseline for rounds 1-4 is different from the best baseline for rounds 5-10. And the best baseline for rounds 11-end is something else again. This is the core of dynamic VBD (dVBD).

My suggestion: Use your dual baseline system in 2011, and perhaps experiment with a triple baseline. In 2012, try creating a rotating series of three baselines for three different segments of the draft (early, middle, and end rounds). And then in 2013, if you want to go further, use what you learned in 2012 to develop a truly dynamic VBD method that changes with each and every round, depending on what positions other teams are selecting and what you expect them to select in future rounds. I know this sounds like a long learning process, but if you really want to understand how all the pieces fit together and ultimately find the right system for yourself, I think that's the best way to go.

Good luck. What you're doing is really keen, I think.

 

[Ignoratio Elenchi]

I don't mean to burst your bubble, but it doesn't appear you're really doing anything different here.

3.) Any player that is a candidate to be drafted has a positive value.Comment: This is one thing I always hated about straight VBD, that non-starter have negative VBD values...this seems wrong. Any player that should get drafted should have positve draft value.4.) Any player that is not a candidate to be drafted has a value of 0.Comment: Worthless is worthless...no sense rubbing it in by saying how much so.

This seems like a meaningless distinction. You're just shifting the values (e.g. instead of the best QB having a VBD of 200 and the worst QB having a VBD of -200, you give the best QB a VBD of 400 and the worst a 0). For the players that aren't draftable, there's no point changing their VBD to a zero, just leave it whatever it is.

Comment: The baseline here is now determined by whom you determine is a candidate to draft. This is subjective and is determined by looking at the forecasted point totals and seeing where the points for a particular postion flat-line (i.e. a great deal of players with nearly the same value), or at least become linear. Note that you can have a position where the the number of draft candidates are less than the number of starters (i.e. you can decide that only 4 kickers are worth drafting so early in the pre-season). Note that in the end the number of draft candidates for all the positions has to equal the number of draft picks in the draft.

This is where the whole thing gets circular - you're trying to factor in a player's "roster value" but this is just being determined by the values you assign at the outset, the same way you would in normal VBD.

8.) The combined value for a player is the average of a player's starting value and a player's roster value.Comment: If you are trying find balance between starter value and roster value, it makes sense to average the two values.

It doesn't always make sense to use an unweighted arithmetic mean to combine two values. I'm not saying that's not the right way to do it in this case, since I'm not entirely sure exactly what you want this final number to represent.Ultimately, it seems to me that all you're doing is shifting the normal VBD of starters up by a constant (specifically, the difference between the worst starter and the worst bench player). You're shifting bench players up by a progressively larger factor, due to the fact that they're not being penalized for not being a starter, like they normally would by by their negative VBD. So basically you're just replicating a normal VBD, but compressing the bench player values upward somewhat.If you disagree, I'd be interested to see a list of players with their VBD (calculated by some "standard" method) and the value you calculate with your system. Then we could see if your system actually exhibits any meaningful differences, or if you end up with essentially the sameresults, just with some different "labels" for the values.

 

[Ignoratio Elenchi]

On further thought, I guess I see how you're trying to quantify whether or not a player is worth drafting. But because you're deciding at the outset how many players at each position are draftable, it doesn't seem like you're adding much info with all the "roster value" calculations. As it gets later in the draft, it doesn't matter to me whether QB36 has a value of 0 or -237; I either want/need a third QB or I don't. At least with normal VBD, those relative values are preserved so that if the draft doesn't go exactly the way you expect, you can still make some projection-based decisions on the fly. With your system, they've all been zeroed out.

For example, you might decide at the beginning that there are only 15 draftable TEs, so TE16 and lower all get zeroes (i.e. they're all "equally worthless" in your system). If for some reason you didn't anticipate, there's a run on backup TEs at some point in the draft, you might get stuck having to draft a player that you previously decided was undraftable.

 

[GregR_2]

On further thought, I guess I see how you're trying to quantify whether or not a player is worth drafting. But because you're deciding at the outset how many players at each position are draftable, it doesn't seem like you're adding much info with all the "roster value" calculations. As it gets later in the draft, it doesn't matter to me whether QB36 has a value of 0 or -237; I either want/need a third QB or I don't. At least with normal VBD, those relative values are preserved so that if the draft doesn't go exactly the way you expect, you can still make some projection-based decisions on the fly. With your system, they've all been zeroed out.For example, you might decide at the beginning that there are only 15 draftable TEs, so TE16 and lower all get zeroes (i.e. they're all "equally worthless" in your system). If for some reason you didn't anticipate, there's a run on backup TEs at some point in the draft, you might get stuck having to draft a player that you previously decided was undraftable.

I don't think it's a matter of coming up with a number to tell you whether to draft a 3rd QB. I think what he wants to do is come up with a number that helps him decide whether he's better to draft his QB2 this round, his RB4 next round, and his WR5 the round after. Or does he draft the WR5 this round, the QB2 next, and the RB4 last, etc.Since the QB2 you can get this round is different than the QB2 that you can get 2 rounds later, part of the answer to that question is how much does the position drop off between now and then, and how does that compare to the other positions. If you have a number that quantifies what you believe their value is, you can do simple subtraction to find out what the best combination of players is.The problems are like I already mentioned... if you have Aaron Rodgers as your QB, your backup QB has a lot less value to you than if you have Josh Freeman as your starting QB. The latter you might start your backup based on matchups. With Rodgers you probably only start the backup on a bye or in case of injury. So there isn't a baseline that gives you a good value that you can apply before the draft. You would need something that takes your team you've already drafted (and still plan to draft) into account.

 

[Ignoratio Elenchi]

On further thought, I guess I see how you're trying to quantify whether or not a player is worth drafting. But because you're deciding at the outset how many players at each position are draftable, it doesn't seem like you're adding much info with all the "roster value" calculations. As it gets later in the draft, it doesn't matter to me whether QB36 has a value of 0 or -237; I either want/need a third QB or I don't. At least with normal VBD, those relative values are preserved so that if the draft doesn't go exactly the way you expect, you can still make some projection-based decisions on the fly. With your system, they've all been zeroed out.For example, you might decide at the beginning that there are only 15 draftable TEs, so TE16 and lower all get zeroes (i.e. they're all "equally worthless" in your system). If for some reason you didn't anticipate, there's a run on backup TEs at some point in the draft, you might get stuck having to draft a player that you previously decided was undraftable.

I don't think it's a matter of coming up with a number to tell you whether to draft a 3rd QB. I think what he wants to do is come up with a number that helps him decide whether he's better to draft his QB2 this round, his RB4 next round, and his WR5 the round after. Or does he draft the WR5 this round, the QB2 next, and the RB4 last, etc.Since the QB2 you can get this round is different than the QB2 that you can get 2 rounds later, part of the answer to that question is how much does the position drop off between now and then, and how does that compare to the other positions. If you have a number that quantifies what you believe their value is, you can do simple subtraction to find out what the best combination of players is.

I don't see how his system is an improvement over any other VBD-type system in this regard. In fact, his system has the built-in flaw that every player after a certain predetermined point has a value of zero, as opposed to their natural relative values.

The problems are like I already mentioned... if you have Aaron Rodgers as your QB, your backup QB has a lot less value to you than if you have Josh Freeman as your starting QB. The latter you might start your backup based on matchups. With Rodgers you probably only start the backup on a bye or in case of injury. So there isn't a baseline that gives you a good value that you can apply before the draft. You would need something that takes your team you've already drafted (and still plan to draft) into account.

I agree with this, which is why trying to quantify "roster value" ahead of time is mostly a pointless exercise. No one drafts strictly by the numbers anyway (e.g. if you've just taken PK6, you're not going to take PK7 off the board next round, even if he has the highest "value" of all the remaining available players). Of course, that's a problem with any static VBD calculation.A truly dynamic VBD system would get much closer, but would still fall short for the reasons you pointed out. The value you assign to a specific player later in the draft can be largely influenced by who you've already drafted, and it would take an enormously complicated set of algorithms to automate that process. The human brain is much better suited to perform that kind of analysis on the fly than any value-calculating system would be.

 

[GregR_2]

...

I don't see how his system is an improvement over any other VBD-type system in this regard. In fact, his system has the built-in flaw that every player after a certain predetermined point has a value of zero, as opposed to their natural relative values.

...

Well, I think the gist of what he is trying to do actually is an improvement over a single baseline VBD.... the backups need a completely different baseline than is used for the starters to get anywhere close to accurate. Or at least, they need a formula that accounts for the fact that not all of their 'value' hits your starting lineup. And you need both to work together to where the best backup is still less valuable than the worst starter, and the gap between them is on par with the difference in value between them.

But that said, while yes it's an improvement, I don't think it's enough of an improvement to be worth doing for anything other than the first couple of backups... whose projected "most likely" fantasy points is the basis for the value of drafting him. A lot of players that I draft it's because of what they might do if the guy ahead of them fails or is injured, not what I expect is most likely they'll do.

 

[Lab]

A truly dynamic VBD system would get much closer, but would still fall short for the reasons you pointed out. The value you assign to a specific player later in the draft can be largely influenced by who you've already drafted, and it would take an enormously complicated set of algorithms to automate that process. The human brain is much better suited to perform that kind of analysis on the fly than any value-calculating system would be.

I agree with your point about the human brain being more adept than calculations, and better able to adjust to intangibles. But I think part of the value of using a VBD (or dVBD) spreadsheet is that it forces you down an objective data-driven path, and better helps you resist natural urges to draft players you "like better." Maybe others are better at resisting the urge to reach for a stud, or resisting the urge to pick a young dynamic rookie over a boring old dude. But I really think the spreadsheet helps keep me focused. It's sort of like following a playbook rather than just improvising -- yes, you need to be able to improvise, but only if the set play breaks down or a special opportunity presents itself.

 

[GregR_2]

A truly dynamic VBD system would get much closer, but would still fall short for the reasons you pointed out. The value you assign to a specific player later in the draft can be largely influenced by who you've already drafted, and it would take an enormously complicated set of algorithms to automate that process. The human brain is much better suited to perform that kind of analysis on the fly than any value-calculating system would be.

I agree with your point about the human brain being more adept than calculations, and better able to adjust to intangibles. But I think part of the value of using a VBD (or dVBD) spreadsheet is that it forces you down an objective data-driven path, and better helps you resist natural urges to draft players you "like better." Maybe others are better at resisting the urge to reach for a stud, or resisting the urge to pick a young dynamic rookie over a boring old dude. But I really think the spreadsheet helps keep me focused. It's sort of like following a playbook rather than just improvising -- yes, you need to be able to improvise, but only if the set play breaks down or a special opportunity presents itself.

I agree with you where that is indeed one of the biggest values of VBD. I'll expound on it and say part of that value is that I can focus all of my attention on a player individually when I make his projection, and capture all of my thoughts on him in that number. And then a week later when I have 90 seconds to make a decision, and my brain won't remember all of those details, I have the benefit of using what I thought about the player when I had the time to think it out.But that's the problem with doing this for backups. A most likely number of fantasy points does a very good job of capturing your beliefs about a starter. It's a lot more difficult to come up with a number to use for the value of backups because it does depend on so many other factors.

You can do it, but I'm not convinced that it should come from projections for a lot of the backups. Unless you play in small leagues and/or with small rosters where everyone rostered is an NFL starter.

 

[Ignoratio Elenchi]

I agree with your point about the human brain being more adept than calculations, and better able to adjust to intangibles. But I think part of the value of using a VBD (or dVBD) spreadsheet is that it forces you down an objective data-driven path, and better helps you resist natural urges to draft players you "like better." Maybe others are better at resisting the urge to reach for a stud, or resisting the urge to pick a young dynamic rookie over a boring old dude. But I really think the spreadsheet helps keep me focused. It's sort of like following a playbook rather than just improvising -- yes, you need to be able to improvise, but only if the set play breaks down or a special opportunity presents itself.

I agree with this in general - it's good to have some projections in front of you to keep you focused. My point is that trying to come up with a system that properly values backups is almost impossible, because the value of, say, Peyton Manning's backup is different for someone who already drafted Peyton Manning than it is for someone who didn't. A static VBD doesn't account for this at all, nor does any dynamic VBD I've ever seen. Imagine writing a computer program that would try to quanitfy that (and every other possible consideration like it) - it would be impossibly complex and totally inefficient compared to the human brain, which can instantaneously reevaluate whether or not it's worth taking Peyton Manning's backup.

 

[Lab]

My point is that trying to come up with a system that properly values backups is almost impossible, because the value of, say, Peyton Manning's backup is different for someone who already drafted Peyton Manning than it is for someone who didn't. A static VBD doesn't account for this at all, nor does any dynamic VBD I've ever seen.

I completely agree with this point. It's made even more complex by the individual specifics of your league. For example, if I've already got Manning as my QB, and I see Brady still sitting there undrafted in round 7, my VBD sheet will certainly tell me to pick Brady even though I don't really need another starting QB. Whether picking Brady is a good choice or not may depend on whether I think I can profitably trade one of those QBs after the draft.But while we all agree that unwavering devotion to a spreadsheet is unwise, it seems like we also all agree that some form of pre-draft calculation and ranking is useful. To get back to the original question, I still think waj707's dual baseline approach has merit, so long as he treats it as a starting point rather than an ending point. I still think though that he can (and should) take it farther to improve the output.

 

[Ignoratio Elenchi]

Note that it the FGB VBD excel spreadsheet makes this easy to calculate...you just insert columns to calcuate starter value, roster value and overall value.

I just did this since I was curious to see the results, and I should take back my earlier comments. It does seem like this produces results more in line with the way you'd want to draft. I set up the VBD sheet with the settings for my small local league, just to keep it simple. 10 teams, 16 rounds of the draft, start 1 QB, 2 RB, 2 WR, 1 TE, 1 PK, 1 TD. I went with the following assumptions for the number of "draftable" players at each position, loosely based on my experience but also rounded to keep the work easy: 20 QBs, 50 RBs, 50 WRs, 20 TEs, 10 PKs, 10 TDs. Here are the results, grouped by position:

Player	VBD Rank DBD RankQB-1	4	4QB-2	6	7QB-3	14	16QB-4	20	20QB-5	23	23QB-6	29	25QB-7	32	28QB-8	46	48QB-9	61	55QB-10	63	57QB-11	64	58QB-12	88	69QB-13	100	73QB-14	105	79QB-15	108	82QB-16	117	85QB-17	125	96QB-18	138	108QB-19	145	123QB-20	152	153RB-1	1	1RB-2	2	2RB-3	3	3RB-4	5	5RB-5	7	6RB-6	9	8RB-7	10	9RB-8	11	10RB-9	13	12RB-10	15	13RB-11	16	14RB-12	17	15RB-13	18	17RB-14	22	19RB-15	24	21RB-16	36	31RB-17	37	32RB-18	40	35RB-19	43	38RB-20	44	39RB-21	45	40RB-22	48	41RB-23	50	42RB-24	58	44RB-25	68	46RB-26	71	49RB-27	77	53RB-28	94	59RB-29	96	60RB-30	102	64RB-31	109	67RB-32	112	68RB-33	126	72RB-34	131	81RB-35	140	86RB-36	141	90RB-37	146	99RB-38	147	103RB-39	148	106RB-40	149	111RB-41	150	112RB-42	151	121RB-43	153	125RB-44	154	129RB-45	155	131RB-46	156	133RB-47	157	136RB-48	158	138RB-49	159	156RB-50	160	157WR-1	8	11WR-2	12	18WR-3	19	22WR-4	21	24WR-5	25	26WR-6	26	29WR-7	27	30WR-8	28	33WR-9	30	34WR-10	31	36WR-11	34	37WR-12	35	43WR-13	38	45WR-14	39	50WR-15	51	62WR-16	52	63WR-17	59	65WR-18	62	66WR-19	66	70WR-20	69	71WR-21	74	76WR-22	76	77WR-23	78	78WR-24	79	80WR-25	84	83WR-26	89	93WR-27	90	95WR-28	92	97WR-29	101	104WR-30	106	110WR-31	110	115WR-32	113	117WR-33	114	118WR-34	116	120WR-35	118	124WR-36	119	126WR-37	120	127WR-38	121	128WR-39	124	132WR-40	127	137WR-41	128	140WR-42	129	142WR-43	130	143WR-44	132	145WR-45	135	147WR-46	136	148WR-47	139	151WR-48	142	154WR-49	143	155WR-50	144	158TE-1	33	27TE-2	49	47TE-3	53	51TE-4	54	52TE-5	57	54TE-6	93	84TE-7	95	88TE-8	97	91TE-9	98	92TE-10	99	100TE-11	103	107TE-12	104	114TE-13	107	116TE-14	111	119TE-15	115	122TE-16	122	134TE-17	123	135TE-18	133	149TE-19	134	150TE-20	137	152PK-1	47	75PK-2	55	89PK-3	56	94PK-4	65	102PK-5	70	105PK-6	75	130PK-7	81	141PK-8	83	144PK-9	85	146PK-10	86	160TD-1	41	56TD-2	42	61TD-3	60	74TD-4	67	87TD-5	72	98TD-6	73	101TD-7	80	109TD-8	82	113TD-9	87	139TD-10	91	159

I omitted the player names in order to protect FBG's paid content, but the "Player" column shows them ranked by projected points. "VBD Rank" is the rank of the player based on the "value" calculated by the VBD spreadsheet, and "DBD Rank" is the rank of the player based on the value generated by OP's system. For example, the VBD spreadsheet would indicate that the top Team Defense (TD-1) is the 41st most valuable player, while OP's system indicates it's the 56th most valuable player. In general, the biggest downgrades went to kickers and defenses, and the biggest upgrades went to QBs and RBs, which seems to make sense. In particular, note that the worst draftable kicker (PK-10) is ranked dead last in OP's system, but 86th in the VBD list. It's far from a perfect system, but I think I'm now convinced that it is an improvement over straight VBD.

 

[GregR_2]

Depending on which baseline you use, a lot of the FBG tools already build the extra value of positions into the baseline. Like Joe's Secret Formula, the default baseline, sets the baseline for RB deeper than it does for the other positions. So be careful you're not building the same effect in twice.

 

[waj707]

Thanks for all the replies! I appreciate the feedback! Couple more comments:

1.) I should have been more clear about that I was shooting for a method to calculate value just based on projections and league rules (i.e. scoring criteria, starter rules, roster size, etc.). This would be more a "static" value as opposed a "dynamic" value that is calculated as a draft proceeds, which would be based on whom you already drafted, what players are available, handcuffing considerations, and ADP data, among other things.

2.) That being said, I do realize that once roster sizes get larger, one would be more compelled to draft higher variance players - those players that have a better chance of hitting it big as opposed to being steady. I think of this as the Kevin Faulk principle: that if you had two players with the same projections at the end of draft, you would rather take a chance on a rookie that has a minute chance of being starter over somebody like Kevin Faulk, who is very steady and has a well-defined role, but has a low chance of busting out. I have tried to figure out a way to factor this in besides factoring the potential breakout in the projections, but haven't found anything I like. Anybody have any suggestions?

 

[Ignoratio Elenchi]

Thanks for all the replies! I appreciate the feedback! Couple more comments:1.) I should have been more clear about that I was shooting for a method to calculate value just based on projections and league rules (i.e. scoring criteria, starter rules, roster size, etc.). This would be more a "static" value as opposed a "dynamic" value that is calculated as a draft proceeds, which would be based on whom you already drafted, what players are available, handcuffing considerations, and ADP data, among other things. 2.) That being said, I do realize that once roster sizes get larger, one would be more compelled to draft higher variance players - those players that have a better chance of hitting it big as opposed to being steady. I think of this as the Kevin Faulk principle: that if you had two players with the same projections at the end of draft, you would rather take a chance on a rookie that has a minute chance of being starter over somebody like Kevin Faulk, who is very steady and has a well-defined role, but has a low chance of busting out. I have tried to figure out a way to factor this in besides factoring the potential breakout in the projections, but haven't found anything I like. Anybody have any suggestions?

You could try to derive some kind of inverse factor of years played in the league, variance in weekly scoring, etc. But that's almost certainly more work than it's worth. That's what I was alluding to earlier, any kind of cheatsheet (however you generate it) is useful as a reference, but at a certain point the human brain is much more efficient than any value model would be. You could drive yourself nuts perfecting a formula that adequately captures something like the Kevin Faulk principle, for example, and you still wouldn't really have any advantage over the owner sitting next to you who can just look at Kevin Faulk and some other rookie RB and decide on the fly which one has more upside potential.

 

[moleculo]

Is there a good working definition for dVBD somewhere? I've got a good working concept for continually adjusting baselines that is pretty slick, but I'd like to se what else has been done.

 

[Ignoratio Elenchi]

Is there a good working definition for dVBD somewhere? I've got a good working concept for continually adjusting baselines that is pretty slick, but I'd like to se what else has been done.

I think the gist of dVBD is that your baseline at any moment in time is the value of the player you'd expect to be available at your next pick (determined by ADP, needs of teams picking before you, etc.) I believe Draft Dominator does this, and then also factors in your team needs (e.g. QB will be valued lower if you already have a starting QB).

 

[waj707]

bump in responder to market driven baseline post.