Let's Talk Value Based Drafting (3 Viewers)

[abbottjamesr]

Joe Edit:  Let's use this thread as a place to discuss Value Based Drafting. 

To make sure we're all on the same page, this is Value Based Drafting Please make sure you're familiar with the concepts there first and then you guys discuss here. Thanks.

J
 

******************

In the "Looking for Links" thread @Ilov80s brought up a podcast that was critical of VBD and its merits for determining draft order in FF.  There was some good discussion going on and it should have its own thread.

When VBD was first published by FBGs several years ago it was pretty revolutionary.  Instinctively we all knew that we should be drafting kickers and defenses late but we didn't have a good model that explained why exactly.  VBD did a good job of showing relative value of one player versus other players of their position and then ranking all of the players together would give you a RB's value compared to a QB. 

There are however some issues with VBD that others have mentioned.  First is, it requires good projections.  Projections are highly subjective and prone to error.  Second it requires you pick a good baseline for each position.  There are several theories on what your baseline should be and is also highly subjective.  Third, VBD does not take into account opportunity cost vs ADP.  Fourth, VBD looks at either points per game or total points for a season and does not account for how points are scored on a weekly basis but instead just the averages.  

My main issues with VBD are

 1. It assumes that the value of a player changes linearly with the decrease in PPG or total points.  The delta in PPG of 18 to 12 is far more significant that 12 to 6 in actual weekly production and impact on your team.  This is mostly due to how variance works.  A player with a higher PPG is more likely to put up a huge week wining score in any given week than a lower PPG player.  Since most FF leagues are played as a series of weekly games the value of the week winning scores is extremely high and is under valued in standard VBD models.

2.  VBD breaks down outside the top 60 or so players.  This is mainly due to the nature of projections being difficult for many players but also in that the difference between the players between 50 and 100 is usually very small.  Historically only 2 points per game separate WR30 from WR60.  The goal of picks after round 5 or 6 should be upside in most league formats.  An average projection is not going to help you determine upside and just lumps all these players together.  It also will over valuate QB's and TE's in this range.  QB's are easier to predict historically and do not break out enough to justify the opportunity cost in those rounds.  

I can copy of some of the previous discussion if its help full.

James

 

[Joe Bryant]

Hi James,

I edited the title and added a note at the top to give folks the link to my article so we can all be on the same page. Thanks.

J

 

[abbottjamesr]

Hi James,

I edited the title and added a note at the top to give folks the link to my article so we can all be on the same page. Thanks.

J

Looks good,  Thanks

 

[Joe Bryant]

I want this mainly to be you guys discussing but one point I sometimes see is people thinking one should stick strictly to the lists for the entire draft. That's not the case at all. In fact, point #7 from the article is:

7. KNOW WHEN TO DEVIATE FROM VBD PRINCIPLES

One of the biggest misconceptions with value-based principles is that you should use them for the entire draft. This simply isn't true. Sure you would like to get value throughout your draft, but after approximately half of your roster is filled, you will generally end up with your best team by using your position lists to address your unique needs. Always look at the best available player by position, but also make sure that person fills bye week needs, etc for your roster.

This includes (but is not limited to) the following not based on X Value:

Covering bye weeks

Handcuffing the backup to a key player on your roster

Looking for favorable match ups for certain tough weeks

For these reasons, we believe the perfect draft will often abandon strict VBD drafting after all baselines have been passed (this is generally around Pick 120 or so). What exactly does this mean? It just means that later in the draft, you'll likely draft from your positional lists and not from your overall list.

Many people like to get off the VBD list and switch to filling positions earlier. That's cool too. 

J

 

[Ilov80s]

Since Joe is here and he's the godfather of VBD, everyone in the podcasts I have heard has referred to Joe as a leader and all time key member of the fantasy football world. However, just like the west coast offense or various offensive approaches in the NBA have evolved, some analysts believe the ideas beyond VBD have developed. 

 

[Joe Bryant]

Since Joe is here and he's the godfather of VBD, everyone in the podcasts I have heard has referred to Joe as a leader and all time key member of the fantasy football world. However, just like the west coast offense or various offensive approaches in the NBA have evolved, some analysts believe the ideas beyond VBD have developed. 

For sure. The biggest change I think has been the ability to add "dynamic" valuation. With today's tech, the ability to refactor value after every pick is incredible compared to a static list. It's why things like our Desktop Draft Dominator and Mobile Draft Dominator apps are so powerful. In essence, it's like refactoring a keeper league after every single pick. But the principles are the same. It's just being able to use an updated available draft pool while knowing what teams have satisfied and relieved their demand for players. In other words, it's a constant pulse on Supply and Demand. A person using a static VBD cheatsheet going against a person using the Draft Dominator is at a huge disadvantage.

J

 

[Borden]

Since Joe is here and he's the godfather of VBD, everyone in the podcasts I have heard has referred to Joe as a leader and all time key member of the fantasy football world. However, just like the west coast offense or various offensive approaches in the NBA have evolved, some analysts believe the ideas beyond VBD have developed. 

I'm just happy that Joe caught your post instead of that half illiterate Borden guy trying to talk about VBD. 

 

[Ilov80s]

I'm just happy that Joe caught your post instead of that half illiterate Borden guy trying to talk about VBD. 

I use VBD more as a tool than a strategy. It gives me a feel for where runs might be, where I can wait a round on a position or player I like, etc. I don't really like it as a draft strategy though. My strategies tend to be more positional based and target specific players that I think can represent difference making value later. For example, this year, I think Decker and Maclin provide potential similar to many of the WRs going in the 3rd-4th rounds so it provides a possible opportunity to leave a hole at WR2 or 3 while stocking up at RB. 

 

[Spike]

In using VBD for years now to aid in draft, I saw something very different this year.  In a league I'm in that starts 2QB/2RB/2WR/2TE/1flex, the quantity of QBs in the top 60 were massively higher this year than in last.

I checked the various positional sheets on the spreadsheet and they're not markedly different than last year.

So why are the QBs SO much higher than before?

 

[Ilov80s]

In using VBD for years now to aid in draft, I saw something very different this year.  In a league I'm in that starts 2QB/2RB/2WR/2TE/1flex, the quantity of QBs in the top 60 were massively higher this year than in last.

I checked the various positional sheets on the spreadsheet and they're not markedly different than last year.

So why are the QBs SO much higher than before?

Interesting. My initial thoughts this year are that QB is as valueless than ever.

 

[Joe Bryant]

In using VBD for years now to aid in draft, I saw something very different this year.  In a league I'm in that starts 2QB/2RB/2WR/2TE/1flex, the quantity of QBs in the top 60 were massively higher this year than in last.

I checked the various positional sheets on the spreadsheet and they're not markedly different than last year.

So why are the QBs SO much higher than before?

Not sure what that would be Spike. I know for sure though that leagues where 2 QBs start have massively higher value for QBs than a 1 QB start league. Because the supply is the same but the demand is twice as much. 

J

 

[Spike]

Not sure what that would be Spike. I know for sure though that leagues where 2 QBs start have massively higher value for QBs than a 1 QB start league. Because the supply is the same but the demand is twice as much. 

J

I get that.  I'm talking about having had the same values on the input page for lineup requirements and scoring rules but last year there were 6 QBs in the top 60 players (in the "Cheat" page's "Overall - Use For Top 50-60 ONLY" column of the spreadsheet) and this year it's 22 of them.

Like I said, exact same lineup requirements and scorning rules.  So what's the difference in the formula?  (In my first iteration of the spreadsheet, I made no tweaks to the projections for the players.)

 

[Arodin]

I'd expect top QBs to be climbing the ranks in dynasty startups...there are a lot of teams where the position isn't settled at all, and many of the go-to, start-and-forget QBs of recent memory are getting old enough to start raising dynasty red flags.  

But this may be the deepest redraft QB crop I've seen in a while.  In smaller bench leagues, I feel like I could grab two guys off the waiver wire after the draft and be comfortable starting them.  

 

[MattFancy]

I get that.  I'm talking about having had the same values on the input page for lineup requirements and scoring rules but last year there were 6 QBs in the top 60 players (in the "Cheat" page's "Overall - Use For Top 50-60 ONLY" column of the spreadsheet) and this year it's 22 of them.

Like I said, exact same lineup requirements and scorning rules.  So what's the difference in the formula?  (In my first iteration of the spreadsheet, I made no tweaks to the projections for the players.)

Most of my leagues are 6pt passing TD leagues. In year's past, there were maybe 3 or 4 QBs listed on the Top 60 list. This year, it's 7. I'm guessing something has been tweaked with the formula to make QBs more valuable.

 

[MattFancy]

I'd expect top QBs to be climbing the ranks in dynasty startups...there are a lot of teams where the position isn't settled at all, and many of the go-to, start-and-forget QBs of recent memory are getting old enough to start raising dynasty red flags.  

But this may be the deepest redraft QB crop I've seen in a while.  In smaller bench leagues, I feel like I could grab two guys off the waiver wire after the draft and be comfortable starting them.  

Agreed. In re-draft this year, there is no point in even thinking about taking a QB until the middle rounds. There's about 15 QBs that I would definitely feel comfortable with as my every week starter. Which is why it seems so strange that there are more QBs in the Top 60 this year than in previous years.

 

[ScottNorwood]

VBD is a sham.  @SeniorVBDStudent has been enrolled for 14 years and still hasn't graduated!

 

[32 Counter Pass]

Can someone post a link to the VBD contrarian podcast mentions above?

 

[abbottjamesr]

Can someone post a link to the VBD contrarian podcast mentions above?

I don't have a link to the podcast mentioned, but here is JJ Zachariason's take on VBD.  This article is a few years old but he is still using similar examples on his numberfire podcasts today.

Value Based Drafting: Version 2

Archived Content Value Based Drafting: Version 2

The goal in drafting a successful fantasy football team is to select a lineup that creates the most value. I talk about this concept a lot throughout The Late Round Quarterback. It’s easily my favorite word to describe logical, sound fantasy football selections.

Joe Bryant created a system a while back that is now referenced by fantasy analysts all over the Net. He coined it “Value Based Drafting”, or VBD for short.

With VBD, Bryant says:

“The value of a player is determined not by the number of points he scores, but by how much he outscores his peers at his particular position.”

And it makes sense. I mean, this is what value is: relative worth. We don’t see Aaron Rodgers as a valuable quarterback in fantasy football simply because he scores the most points. We see him as valuable to the game because he is much better than the rest of the players at his position.

We should be concerned about how a particular player is performing compared to the rest of his position, not how that particular player is performing compared to all positions. That, in a nutshell, is what VBD speaks to.

But this, ladies and gentlemen, is where people stop their analyses. You’ll notice that the growing trend in fantasy football is to draft a quarterback early, because, based on VBD, quarterbacks are now the most valuable position. The difference in points between Aaron Rodgers to the worst starting quarterback in your league last year was larger than the difference between Ray Rice and the worst starting running back in your league.  Running backs, according to many experts, are becoming more and more obsolete because of this.

It all seems very reasonable. If you have the biggest difference maker at a position, then you should have the most valuable team. And if you have the most valuable team, then you should have the best shot at winning, right?

Well, the second part of that is true. The first, regrettably, is not. Even if the point total of Aaron Rodgers to his position is greater than Ray Rice to his, it doesn’t mean Rodgers is more valuable.

And it’s the reason VBD is flawed.

Let’s introduce an addition to the old idea. I’ll call it VBDv2.

If we continue with Joe Bryant’s philosophy, we would conclude that the top ranked player at the position with the largest point differential from best starter to worst starter is the most valuable player in the draft. Woah, that’s a lot to take in. Another way to put this is:

The MVP of fantasy football is the player who is far and away the best at his position. He scores more points compared to the rest of the guys at his position than anyone else.

In today’s world, let’s pretend that this guy is Aaron Rodgers. Remember, this is what experts are telling us all to do – to draft an elite quarterback early. Ideally, this quarterback would be Aaron Rodgers.

Aaron Rodgers is being drafted, on average, in the early to mid first round of most drafts using standard scoring. And moreover, the 12th ranked quarterback in a 12-team league is being drafted in Round 8. This 12th ranked quarterback is the hypothetical worst starting fantasy quarterback in the league.

If the difference between Aaron Rodgers and the 12th quarterback drafted (or the worst starting quarterback in your league) is X, then you would expect to lose X/8 points at each quarterback rank, or pick, in a draft. Another way to put this: X points are being spread across 8 rounds of your fantasy draft at the quarterback position.

Looking at running backs, we can conclude that Arian Foster, the top ranked running back in 2012, is Y amount of projected points better than the 24th ranked back. Remember, if you’re in a 12-team league, your league is more than likely starting at least 24 running backs in a given week. When that 24th ranked running back is being drafted in round 5, then we would expect to lose Y/5 points at each running back rank, or pick, in a draft. Again, another way of wording this is to say that Y points are being extended through 5 rounds of a fantasy draft.

We now have two formulas: X/8 and Y/5. These formulas represent the number of points lost at each round of a fantasy draft until all starters at a particular position are drafted.

Let’s plug in some numbers. Currently, according to ESPN, Aaron Rodgers is expected to score 146 (381-235) points more than the 12th ranked quarterback, Ben Roethlisberger. If Big Ben goes in round 8 of a 12-team draft, which is his expected position, then you’d be losing 18.25 points each round at quarterback.

At running back, Arian Foster is projected to score 296 points. That’s 152 points more than Jonathan Stewart (296-144). With the round and formula in place, we’d expect to lose 31.2 points each round at running back until all 24 backs are chosen.

The common argument in response to this assumption is “how do you know the 1st ranked running back will finish as the 1st ranked running back, and the 24th ranked running back as the 24th ranked running back? Aren’t we more willing to assume Aaron Rodgers will finish at his expected rank compared to someone like Arian Foster?”

You would think so, but history tells us otherwise. Yes, Aaron Rodgers has consistently and historically been a top quarterback, and we’d expect him to do the same this year. But since 2006, the overall probability that a quarterback in the top-12, -6, and -3 pre-season rankings finishes in the same rank category at season’s end is no different than a running back in the top-24, -12, and -6 pre-season rankings finishing with the same or better expected rank. In other words, the running back and quarterback positions are equally inconsistent. If you don’t believe me, then buy my book. I’ll prove it to you.

The fact is, people aren’t analyzing rankings from 2008 to make their decisions in a 2012 draft. We’re looking at rankings for 2012, and making draft choices based on them.

We all need to analyze players using VBDv2 on a case-by-case basis. If you have the 11th pick in a 12-team draft, and not a single quarterback has been taken, then yes, you should draft Rodgers. And inversely, if everyone is drafting quarterbacks during the front end of the first round, you should most definitely snag a guy like Ray Rice or Arian Foster with your pick.

The fact, however, is that the majority of your fantasy leagues are going to be basing their draft decisions on rankings from websites, magazines and books. And when each of these are fairly similar, the rounds these players are taken don’t fluctuate all that much.

In my analysis with all of these posts, and even in my book, I’m not concluding that Aaron Rodgers is a bad draft choice at any point in the draft. Heck, I think he will become the best quarterback in fantasy football history. I’m saying that Aaron Rodgers, being drafted in the mid to early first round, isn’t worth the pick. He only becomes worthwhile towards the end of the first round of your draft.

Written by JJ Zachariason

 

[Fred Lane]

Like many others, the VBD strategy is like the canvas onto which one develops his team.  Many other factors come into play when drafting the "best value" player, such as these already mentioned by the FBG article:  Covering bye weeks, handcuffing the backup to a key player on your roster, looking for favorable match ups for certain tough weeks.

In this regard, I would like to add a few other factors that come into consideration before I draft:

1)  Overall drafting strategy of the other owners:  Sometimes leagues are RB heavy, sometimes the lack of top end WRs gives them more prominence.  It is one thing to get a feel on how most seasons are starting by looking at mock drafts; however, each draft is also individualistic and will take on its own divergent path.  Look for these trends and react.  Sometimes you need to take a player that is not as highly valued, but in a position of greater scarcity.

2)  Keep an eye on other teams selections and needs.  Especially important when you are near the top or bottom of the draft order.  This is another instance where you may pick a lower based player before higher one, knowing that gives you the best chance for selecting both.

3)  Try to secure your safe points in volume positions (QB/RB) and go for breakout players in your other positions (WR/TE), generally; however, that is not always feasible.  For the positions of risk, try to compensate with volume (extra players).

 

[Ilov80s]

Can someone post a link to the VBD contrarian podcast mentions above?

Numberfire football podcast, The Late Round QB. The episode was a couple weeks ago, episode is called "the flaws of vorp".

 

[Slapdash]

Not sure what that would be Spike. I know for sure though that leagues where 2 QBs start have massively higher value for QBs than a 1 QB start league. Because the supply is the same but the demand is twice as much. 

J

I get that.  I'm talking about having had the same values on the input page for lineup requirements and scoring rules but last year there were 6 QBs in the top 60 players (in the "Cheat" page's "Overall - Use For Top 50-60 ONLY" column of the spreadsheet) and this year it's 22 of them.

Like I said, exact same lineup requirements and scorning rules.  So what's the difference in the formula?  (In my first iteration of the spreadsheet, I made no tweaks to the projections for the players.)

I noticed similar changes in VBD for a 1QB auction league I am in.  Luckily, I ignored the QB values, but I have found VBD less helpful in auctions anyways. 

 

[shadrap]

I get the gist of it, but way too much for me.    I use ADP.  If I want a player I have ranked high I just get him one round early.   It's mindboggling how a draft can follow ADP almost to the letter in the 1st four rounds.   That's where I see an advantage.  My style anyway.  That's why this game is so cool.

 

[cloppbeast]

Value based drafting? Never heard of it.

 

[Joe Bryant]

JJ is a super smart and great guy. And I'm not sure if he still agrees with what he wrote several years ago. We disagree on factoring the round into the value equation. 

When he said in his summary that he doesn't like that it showed Rodgers as 4-6 picks too early, that sounds like we're not far apart. But that brings up another point from the article.

The VBD list you'll get is where we see players valued. NOT where you should draft them. You need to be doing better than just drafting players at the "fair" value. From the article:

5. DETERMINE THE AVERAGE DRAFT POSITION OF ALL THE PLAYERS

Now you have a ranking of players by value. But that's not enough. When you're drafting, the goal is to maximize value with every single pick. And to do that, you need to know where the other owners in your league are likely to draft each player.

Your info may show Julio Jones is your No. 5 overall ranked player by value. But all the average draft position data (ADP) shows Jones is being drafted at No. 8 in leagues across the country, you may be able to trade down to 8 and get Jones plus something. In the same way, if your VBD ranking shows Jones as the number 8 overall player and ADP shows he's being drafted at number 5 overall, that means you're likely not going to be drafting Julio Jones this year.

An excellent source for average draft value is our ADP on Footballguys.com. We compile a huge number of drafts by date and allow you to customize a list based on a few variables. If this sounds like too much work, do not fret - our VBD Excel App and the Draft Dominator software contain all of this data.

Remember that your X numbers represent worth to you. But if you draft everyone at worth, your draft will be average. The goal is to squeeze value with every single pick. In other words, with the 24th pick of the draft, you'd like to be drafting a guy that is in your top-15 overall. And that is done by keeping a pulse on what an average draft is likely to do.

J

 

[cloppbeast]

My problem with VDP: the fantasy player can select his own X number. The whole thing is worthless without having a precise and accurate way to figure out the X number. Not a matter of preference; a matter of fact.

 

[Long Ball Larry]

I agree with a lot of what you are saying, but I still think that it is a useful base to start with.  Here are a few thoughts

Second it requires you pick a good baseline for each position.  There are several theories on what your baseline should be and is also highly subjective.

I'm not sure that this matters so much because you are looking for a relative value.  You are still going to get a sense of how the players and positions relate to one another if you are using the same methodology across positions.

Third, VBD does not take into account opportunity cost vs ADP.

Agreed, which is why I always do my VBD chart and then pull in the current ADPs and highlight any spots where there seem to be clear discrepancies to find values and to see what really matches up to know when I have to take players.

VBD breaks down outside the top 60 or so players.  This is mainly due to the nature of projections being difficult for many players but also in that the difference between the players between 50 and 100 is usually very small.  Historically only 2 points per game separate WR30 from WR60.

Agree, which is why recently I've basically tried to use VDB to target roughly the most efficient starting lineup based on positional availability through the first 6-7 rounds and then look to fill out the rest of the roster with higher upside players.

I do agree that the volatility and variance that are in fact desirable for HTH match-ups is the biggest weakness, but I think that blending VDB as a base with a solid bench of hopeful breakout guys (or those that you can churn) will still lead to the best outcomes.  Especially in leagues where more teams make the playoffs, getting a consistent base will help you get there and then it is going to be somewhat of a crapshoot anyway.  As for challenges with projections, of course no one can predict the future, but I think that it makes more sense to try to get baseline projections that incorporate many different factors in order to anticipate a more likely outcome as opposed to just trying to find all the high-variance guys.  More often than not, the high-variance strategy is likely to bust.  I guess it is a question of whether you are trying to get a championship or bust every year.  I would like to know how many of the high variance sleepers over time have even been drafted and held vs. waiver wire pickups at some point.

 

[Joe Bryant]

I agree Larry. One thing we're continuing to work on is a "confidence" or "upside" factor that takes into account likely floors and ceilings. A RB that has all the tools and just needs a starter to stumble or get hurt is a different situation than an average talent role player. Both might see the same number of carries, but one has much more potential. Now, experienced players do those valuations in their head. What we're working toward is a way to make them show up in the lists better for people that may not be as experienced. 

J

 

[Simon Shepherd]

Our new Draft Dominator already supports upside too of course! The upside factor increases as the draft progresses and those numbers are applied against the original VBD baselines. It also projects picks between your current and next pick to calculate positional dropoffs. There's a bunch of other tweaks we do in there too on top of the basic VBD, like a separate set of baselines for bench players to rank players after the starter. 

http://www.footballguys.com/footballguys-dominator-mobile.php

 

[Joe Bryant]

Of course. As Simon stated, we're starting to see an offshoot in VBD where our tools like the Mobile Draft Dominator App and the Desktop Draft Dominator are taking VBD and pushing them even further in the direction we think is best. Sometimes that can be a different discussion than the basics or principles of VBD. But for us, it's how we've naturally advanced it.

What it does bring to light though is the fact that not all our products are equal. As much as we love the old VBD Excel App, it does not have the "confidence" or "upside" factors figured in. That's more advanced and something we do in the Mobile Draft Dominator and the Desktop Draft Dominator. 

J

 

[Joe Bryant]

One thing I want to do a better job of this year is making it more clear for users to get the very latest thinking behind what we're doing with our products. Almost every time I see a discussion of VBD, the topics brought up that people wish we accounted for are things we're already doing or have addressed. But that's on us to make sure we get our message out and have a clear understanding. Thanks.

J

 

[abbottjamesr]

JJ is a super smart and great guy. And I'm not sure if he still agrees with what he wrote several years ago. We disagree on factoring the round into the value equation. 

When he said in his summary that he doesn't like that it showed Rodgers as 4-6 picks too early, that sounds like we're not far apart. But that brings up another point from the article.

The VBD list you'll get is where we see players valued. NOT where you should draft them. You need to be doing better than just drafting players at the "fair" value. From the article:

J

Joe,

You are right that the current version of VBD that is presented in the article addresses many of the concerns that critics have had since its development.  I don't think any one really disagrees with the general principles of VBD that players should be ranked on value relative to their own position.  I think that your way of determining the baseline players for each position by using the top 100 players in adp is probably the best way of addressing the concerns that JJ raises with regard to replacement player cost being a factor.  

Where I think the issue is and what leads to the results not matching what most "feel" is the correct valuation for players is that, total points or even PPG is not the best way to determine how beneficial a player is to your fantasy team.  I know that sounds crazy but bear with me for a minute.

Total points or even PPG does not take into account how points are scored on a weekly basis and what those weekly points mean to your chances of winning that week.  FF is most often played as a series of weekly contests and your record over 13ish games determines if you will make the playoffs.  Your chances of winning weeks is greatly improved by having guys that have high weekly potential.  In non PPR having a RB or WR that scores over 25 points is often enough for you to win the week.  Some what obviously the players with the highest projected PPG are the most likely to have weeks that will have the greatest impacts and least likely to have dud weeks.  Where the current way of calculating VBD breaks down is that for each position the rate of the game winning scores is different.  For instance, the QB12 is more likely to have a 40+ point week than RB12 is to have a 25+ point week.  This means that the value for high end QB's is less relative to baseline than high end RB's to their baseline.

I went through the last two years of players and counted the numbers of games that each player had in the following scoring categories and gave each category a point value.  From this I calculated a "value" for each player.  Turns out over the last two years that this " value" correlates very highly to PPG at R^2 values over .9 for all positions.  The values can be calculated from projected PPG and then used in traditional VBD calculations to give a ranking or auction value. 

All of my data is for Non-PPR as that's what I typically play and had easiest to generate data for.  Scoring categories for each position.  These numbers are based off of historical values for the average weekly top 36, 24, 12, 6, 1 finishes at each position. Values for each category were given as listed below.  The general idea on the values was that one 25+ point performance will offset a 4 or so flex worthy performances relative to the average team of RB1,RB2,WR1,WR2, flex.  I don't know that the values are perfect but they correlate well to PPG and it seems close.

QB:         -15 / 15-19 / 19-23 / 23-27 / 27-35 / 35+

RB/WR:   -6 / 6-10 / 10-14/ 14-18 / 18-25 / 25+

TE :         -3 / 3-7 / 7-11 / 11-15 / 15-22 / 22+  

Value:       -1 /  1   /   2   /     3    /    4     /   6

Formulas to calculate "value" from PPG

QB:  3.505 * PPG - 40.519    R^2 = 0.96 (correlation between "value" and PPG)
RB/WR:   -0.0833 * PPG^2 + 6.3496* PPG - 34.755   R^2 = 0.95 (correlation between "value" and PPG)
TE:   4.544 * PPG - 13.4332     R^2 = 0.95 (correlation between "value" and PPG)

All this might be for nothing, but it creates auction values and rankings that feel more like what we expect in modern FF with QB values much lower as the  QB's are more valuable on a weekly basis than their PPG will lead you to believe, which is the basis for JJ's whole podcast on streaming free QB's.  

Baseline methodology.  I use a slightly different method of determining baseline for my version of VBD.  I looked historically at the number of games played over a 16 game season for the 60 highest PPG players over the last 5 years for RB's and WR's and the top 24 for QB and TE to determine how many of the 13 regular season games a typical FF relevant player will play.  Results were for a 13 game FF season

The average top 24 QB will play 10.28 games
 The average top 24 TE will play 10.34 games
The average top 36 RB will play 9.55 games
The average top 36 WR will play 10.92 games

From this I calculate the number of each position are "in play" for a season with the starting requirements for that league and use that as my base line.  For my start 2RB, 3WR, 1Flex league it is top 92 of the combined group of RB's and WR's.  The top 15 QB and TE. I kind of like using the 12th TE better but that is just preference.

James

 

[Joe Bryant]

Absolutely James. Playing around with different baselines as you're doing is a fundamental part of developing a style that works for you. I've always encouraged that type of thing. Great work there and good thinking!

J

 

[cloppbeast]

Absolutely James. Playing around with different baselines as you're doing is a fundamental part of developing a style that works for you. I've always encouraged that type of thing. Great work there and good thinking!

J

So you can play around with baselines until you get the result you want?

 

[32 Counter Pass]

Per the Principles of Value Based Draft: "We personally use 100 players as this basis..."

I am curious how you arrived at this conclusion. For a standard 12 team draft that is 4 picks into rd 9 which feels a little arbitrary. Maybe I am missing something. 

 

[Joe Bryant]

Per the Principles of Value Based Draft: "We personally use 100 players as this basis..."

I am curious how you arrived at this conclusion. For a standard 12 team draft that is 4 picks into rd 9 which feels a little arbitrary. Maybe I am missing something. 

Hi 32. We arrived there mainly through just trial and error and experimenting with different values and then mock drafting with the values. 

J

 

[32 Counter Pass]

How much difference would the results be if you just extended baseline to 9 full rounds? 

 

[DropKick]

I get that.  I'm talking about having had the same values on the input page for lineup requirements and scoring rules but last year there were 6 QBs in the top 60 players (in the "Cheat" page's "Overall - Use For Top 50-60 ONLY" column of the spreadsheet) and this year it's 22 of them.

Like I said, exact same lineup requirements and scorning rules.  So what's the difference in the formula?  (In my first iteration of the spreadsheet, I made no tweaks to the projections for the players.)

22?

That's a mistake in the calculations somewhere.  Bad cell reference, etc.  It isn't possible to have that many in the top 50 or 60.

 

[Joe Bryant]

How much difference would the results be if you just extended baseline to 9 full rounds? 

Hi 32. Not a lot. I usually like to talk about number of picks instead of rounds as the different league size makes rounds different for different users. 

J

 

[KingPrawn]

Not sure if this is the right spot for this but perhaps Joe can help out. Years ago when there still were magazines I read an article that expanded on VBD and converting the fantasy points to auction dollars. Would love to see this article again. It had something to do with figuring how many fantasy points were to be scored by your starters and then dividing that by how much money you allotted for your starters. You would then multiply the fantasy points by the dollars per point. Not sure of the details anymore so would appreciate it if someone could refresh my memory or link the article.

 

[Dez]

Take the players that score the most points

 

[32 Counter Pass]

Take the players that score the most points

So I should take QBs with my first five picks. Thanks D! 

 

[Biabreakable]

I think Joe has done a good job of communicating some of the misunderstandings or issues that people may have with VBD and points out that VBD isn't the only thing that one will consider when drafting. Obviously you would consider average draft position, roster management decisions, among other things when making your decisions.

I would describe most of the criticisms of VBD to be straw man arguments, such as JJ's claims that VBD tells people to draft QBs too high, when in my experience VBD suggests the opposite. It just a false claim about VBD and how people use it, to support his point of view. I mean were people seriously drafting multiple QB in round one as he claims? I guess there was a year when QBs went that high based on ADP but I don't think it was VBD driving that, it was the end results of the previous season and how so many rankings do not consider much more than that. Then people just drafted based on what those rankings were and more frequently just because of ADP.

It all seems very reasonable. If you have the biggest difference maker at a position, then you should have the most valuable team. And if you have the most valuable team, then you should have the best shot at winning, right?

Well, the second part of that is true. The first, regrettably, is not. Even if the point total of Aaron Rodgers to his position is greater than Ray Rice to his, it doesn’t mean Rodgers is more valuable.

And it’s the reason VBD is flawed.

This argument turns VBD into drafting the player expected to get the most total points, which is the exact opposite of what VBD is informing you about. He is basically making the same argument that VBD does in regards to not just drafting the player expected to get the most total points, just turning total points into VBD instead.  He is making the same argument that VBD does and has for decades.

That isn't the reason VBD is flawed but JJ's argument is. And he clearly knows better so why the face? 

What is more flawed is NEP that has no utility whatsoever in the real world of fantasy football.

 

[Ilov80s]

I think Joe has done a good job of communicating some of the misunderstandings or issues that people may have with VBD and points out that VBD isn't the only thing that one will consider when drafting. Obviously you would consider average draft position, roster management decisions, among other things when making your decisions.

I would describe most of the criticisms of VBD to be straw man arguments, such as JJ's claims that VBD tells people to draft QBs too high, when in my experience VBD suggests the opposite. It just a false claim about VBD and how people use it, to support his point of view. I mean were people seriously drafting multiple QB in round one as he claims? I guess there was a year when QBs went that high based on ADP but I don't think it was VBD driving that, it was the end results of the previous season and how so many rankings do not consider much more than that. Then people just drafted based on what those rankings were and more frequently just because of ADP.

Obviously that article by JJ was written awhile ago. In 2012, the ADP for QBs had 4 in the first round and a 5th in the 2nd round. Here is a VBD table for the 2011 season which is part of the reason many people were suggesting drafting QBs so high. Rodgers was 3rd, Brees was 5th, Cam was 8th, Brady was 9th and  Stafford was 14th. So ADP for 2012 was actually pushing against the VBD results of 2011. Obviously projections change year to year, but if people were looking at 2011 VBD, they would have been more inclined to take the top QBs earlier than the 2012 ADP suggested. 

QB is a spot where VBD totally misses. In most leagues, the supply/demand of QB is so out of whack that quality starters can be had on the WW as the year goes on. Also, you don't just have to draft 1 QB and play him every week. Below average to bad QBs  take advantage of bad defenses on a weekly basis and as JJ has documented, an owner can use a frankenstein's monster of QBs throughout the year to achieve top 8QB numbers without spending almost any draft capital. With QB, looking at end of the year numbers isn't all that important.

 

[Joe Bryant]

QB is a spot where VBD totally misses. In most leagues, the supply/demand of QB is so out of whack that quality starters can be had on the WW as the year goes on. Also, you don't just have to draft 1 QB and play him every week.

Hi Iluve80s,

VBD is completely driven by supply and demand. If a position is deep with talent (as QB can often be), the value is minimized due to supply and demand. VBD puts actual numbers from projected points into the equation. Not just feelings or assumptions. In other words, it's not a "QBs are deep this year so I can wait", it's more a "We project this QB to score ___ points and that is ____ points above the worst or average starter I could replace him with so I can wait". (Or not, depending on your league factors) It's completely dependent on the scoring system and league size and roster requirements as all those affect the demand. And keepers of course affect the supply. And if you're using something like the Mobile or Desktop Draft Dominator, you can make those supply and demand adjustments dynamically on the fly after every pick. 

J

 

[Ilov80s]

Hi Iluve80s,

VBD is completely driven by supply and demand. If a position is deep with talent (as QB can often be), the value is minimized due to supply and demand. VBD puts actual numbers from projected points into the equation. Not just feelings or assumptions. In other words, it's not a "QBs are deep this year so I can wait", it's more a "We project this QB to score ___ points and that is ____ points above the worst or average starter I could replace him with so I can wait". (Or not, depending on your league factors) It's completely dependent on the scoring system and league size and roster requirements as all those affect the demand. And keepers of course affect the supply. And if you're using something like the Mobile or Desktop Draft Dominator, you can make those supply and demand adjustments dynamically on the fly after every pick. 

J

I understand this, but VBD is basing it on year end projections. There are going to be 45-60 QBs that post at least one QB1 week each year. Since most leagues only see 16-24 QBs rostered at any given time, that means there are always QBs capable of QB1 performances available on the WW and going in the latest of fantasy draft rounds/$1. Knowing which player to start from week to week isn't as easy as just plugging Drew Brees in every week, but it's also not that difficult to identify QBs playing weak defenses. Not to mention, every year there are several of these late QBs that just end up emerging as weekly starters. So with QB, I think VBD just misses because it is based on the idea someone will draft a QB and play that QB all year. 

 

[Joe Bryant]

I understand this, but VBD is basing it on year end projections. There are going to be 45-60 QBs that post at least one QB1 week each year. Since most leagues only see 16-24 QBs rostered at any given time, that means there are always QBs capable of QB1 performances available on the WW and going in the latest of fantasy draft rounds/$1. Knowing which player to start from week to week isn't as easy as just plugging Drew Brees in every week, but it's also not that difficult to identify QBs playing weak defenses. Not to mention, every year there are several of these late QBs that just end up emerging as weekly starters. So with QB, I think VBD just misses because it is based on the idea someone will draft a QB and play that QB all year. 

For sure. All positions benefit from streaming options. Our Sigmund Bloom's made a career out of streaming defenses. For early in the draft, most owners are assuming a full season for the player. But as the draft unfolds, punting a position and streaming can be viable. But again, it all depends on the league. In a 14 team league starting 2QBs, streaming QBs is going to look very different than a 10 team league starting 1 QB. VBD accounts for all of that. Thanks for the discussion.

J

 

[Ilov80s]

For sure. All positions benefit from streaming options. Our Sigmund Bloom's made a career out of streaming defenses. For early in the draft, most owners are assuming a full season for the player. But as the draft unfolds, punting a position and streaming can be viable. But again, it all depends on the league. In a 14 team league starting 2QBs, streaming QBs is going to look very different than a 10 team league starting 1 QB. VBD accounts for all of that. Thanks for the discussion.

J

I think one of the most helpful things with DD and VBD is for people that play in multiple leagues that have varying scoring formats and lineup requirements. Like you said, there's a lot of possible variance in value in different leagues. Having a program sort out the differences between leagues is important. 

 

[abbottjamesr]

I understand this, but VBD is basing it on year end projections. There are going to be 45-60 QBs that post at least one QB1 week each year. Since most leagues only see 16-24 QBs rostered at any given time, that means there are always QBs capable of QB1 performances available on the WW and going in the latest of fantasy draft rounds/$1. Knowing which player to start from week to week isn't as easy as just plugging Drew Brees in every week, but it's also not that difficult to identify QBs playing weak defenses. Not to mention, every year there are several of these late QBs that just end up emerging as weekly starters. So with QB, I think VBD just misses because it is based on the idea someone will draft a QB and play that QB all year. 

This was kind of the point I was trying to make with my way to long post before.  In 2016 there were 16 different QB's with at least 6 top 12 weekly finishes to put that in perspective there were only 10 RB's with at least 6 top 12 weekly performances and we start 2-3 times as many of them.  These ratios are pretty consistent over the last several years and in my opinion the one issue with using VBD in its current form of  using PPG or season total points.  

 

[Joe Bryant]

I think one of the most helpful things with DD and VBD is for people that play in multiple leagues that have varying scoring formats and lineup requirements. Like you said, there's a lot of possible variance in value in different leagues. Having a program sort out the differences between leagues is important. 

For sure. I know it seems crazy to you guys here that are advanced players but I'm always surprised how many casual players don't understand the impact of their specific league's scoring system and rules and league size. We currently have Drew Brees projected for 33 TDs and about 4900 yards. Those stats in a 10 Team / 1 starting QB league would put Brees in a radically draft value than Brees in a 14 team  / 2 starting QB league. You guys totally get that. But I see tons of casual players who don't. 

It's why I think we're seeing such a push towards "unusual" league formats like Superflex and large (or small) leagues. If I wanted to win my local league, I'd do everything I could to push for out of the box format to separate from the people using generic cheatsheets from our competitors...

J

 

[Ilov80s]

For sure. I know it seems crazy to you guys here that are advanced players but I'm always surprised how many casual players don't understand the impact of their specific league's scoring system and rules and league size. We currently have Drew Brees projected for 33 TDs and about 4900 yards. Those stats in a 10 Team / 1 starting QB league would put Brees in a radically draft value than Brees in a 14 team  / 2 starting QB league. You guys totally get that. But I see tons of casual players who don't. 

It's why I think we're seeing such a push towards "unusual" league formats like Superflex and large (or small) leagues. If I wanted to win my local league, I'd do everything I could to push for out of the box format to separate from the people using generic cheatsheets from our competitors...

J

Totally. I play in an online league made up of people from a beer related website. They are obviously pretty basic level players. For whatever reason, the commish of that league made it point per carry. For 2 years in a row, the drafters have all completely failed to realize the monster impact that has. That makes even mid level RB1s out scoring QBs. Plus you can start 3 of them. It's a pretty easy league to draft in as I just hammer RB. It was especially easy last year as half the teams went "zero RB" which destroyed their teams.