Rating the NFBC drafts

ToddZ · Post by **ToddZ** » Fri Apr 22, 2005 4:45 am

Second, as I mentioned earlier I believe auctions are more representative of people’s player preferences because they bid all the way up to how much they think the player is worth while in drafts they try to draft their sleepers later than their true value. This really isn't true. In auctions, there are two ways to get profits.

1. Money management--realizing players are going for inflated prices and holding back until players go for deflated prices

2. Rostering sleepers at a price $1 more than the next guy would pay. A loose definition of sleeper is someone you believe will do better than others believe so it will only cost you $1 more.

The dynamics of an auction are so much different than a draft. The prices in an auction are controlled by the order the players are tabled (as suggested) and dictated by the strategies of the participants. All it takes is a couple of guys who want to build their MI with studs and the price of 2B and SS are way above projected value. The amount overpaid HAS to be underpaid elsewhere. So now you are doubling the error in the analysis as the "value" of the studs is too high and the "value" of the mid tier is too low.

Chest Rockwell · Post by **Chest Rockwell** » Fri Apr 22, 2005 5:07 am

It was not a compliment...

John Patterson for CY Young sorry that was another thread.

Chest Rockwell · Post by **Chest Rockwell** » Fri Apr 22, 2005 5:13 am

Also on the flawed logic concept, it is your opinion not fact. I like his analysis for what it is one approach of many of determining who did well on draft day. Just curious what you saw in Sean Casey besides injury prone, inconsistent, and extremely poor power numbers relative to position and pick.

hughes · Post by **hughes** » Fri Apr 22, 2005 5:35 am

If I were a betting man, I'd love to take anyone who thought there was zero correlation here on on a bet that 11 or more (10 as push) of the first place teams come from the top 150 on this analysis. If you assume the auction leagues knew what they were doing at some level, the idea that this has zero correlation is almost surely wrong.

Regardless of whether this correlates to actual outcome, I have a strong hunch that a correlation exists between those who paid less attention to positional scarcity and those higher on the analysis that is the subject of this thread. I picked zero mis or cs (or for that matter, cis) in first ten rounds (5 ofs, 3sps, 2 rps), and am ranked 8th on this analysis. This strategy was not intentional (co-manager was ready to beat me with computer), but is one I see myself doing often. As a huge generalization, it seems to me that, in auctions, while anyone is willing to throw $20 at on of like Giles or Gonzo, they are more willing to use the corresponding draft pick to go after a guy because of position. Stated another way: auctions value positional scarcity less. My theory of why is that auction participants are more keenly aware of $'s earned based on projections and history and don't vary from it as much when picking teams. Drafters say "I need a ss" more readily and throw Orlando Cabrera's name out in the eight. And if you're in an auction, it's easier to overpay by just $1 when you think you really need someone at a position, but in a draft if you actually fear a run, the opportunity cost can be much greater. Plus, auction people often say at some lever "I'm going to spend $100 on my outfield/5 category bats."

The why theory/ies are irrelevant, but I think the observation re draft/auction is right. Seems that someone could compare the positions to the dollar values/draft positions and confirm whether this is right.

Now what do I do with Lance Berkman returning as my 7th outfielder?

Dyv · Post by **Dyv** » Fri Apr 22, 2005 5:55 am

Hughes - I have the same problem with Berkman, but somehow I'll find room for his bat

Now if Griffey goes out so Wily Mo gets full time, plus Berkman and Bonds both come back then my decisions are a major crunch weekly.... but I'll gladly take that problem (as I'm sure you would!)

I believe that if someone is in the top 150 on your list AND has selected good value then they appear to have a leg up on someone in the bottom 150 who has selected good value. That just rings as common sense to me. Not to say that anyone should give up as the value of administering your FA and how you built your team is not accounted for in these calcs and is at least 1/3 of your final ranking.

The fact I got Wily Taveras in round 28 isn't garnering me any credit on your rankings, but his stolen bases and even a HR are helping me so far.

I'll bet up to $100 that Puffin's correlation is correct. I'll bet that of the 60 league money winners (20 leagues x 3 spots) you'll find at least 40 come from the top 150 on these preseason rankings. Any takers?

Dyv

ToddZ · Post by **ToddZ** » Fri Apr 22, 2005 5:56 am

It was not a compliment... Didn't think so...

Thank your friend for me though.

John Patterson for CY Young sorry that was another thread. Cy Young? Excellent use of a straw man's argument. I never claimed he would challenge for the Cy.

Was this the same thread I also pointed out that yesterday's performance, as good as it was, was just one game and the jury is still out on both his future performance and the validity of my bid?

Also on the flawed logic concept, it is your opinion not fact. I like his analysis for what it is one approach of many of determining who did well on draft day. It is my opinion that it is a fact the logic is flawed. No, forget that, it's wrong.

Find an independent projection set you trust, with values customized to a 15-team mixed league and compare those values with the values utilized in this study.

The use of auction values as a means to judge how well drafters did is not a valid means to conduct the analysis.

A is wrong.

B matches A.

Because B matches A, B is right.

Sorry, doesn't work for me.

The presumption is A is right--it is not, it is wrong.

{Just curious what you saw in Sean Casey besides injury prone, inconsistent, and extremely poor power numbers relative to position and pick. Casey has played 146 and 147 games the past two season, getting 571 and 573 AB. The injury-prone label is more perception than reality. He has basically missed 15 games each of the past 2 season, equivalent to one stay on the 15-day DL. The shoulder woes are a thing of the past. The calf injury he suffered last year was unrelated to anything previous plus he had his knee scoped in the offseason which should help keep him healthy. So even if we projected the same rate of performance for Casey, my raw numbers would be higher as I would have him playing more games.

But I also see an uptick in his rate of performance. His ability to hit for average has never been in question, only his ability to hit HRs. At the beginning of his career, it was assumed some of the doubles he smacked would eventually turn into homers. Some novel research by Mat Olkin showed Casey tended to hit more grounders than fly balls, meaning his HR total may not see the expected spike. Then he got hurt so it was difficult to tell if he would make any adjustments, or remain a Mark Grace/John Olerud type hitter--high BA, lower power for a corner position. But last season his GB/FB ratio tended a bit more towards fly balls, something I am hoping is a sign there is still a chance for increased power.

Last season he ended with 24 HR and 99 RBI. My expectations this season were nothing more than that plus maybe 2-3 HR and 10 more RBI, hoping he was healthy for 155 or so games.

By saying I liked Patterson, the assumption was made I predicted he would get a Cy Young. I'll save you the hyperbole this time and say I never predicted a triple crown or MVP for Casey, I only stated I like him more than others like him--that's all.

[ April 22, 2005, 12:04 PM: Message edited by: ToddZ ]

newkidintown · Post by **newkidintown** » Fri Apr 22, 2005 6:08 am

Where is this list and how can I get access to it? Thanks.

nydownunder · Post by **nydownunder** » Fri Apr 22, 2005 6:13 am

Puffins

Well done! I think your value methodology is no worse than anything else out there. The dollar value averages would probably come pretty close to the average stat projections of all those managers players anyway, which saves time....and those managers would probably use some sort of average form 4-8 publications anyway.

Dyv · Post by **Dyv** » Fri Apr 22, 2005 6:23 am

Originally posted by newkidintown:

Where is this list and how can I get access to it? Thanks. At the start of this thread....

nydownunder · Post by **nydownunder** » Fri Apr 22, 2005 6:35 am

ToddZ

Keep in mind his analysis is a baseline analysis of pre-season projections (that includes mangers directly and publications indirectly). And no one manager or publication is 100% accurate. All the analysis is saying is that based on publicly available information (ie player projections), which is then translated my managers into $ auction values, this is how everyone would perform if every player meets their average projection. And whether you believe it or not it translates very well in how these mangers will perform for the year because on average about 1/3 of your team will outperform, 1/3 underperform, and 1/3 perform. If you players have well established track records you will come pretty close to estimates barring injury. If you have taken risks on both ends (ie youth and age) you could end up far over or under achieving.

In a nut shell, the winner this year is somewhere in the top 40 of this - that you can take to the bank.

And if you don't agree with his analysis, stop picking it apart, do your own, and publish it like Puffins did.

nydownunder · Post by **nydownunder** » Fri Apr 22, 2005 6:40 am

[

[ April 22, 2005, 12:40 PM: Message edited by: nydownunder ]

Puffins · Post by **Puffins** » Fri Apr 22, 2005 9:55 am

Originally posted by ToddZ:

Great--therefore you can't assume the winners will emanate from the top-150 as it is the value obtained from superior knowledge that separates the winners from the losers. Your superior knowledge will be reflected by having worse results by this means of evaluation, hence my contention that I would not bet the eventual winners come from the top-150. There are two ways you can get value in the draft: (1) drafting players later than the pool of NFBC players thinks they’re worth, (2) drafting players later than they’re truly worth using superior knowledge. I measured the first part and believe that is the greater part of the total. It is not clear to me that superior knowledge exists. If you’re the only one with a certain opinion, there’s a good chance you’re wrong. If 2 people in an auction know that Sean Casey is going to hit .324 with 27 homers and 109 RBI, they will bid him up to top dollar. You have to be the only one with certain knowledge for superior information to exist. There may be cases where superior knowledge does exist (knowing that Lyon was the definite closer in Arizona) and those drafters were penalized, but I expect that those cases are rare.

If you believe Sean Casey is better than everyone else thinks he is, maybe you do have some superior knowledge, but it is possible that you’re just wrong. When I look at Sean Casey, I see someone who only hit one home run measured over 405 feet last year and expect a drop in home runs.

Originally posted by ToddZ:

All it takes is a couple of guys who want to build their MI with studs and the price of 2B and SS are way above projected value. Since I’m using the median auction values, this has to happen in 2-3 of the auction leagues to have an effect.

ToddZ · Post by **ToddZ** » Fri Apr 22, 2005 10:29 am

(1) drafting players later than the pool of NFBC players thinks they’re worth, (2) drafting players later than they’re truly worth using superior knowledge. I measured the first part and believe that is the greater part of the total. No, I'm sorry, but you didn't. The foundation for the analysis is the median value of the 4 auctions. This is not representative of the projected values of all 345 players. There is a big difference between projected value and bid value. If there were not, wouldn't the bid value for all 4 auctions be basically the same? Are they? You yourself admit there are flaws to the system--from the page linked in the initial post of this thread...

There are flaws with using the auction values to value the players, but I am going to ignore them in order to keep the player valuation system simple. Below are the dollar being used for each player: Analysis pertaining to and conclusions drawn from flawed data, are by definition flawed.

We are expected to ignore the flaws in the system, but yet I am the bad guy for suggesting we also ignore the resultant conclusions.

[ April 22, 2005, 04:31 PM: Message edited by: ToddZ ]

Puffins · Post by **Puffins** » Fri Apr 22, 2005 10:34 am

Originally posted by hughes:

As a huge generalization, it seems to me that, in auctions, while anyone is willing to throw $20 at on of like Giles or Gonzo, they are more willing to use the corresponding draft pick to go after a guy because of position. Stated another way: auctions value positional scarcity less. My theory of why is that auction participants are more keenly aware of $'s earned based on projections and history and don't vary from it as much when picking teams. Drafters say "I need a ss" more readily and throw Orlando Cabrera's name out in the eight. And if you're in an auction, it's easier to overpay by just $1 when you think you really need someone at a position, but in a draft if you actually fear a run, the opportunity cost can be much greater. Plus, auction people often say at some lever "I'm going to spend $100 on my outfield/5 category bats."

The why theory/ies are irrelevant, but I think the observation re draft/auction is right. Seems that someone could compare the positions to the dollar values/draft positions and confirm whether this is right. I just did this and your theory appears to be correct. I looked at the dollar values of the average draft positions for middle infielders and they were higher than the auction values for nearly every middle infielder (Furcal being an exception).

Puffins · Post by **Puffins** » Fri Apr 22, 2005 10:58 am

Originally posted by ToddZ:

There is a big difference between projected value and bid value. If there were not, wouldn't the bid value for all 4 auctions be basically the same? Are they?That’s why I use all 4 auctions to determine the value. With 1 auction the values would be way off. Four auctions and the values are better, but still not perfect.

Originally posted by ToddZ:

You yourself admit there are flaws to the system Yes, there are flaws and I fully admit that. But there still is predictive power in my data. I could calculate batting average by counting doubles as outs and that would be a flawed system. Lyle Overbay would be cheated with that flawed system, but there still is predictive power in the flawed batting average.

I also want to repeat what Dyv said earlier: “Puffins' analysis is for amusement only.” I am not claiming these as the true values, but I guarantee you there will be some positive correlation.

If you would like to prove that your player values are superior, send me your draft day values, and I’ll rank the teams according to your dollar values. We can then see how much stronger the correlation is when using your dollar values.

ToddZ · Post by **ToddZ** » Fri Apr 22, 2005 11:42 am

I could calculate batting average by counting doubles as outs and that would be a flawed system. Lyle Overbay would be cheated with that flawed system, but there still is predictive power in the flawed batting average. Sorry, but I philosophically disagree with this concept, which is perhaps the root of our disagreement--which is perfectly fine. I'm a trained scientist and cannot draw conclusions based on flawed data. The researchers and pharmaceutical companies that buy the product I make are probably happy that is the case.

If you would like to prove that your player values are superior, send me your draft day values, and I’ll rank the teams according to your dollar values. We can then see how much stronger the correlation is when using your dollar values. I appreciate your offer (not saying this sarcastically). But my intent has never been to prove my values are superior, just to suggest that the conclusions people are drawing from this work are not particularly valid. Believe me, I appreciate the effort that goes into this type of thing more than anyone.

I did a series of valuation based studies based on 2004 NFBC data which I will repeat and expand upon this off season.

It is difficult to find a set of auction values customized for a 15-team mixed league. But I assure anyone who cares that the majority of the players in the top-20 in this study have values over $5 higher than would be projected. And as has been explained, an auction is a sum-zero economy, so for every buck misallocated at the top, there will be a buck misallocated in the other direction elsewhere.

I found it pretty interesting that using your median values, the sum of the top 345 players was about $3840, or $60 short of the proper economy. That's not too much at all. Equally distributed over the entire population, that is about 17 cents a player, so 1 in 6 players will have their value rounded up if given those 17 pennies.

hughes · Post by **hughes** » Fri Apr 22, 2005 12:04 pm

It seems to me that Puffins' analysis extracts value from four auction leagues (throw out top and bottom, use average of middle two) and compares it against each individual pick. A similar tool could be used by taking 20 drafts, throw out top 5 and bottom 5, average middle ten draft slots, and compare against the auctions to see similarly "evaluate the auctions did." The latter would be better bc 20 against 1 is better than 4 against 1. There weren't, however, 20 auctions. To compare 1 on 1 is stupid: there's no way to know which is correct. To compare 4 on 1 by throwing out outliers is what it is. I personally think it's likely to be pretty good. It's certainly likely to be somewhat predictive. By Todd's theory, it wouldn't really matter 2, 20, or 1,000 to 1 bc what the consensus predicts is not likely to bear any correlation to value. There is of course an underlying prejudice that his (or any individual's) predictive value is superior to the consensus of a large group of people willing to shell out a grand plus and spend the associated time doing this. I think that he's wrong. Of course there was another guy on the site bragging about getting 17 of his top 21 players in a mixed draft, and it turned out he picked sheets and peavy first two. Similar logic. Flawed if you ask me. Let's see how many of the top 60 slots end up in the first 150 picks on puffin's analysis. Anyone who'd bet money against is someone I'd like to sit down at a card table with.

ToddZ · Post by **ToddZ** » Fri Apr 22, 2005 12:26 pm

By Todd's theory, it wouldn't really matter 2, 20, or 1,000 to 1 bc what the consensus predicts is not likely to bear any correlation to value. There is of course an underlying prejudice that his (or any individual's) predictive value is superior to the consensus of a large group of people willing to shell out a grand plus and spend the associated time doing this. I think that he's wrong. Sorry, but this is NOT my point. The 2, 20, 1000 part is true, except not for thereason you suggest. The underlying prejudice has nothing to do with my point. Simply put, the dynamics of an auction are vastly different from the dynamics of a draft. That is the crux of my argument. You cannot compare data generated from one to data of the other, as the data does not represent the same thing.

One represents the median of how four people valued a player in an auction format, the other how people valued players in a draft format. As I said before, this is apples and oranges.

The defining element of an auction is everyone has a shot at every player. You can design strategies accordingly. The most extreme strategy, and the one which skews this analysis the most is stars and scrubs. The idea is to overpay for the top players and fill in the back end with scrubs. Over the course of the season, you then use the fairly deep free agent pool to upgrade the scrubs, earning your team profit. The stars and scrubs strategy is not available in a draft.

That is my argument, not that my values are better than the median of the four auction drafts.

To quote myself, "For kicks, I just compared my dollar values versus those used in this study and there is a huge difference. Ok, they're my values which are worth as much as the paper I printed them out on"

If this implied my intention was to show my values are better, I apologize. My intention was to show my values are very representative of how most would value the majority of the players for an auction, and the bid values at the top end clearly showed stars and scrubs was the prevailing strategy.

[ April 22, 2005, 06:27 PM: Message edited by: ToddZ ]

hughes · Post by **hughes** » Fri Apr 22, 2005 12:45 pm

Point is to get the highest value (i.e., most production in ten categories from 23 players) period. Draft or auction irrelevant, just vehicle to get the players (really their stats). Once you start talking re strategy being different, if it speaks to how players should be valued and a claim that it is different bbased solely on whether you've bid or picked them, you, are either speaking to a nuance that I don't believe exists/fail to comprehend, or you're thinking too much on this. Unless it speaks to different league depths, and even that is often overrated. I could be wrong, there is no website under my name, but I think on this not.

JohnZ · Post by **JohnZ** » Fri Apr 22, 2005 1:22 pm

Interesting thread....

I enjoyed the site.

I love doing research like this myself, and after two decades have come to realize that all fantasy analysis is flawed to some degree or another. That's what makes the game so fun if you think about it. Those that search for the holy grail on fantasy numbers crack me up. They usually talk in "absolutes".

I have a little bit of a hard time with the fantasy auction values in those 4 drafts. I can see why he used them, and I can see why they are flawed.

No one can tell me that my draft is flawed because I took Tejeda #3. If his numbers were labled "Alex Rodriguez", he'd been #1 in every daft. JMO.

What value do I get for my G.Zaun pick, laughed at by DYV

, but was the third from last pick of the entire draft? Does that offset the loss I get on Tejeda?

Zaun just hit another RBI double. My analysis is awesome

When will Jeremy Reed steal a base so I can get in the Top 20?

[ April 22, 2005, 07:27 PM: Message edited by: UFS ]

ToddZ · Post by **ToddZ** » Fri Apr 22, 2005 1:23 pm

After driving home (amazing, I seem to do my best thinking at traffic lights) I think I have a way to present my case. Some of what Hughes just posted is valid about the manner I am approaching this.

Let me introduce four sets of player rankings:

1. A list of 345 player assigned value randomly so that the sum is equal to $3900 (15 x $260)

2. An end-of-the-season ranking based on actual stats

3. The median values discussed here

4. An list based upon the average values of your favorite 10 fantasy baseball prognosticators.

It is obvious that the winners will have the most correlation to the second list (actual year end values) and the worst correlation to the first list (random).

My point is I feel the 4th list is a better indicator than the third list. The median list will in all probability be based upon the prognosticator's list, but normal auction dynamics will introduce some fluctuations, most notably the stars and scrubs effect.

But will puffins list have some sort of positive correlation to winners? Of course it will because even with the flaws, ir is still a much better ordering of player's value than a randomly generated list (using an extreme example). Perhaps in my zeal to point out what I perceive to be flaws, I overlooked the positive correlation it will have.

Truth be told, I think this is a really neat exercise if thought of in the vein "let's see which is better, median values of NFBC auction drafts or an average of these 10 published value sets." Let's not presuppose anything and see what happens (which I suspect is the primary intent anyways).

The Lollygaggers · Post by **The Lollygaggers** » Fri Apr 22, 2005 1:40 pm

Very interesting work, Michael. Nice job!

I was curious to see how last year's top finishers ranked on your list. I've listed the returning league champs plus the top 2 overall finishers who weren't league champs (got the names from the NFBC Mock Draft in December). Not sure what it means, but here is the data:

248 Artie Rastelli

208 Aaron Seefeldt

210 Stanley Kaye

120 Jim Ruland

63 Leonard Ringle

102 Stephen Jupinka

42 Dave Cushard

221 Jeff Thomas

96 Eric Peden

17 Terrance Haney

77 Shawn Childs

132 Clark Olson

141 Ari Benjamin

100 Karl Mische

??? Bill Begley

Median rank is 111. 10 of the 14 (71%) rank in the Top 50% (ie, top 150). It will be interesting (although I'm not sure how meaningful) to see at the end of the year if the above list correlates at all with repeat success.

mulberry · Post by **mulberry** » Fri Apr 22, 2005 1:50 pm

Funny I just looked at the same thing. How can the top 2 overall finishers last year rank 248th and 208th coming out of the draft and all the league winners be that far down as well. Plus the straight draft guru for hq is 245th? One would think these guys come prepared to the draft and knew what they were doing enough to make this data flawed. Interesting read though, Puffins.

[ April 22, 2005, 07:51 PM: Message edited by: mulberry ]

Dyv · Post by **Dyv** » Fri Apr 22, 2005 3:08 pm

Originally posted by mulberry:

Funny I just looked at the same thing. How can the top 2 overall finishers last year rank 248th and 208th coming out of the draft and all the league winners be that far down as well. Plus the straight draft guru for hq is 245th? One would think these guys come prepared to the draft and knew what they were doing enough to make this data flawed. Interesting read though, Puffins. What makes you think the straight draft guru from HQ is any good? I have no idea who this guy is or what his credentials are, just curious why he's given special note?

Does anyone have the number crunching acumen to compare Puffins top 150 to the current top 150 and see how close it is RIGHT NOW ? Obviously this has little real value, but just curious... is there any trending already noticeable?

Dyv