Originally posted by KJ Duke:
quote:Originally posted by bjoak:
Not angry. My model is perfect in terms of valuation if using last year's projections and 2005's NFBC results. Adding pitchers, 3 year NFBC results, and/or 2006 real stats would have you more updated and accurate than me.
But, uh, #24 is exactly where he was taken this year and I don't know that his ADP last year was any higher so I don't see how he "always goes 5-10 spots higher than he should."
My point exactly is that it's no big deal to have him 5 positions later. Everyone is 5-10 positions higher or lower so there's no reason to make a big deal about one specific guy. Wasn't a big deal, just a point of view. Not sure what his ADP was, I recall him going around the first turn in several drafts which seemed early given more proven guys. I just sense Sizemore-love from a lot of fantasy players, a guy that draws oohs and aahs from the crowd when he's drafted; I don't get it. But you're right, his selection here probably wasn't that far off - maybe it was the exclamation point that got me!
If you have any kind of mathematical, objective valuation model (even if it's incorrect), you already know that there is not a linear relationship between ADP and projections.
Well, everyone's projections are different anyway, so I don't know how you could judge this unless you are assuming last year's value or some stat service as a consensus. But I would say if there was a consensus projected value that range would be much tighter in the first couple of rounds. [/QUOTE]Sizemore's ADP was 28 and even if 29 was correct he was still better than many of the guys who were drafted before him but had less value (all 27 guys drafted ahead of him probably didn't make up 27 of the 28 who ended up with more value at the end of the season).
As for projections, you can make assumptions based on average draft results and how different they are. Many people have different projections but for the most part they won't be ludicrous.
But ADP's don't hold true regardless of the values you place on individual stats. For example, Let's assume that a run is worth 1, a HR 5, an RBI 1, a SB 4, and we'll leave out average to make it simpler (these values are arbitrary):
Now let's take two sets of players next to each other in the ADP's:
Tex 105, 41, 114, 4.
105+205+114+16=440
Crawford 103, 15, 65, 48
103+75+65+192=435
If we assumed these values to be true Crawford would be slightly overvalued because of speed, but now let's look at Rollins and Tejada:
Tejada: 85, 25, 104, 6
85+125+104+24=338
Rollins: 105, 13, 67, 40
105+65+67+160=397
By this estimate, first of all, the 2 players have vastly different values. On top of that, Tejada's power is the overrated ability this time.
But no matter what valuation model you use it won't come out right. If you used a model that rewarded Tejada's talents more, then they would have similar values, but when you go back and apply it to Crawford and Tex, their values will be skewed and there will be a huge gap between them. I suppose you can argue projections here, but as I said the projections would have to be pretty wild to make these draft positions sensical.
[ November 08, 2006, 06:45 PM: Message edited by: bjoak ]
Sizemore and stat valuation thread
Sizemore and stat valuation thread
Chance favors the prepared mind.
Sizemore and stat valuation thread
Originally posted by bjoak:
Sizemore's ADP was 28 ... Surprised he went that low, hadn't checked the data, the "market" overall got it about right then.
Sizemore's ADP was 28 ... Surprised he went that low, hadn't checked the data, the "market" overall got it about right then.
Sizemore and stat valuation thread
There is a link for the ADP's off of the main BFBC site.
Chance favors the prepared mind.
Sizemore and stat valuation thread
Originally posted by bjoak:
... ADP's don't hold true regardless of the values you place on individual stats. For example, Let's assume that a run is worth 1, a HR 5, an RBI 1, a SB 4, and we'll leave out average to make it simpler (these values are arbitrary):
Now let's take two sets of players next to each other in the ADP's:
Tex 105, 41, 114, 4.
105+205+114+16=440
Crawford 103, 15, 65, 48
103+75+65+192=435
If we assumed these values to be true Crawford would be slightly overvalued because of speed, but now let's look at Rollins and Tejada:
Tejada: 85, 25, 104, 6
85+125+104+24=338
Rollins: 105, 13, 67, 40
105+65+67+160=397
By this estimate, first of all, the 2 players have vastly different values. On top of that, Tejada's power is the overrated ability this time.
But no matter what valuation model you use it won't come out right. If you used a model that rewarded Tejada's talents more, then they would have similar values, but when you go back and apply it to Crawford and Tex, their values will be skewed and there will be a huge gap between them. I suppose you can argue projections here, but as I said the projections would have to be pretty wild to make these draft positions sensical. bjoak, there are ways to equalize data across different categories. Without diving fully into details of how I run this, it is basically converting raw data into proportional data as measured by points of contribution to league standings and then coming up with the equivalent of auction relative values. As such, it is dependent not only on lge structure, but on various assumptions, such as overall lge data.
That said, I still don't get your point about ADP matching up with expected player value. Different people have a) different evaluation methods, b) different player expectations, c) different positional relative values, and d) just plain personal preferences in many aspects of assembling a team. Thus, I don't quite understand how you come to the conclusion that if a player is worth $X, he will go within 5-10 spots of that value.
... ADP's don't hold true regardless of the values you place on individual stats. For example, Let's assume that a run is worth 1, a HR 5, an RBI 1, a SB 4, and we'll leave out average to make it simpler (these values are arbitrary):
Now let's take two sets of players next to each other in the ADP's:
Tex 105, 41, 114, 4.
105+205+114+16=440
Crawford 103, 15, 65, 48
103+75+65+192=435
If we assumed these values to be true Crawford would be slightly overvalued because of speed, but now let's look at Rollins and Tejada:
Tejada: 85, 25, 104, 6
85+125+104+24=338
Rollins: 105, 13, 67, 40
105+65+67+160=397
By this estimate, first of all, the 2 players have vastly different values. On top of that, Tejada's power is the overrated ability this time.
But no matter what valuation model you use it won't come out right. If you used a model that rewarded Tejada's talents more, then they would have similar values, but when you go back and apply it to Crawford and Tex, their values will be skewed and there will be a huge gap between them. I suppose you can argue projections here, but as I said the projections would have to be pretty wild to make these draft positions sensical. bjoak, there are ways to equalize data across different categories. Without diving fully into details of how I run this, it is basically converting raw data into proportional data as measured by points of contribution to league standings and then coming up with the equivalent of auction relative values. As such, it is dependent not only on lge structure, but on various assumptions, such as overall lge data.
That said, I still don't get your point about ADP matching up with expected player value. Different people have a) different evaluation methods, b) different player expectations, c) different positional relative values, and d) just plain personal preferences in many aspects of assembling a team. Thus, I don't quite understand how you come to the conclusion that if a player is worth $X, he will go within 5-10 spots of that value.
Sizemore and stat valuation thread
Originally posted by bjoak:
There is a link for the ADP's off of the main BFBC site. I know, just being lazy.
There is a link for the ADP's off of the main BFBC site. I know, just being lazy.
Sizemore and stat valuation thread
Thus, I don't quite understand how you come to the conclusion that if a player is worth $X, he will go within 5-10 spots of that value. When did I ever say that a player *will* go where he should? That is the exact opposite of my point. My thesis was that ADP's do *not* reflect either the actual value of a player or even the value of a player that his typical projections would create.
W/R/T Sizemore, you were making the point that he went too early (and thus that he should have gone near his actual value) and I was arguing that was a silly reason to pick on the guy who took him, considering you could have argued the same thing for no less than half a dozen guys beforehand and more accurately at that.
BTW, I didn't mean to say the Bicycle-Friendly Berkeley Coalition had average draft position numbers. That should have said NFBC.
[ November 08, 2006, 09:05 PM: Message edited by: bjoak ]
W/R/T Sizemore, you were making the point that he went too early (and thus that he should have gone near his actual value) and I was arguing that was a silly reason to pick on the guy who took him, considering you could have argued the same thing for no less than half a dozen guys beforehand and more accurately at that.
BTW, I didn't mean to say the Bicycle-Friendly Berkeley Coalition had average draft position numbers. That should have said NFBC.
[ November 08, 2006, 09:05 PM: Message edited by: bjoak ]
Chance favors the prepared mind.
Sizemore and stat valuation thread
Originally posted by bjoak:
quote: Thus, I don't quite understand how you come to the conclusion that if a player is worth $X, he will go within 5-10 spots of that value. When did I ever say that a player *will* go where he should? That is the exact opposite of my point. My thesis was that ADP's do *not* reflect either the actual value of a player or even the value of a player that his typical projections would create.
W/R/T Sizemore, you were making the point that he went too early (and thus that he should have gone near his actual value) and I was arguing that was a silly reason to pick on the guy who took him, considering you could have argued the same thing for no less than half a dozen guys beforehand and more accurately at that.
BTW, I didn't mean to say the Bicycle-Friendly Berkeley Coalition had average draft position numbers. That should have said NFBC. [/QUOTE]ok, then I guess we're not disagreeing on anything there ... wasn't picking on the Sizemore guy, just giving an opinion - if I was picking on someone, I would've picked on the guy who took Abreu a round earlier but neither were HORRIBLE picks, its not like someone took Barry Bonds in the first rd
quote: Thus, I don't quite understand how you come to the conclusion that if a player is worth $X, he will go within 5-10 spots of that value. When did I ever say that a player *will* go where he should? That is the exact opposite of my point. My thesis was that ADP's do *not* reflect either the actual value of a player or even the value of a player that his typical projections would create.
W/R/T Sizemore, you were making the point that he went too early (and thus that he should have gone near his actual value) and I was arguing that was a silly reason to pick on the guy who took him, considering you could have argued the same thing for no less than half a dozen guys beforehand and more accurately at that.
BTW, I didn't mean to say the Bicycle-Friendly Berkeley Coalition had average draft position numbers. That should have said NFBC. [/QUOTE]ok, then I guess we're not disagreeing on anything there ... wasn't picking on the Sizemore guy, just giving an opinion - if I was picking on someone, I would've picked on the guy who took Abreu a round earlier but neither were HORRIBLE picks, its not like someone took Barry Bonds in the first rd