Wanted: Review of "How to Value Players for Rotisserie Baseball"

[stevusmagnus]

Has anyone read this book by Art McGee?

It sounds intriguing but I want to know if it is any good and a description of some useful things I may find in there. Does it teach you how to rank your own draft cheat sheet? Why is this advantageous over BBHQ's rankings? Is it valuable for drafts and head to head leagues? Would you recommend this book to someone who has never made personal evaluations of players before? Thanks!

Steve

[Apple Pie]

Yes, I've read the book. Several times.

No, it does not teach you to rank your own draft day cheat sheet. Not exactly.What it DOES do, as the title suggests, is teach you a method to assign value to players. You can use that to build a cheat sheet. It's a little heavy on math, going through the definition and calculation for SGPs -- Standings Gain Points. You need to be comfortable in Excel.

The advantage over HQ or other lists is that this one will by YOURS. It'll be based on your league parameters, your SGP denominators, etc. The advantage is in really understanding what you're buying, not just bidding to an amount on a list.

Of course I recommend this book to anyone who wants to make valuations on players. Again, there's some math and you need to be comfortable in Excel.

[Old Tom]

Again, there's some math and you need to be comfortable in Excel.

:Slams money down on kitchen counter:

I'm out!

[NICK@HQ]

I've read it. It's like having a detailed manual on how to rebuild your car's engine. Do you need to know this in order to drive a car? Nope. But if you do know how to build an engine, not much will happen to your car that will be a mystery to you.

[bonehead]

I just got mine - thanks HQ for getting this back in print.

While a lot of it is over my head, a couple of things stand out, and I think it would be great to get some perspective on it from some of you valuation junkies that cook your own, like Jayhawk, Merv, Old Tom Andy A., etc...


My observations/surprises/epiphanies/puzzlements:

1. With the SGP method, THERE IS NO FIDDLING WITH THE Hitter/Pitcher SPLIT! (that really surprised me!) Unlike the SCARCITY method, where you have to make a decision in the valuation model on how much $$ to allocate to hitting vs. pitching, the SGP method does it automatically. Merv, is that how Rotolab works when the SGP method is selected?

2. I was also surprised to see in the methodology that ratio stats use the LEAGUE AVERAGE as a baseline stat - and any player that generates a ratio stat BELOW the league average earns NEGATIVE sgp's. I've never fully bought into that method, and always preferred comparing a player to the 'replacement level' stat, rather than the 'league average' stat...

...and here's why: imagine a league that is constructed with ONLY ratio stats. If every player was evaluated against the league average, 1/2 the players taken *could* have negative SGP's - leading to a negative value, - yet certainly, a collection of those players could certainly allow a team to end up higher in the standings than a team made up entirely of "replacement level" players...which would imply to me that they would have both positive SGP's, and positive value...

of course, I doubt anyone plays in a league like that...

Thoughts?

[pate]

THERE IS NO FIDDLING WITH THE Hitter/Pitcher SPLIT! (that really surprised me!) Unlike the SCARCITY method, where you have to make a decision in the valuation model on how much $$ to allocate to hitting vs. pitching, the SGP method does it automatically. Merv, is that how Rotolab works when the SGP method is selected?

I could have made it work that way, but I chose to use the split. If the SGPs are split say 66 /34 but your league has a lot of owners that really like low-end $1 pitchers and invest the money in Hitting, you better have a way to adjust for that. Not every group of 12 owners will bid the same given the same set of rules and league paramaters.

-Merv

[Randall@HQ]

...and here's why: imagine a league that is constructed with ONLY ratio stats. If every player was evaluated against the league average, 1/2 the players taken *could* have negative SGP's - leading to a negative value, - yet certainly, a collection of those players could certainly allow a team to end up higher in the standings than a team made up entirely of "replacement level" players...which would imply to me that they would have both positive SGP's, and positive value...

My understanding (which may not be correct) is the following --

When you are calculating SGPs at the beginning of the season, you're using previous season's stats. When you look at the actual standings at the end of the season, you're looking at different stats and different SGPs.

Suppose based on prior years, league average ERA = 4.50. So a pitcher projected for 4.75 ERA has negative SGP projected for this season.

But, after we play this year, it turns out that league average ERA = 5.00. So the same pitcher with 4.75 ERA really has positive SGP for this season.

[pate]

2. I was also surprised to see in the methodology that ratio stats use the LEAGUE AVERAGE as a baseline stat - and any player that generates a ratio stat BELOW the league average earns NEGATIVE sgp's. I've never fully bought into that method, and always preferred comparing a player to the 'replacement level' stat, rather than the 'league average' stat...

I think in one of Patton's early books he mentions this, about where the base line is chosen in ratio based stats. Some professor from North Carolina (I think) told him that it was incorrect to use the league average and that the bottom of the category should be used (paraphrasing). I'll try to dig up the section later tonight. After studying different value systems for close 20 years and coding them for over 15, I have come to the conclusion that there isn't one right way to do it. Or rather, I'm not convinced a particular method produces the best values for all owners. And I am even more convinced that there will never be a consensus reached.

-Merv

[usualsuspects]

The hard thing about the ratio categories is that not only does the baseline change, but the proportions change. For instance if you set a .250 baseline and are valuing a .280 hitter with 500 AB and a .300 hitter with 500 AB, there will one delta between their AVG SGP's. But if you set the baseline at .270, not only do their values change, but their relative values also change, which is partly what makes it such a mess.

As for the book, don't ask me why but in a crazed sleep deprived moment I ordered the new edition even though I had the old one. You do NOT need to do that-unless you are related to Nick Markakis who's in many of the examples. Anyhow, it is worth reading although there might be room to argue some of the discussion, such as position scarcity and valuing the ratio categories.

[GoldenEagle]

As for the book, don't ask me why but in a crazed sleep deprived moment I ordered the new edition even though I had the old one.

You must have been really sleep-deprived to not have put the old one on ebay when the prices were like $1,750.

[teddyballgame]

and did any of those babies sell at $1750?

I just about fell off my chair when I saw those $1000 price tags on amazon for the old edition.

I figured Shakespeare must have had a hand in those!

[Apple Pie]

You must have been really sleep-deprived to not have put the old one on ebay when the prices were like $1,750.

I sent inquires to a couple rare book sellers during that time but didn't get any bites. Supposed I could've used a $30 reserve on eBay or something to protect myself. Anyway, I suspect those were advertised prices, not actual.

[pate]

I was also surprised to see in the methodology that ratio stats use the LEAGUE AVERAGE as a baseline stat - and any player that generates a ratio stat BELOW the league average earns NEGATIVE sgp's. I've never fully bought into that method, and always preferred comparing a player to the 'replacement level' stat, rather than the 'league average' stat...

From the 1991 Patton book, Patton commenting on a letter from a N.C. professor:

"This is a rebuttal to your basic claim that 'hitters overall have to be of no value overall in batting average.' " A team comprised of average hitters, he points out, does not get zero in batting average. "Such a team in the AL would get 6.5 points in batting average... The 'zero point' you've chosen (.264) is not worth zero. (if it were, such teams would come in last in the BA category)... I think you have chosen the wrong average. The average BA of the last place team is .2494... BA value should be based on this as the zero point, because a hitter only has value to the extent he helps pull a team out of last place.

I REALLY miss Patton's writing style in his early books, still have all of them on my shelf for reference.

-Merv

[usualsuspects]

You must have been really sleep-deprived to not have put the old one on ebay when the prices were like $1,750.

In fact, when I saw those prices in the $400's, I DID put it on Amazon or eBay, forget which, but no nibbles at $50........

[bonehead]

Merv - thanks for digging out that reference.

Anyway - back to the thread - the parts of the book that I could understand, I enjoyed. Creating values based on SGP's turns out to be MUCH easier than with the scarcity/PVM method...and the added bonus, I found out, is that it is MUCH easier to adjust values for POSITIONAL SCARCITY with the SGP method - it's a nightmare with PVM! And, as a note, McGee is a proponent of adjusting the value of the catchers (and all positions, for that matter) to account for the positional scarcity....

Merv is right - nearly 30 years of fantasy baseball and there is no definitive 'right' way to value players.

Unless, of course, you use RotoLab!!!