clock menu more-arrow no yes

Filed under:

roto corner: Q+A with art mcgee

New, 41 comments

today's game is on mlb.tv, beginning at 12:10 central time --- carpenter v glavine. if you're not an mlb.tv subscriber, you can also follow via gameday; liam provided the links in yesterday's thread. i've been unable so far to find a list of the pitchers who'll follow carpenter to the mound; anybody come up with that, please put it in the comments. Update [2007-3-1 9:43:58 by lboros]: gdowdy3 did, indeed, find that info: RHP Chris Carpenter (40 pitches). Followed by: RHP Blake Hawksworth (40), RHP Mike Parisi (30), RHP Andy Cavazos (30), LHP Tyler Johnson (30) and RHP Mark Worrell. [end update]

and while you're watching the likes of travis hanson and rick ankiel and mike parisi in today's game, check out erik's interview with chris costancio, who operates the minor-league site FirstInning.com. costancio has developed a PECOTA-like system for assessing minor leaguers' development arcs; it's a good read.

and, oh yeah --- if you haven't filed your line in the anthony reyes community projection, there's still time; we'll look at those results tomorrow.

for five years in the 1990s, i played in a rotisserie keeper league with an old friend named art. we've known each other since we were 9 years old or thereabouts. art was more or less the walt jocketty of the league --- he was extremely disciplined in his spending at each year's preseason auction and as a result almost never acquired the sexiest players in the talent pool. he'd let other teams overbid on the superstars and then grab less celebrated players at below-market rates. after the auction you'd look over his team (dubbed the Art Institutes), and the roster hardly ever impressed you; often looked pretty rag-taggy. but by midseason you were dumping your players to him for prospects, and at the end of the year he'd finished in the money and you hadn't.

i don't exaggerate at all when i say that art was practicing this approach long before jocketty. at the time i joined the league in 1993, art had already perfected his auction strategy (having developed it during the league's formative years in the late 1980s) --- and jock was an assistant gm with the rockies. by the time i left the league in 1998, jocketty was in his 3d year as the st louis gm, still just getting things figured out.

in 1997 art authored and self-published a book, How to Value Players for Rotisserie Baseball; i did some line editing of the text and offered a few pointers re printing and publishing. he produced 1,000 copies and sold out his inventory over a period of years; in time the book, to art's own surprise, became something of a cult classic, with hard-to-find copies selling for hundreds of dollars online. the book's reputation was such that art drew an invitation to participate in the experts-only NL Tout Wars, which sam walker wrote about in Fantasyland.

last month art released a revised and updated second edition of How to Value, this time published by rotoball guru ron shandler, who puts out the annual Baseball Forecaster and Graphical Player. you can check it out at shandler's site and, if'n you want, order a copy.

i asked art if he'd be interested in doing an interview, and we've known each other for too long for him to say no. so those of you preparing for auctions later this month, pay attention --- and learn something from the expert.

describe your career background as it relates to How to Value Players.
my day job for 20 years now has been trading in the financial markets based in chicago. chicago's really the global home of trading in options and futures, which means that a lot of the mathematical models developed for valuing these kinds of instruments were first put to work in the markets here in chicago. from the start of my career i have been exposed to that kind of thinking and that kind of mathematical modeling. when i started getting involved with rotisserie baseball, it was just a natural that i started applying that kind of thinking.

why'd you decide to write the book?
i think that i had tried to apply all these concepts in our league, and i just thought i had come up with some ways of analyzing things that i hadn't seen in print anywhere else. so, like a lot of people who, i guess, think they have got some interesting insights, i just wanted to get them out there and share them with other people. in the very narrow sense i was compromising my competitive advantage in my own league, but i guess my desire to get the ideas out there in the world outweighed any competitive advantage i was giving up.

we're talking about the mid-90s or thereabouts, correct?
the first edition came out right at the start of the 1997 season, so this 2d printing is somewhat of a 10th anniversary.

describe the distinction between this book and other roto books that are out there.
i think very little that's been written has been focused on these questions of how you translate player performance into the right price to pay for a player. the early rotisserie authors back in the early 90s --- alex patton and john benson were really the pioneers in writing about this, and then i came out with my book --- it's like the ideas got to a certain point and they weren't developed much further, and i felt the ideas could be taken a lot further and refined a lot more, and that was really still a lot of value to be gained by doing that. the traditional methods were sort of crude enough that there was still a lot of inaccuracy.

the way you phrase that is somewhat striking, because it's a favorite subject among sabermetricians and gms these days --- what's the "right" price point for a given player, based on his projected performance. they're asking the same question you did, but applying it in the real-life player market as opposed to the rotisserie market.
there are certainly a lot of analogies there, and i think that's part of the appeal of rotisserie --- to play out that fantasy of being a major league general manager. it doesn't translate entirely; there are some artificial constraints you have to deal with in rotisserie, one of the most obvious being that all the teams have the same fixed pool of money to start with. and you can't get good 2d- and 3d-year players at artificially low prices, the way real-life gms can.

the system you use is called marginal SGP pricing, where "SGP" stands for "standings-gain points."
that's right.

describe what that is.
under the rules of rotisserie baseball, every team has to spend at least a dollar on every player. so even if you only spend a dollar for every roster spot you have in the draft, you'd still end up with a certain amount of value on your roster --- value being defined as points in the standings, SGPs. the challenge you have when you go into the auction is figuring out how to allocate all the rest of your dollars, which i call your marginal dollars, to get as many marginal points as you can. that's the idea of marginal SGPs --- to figure out how many points above that baseline a player can be expected to produce for you, and then allocating dollars to players accordingly.

and how do you convert the basic roto stats into SGPs? a guy who you project to 30 hr and 100 rbi --- how many SGPs are those numbers worth? and how do you determine that?
if you're in what we can call a "standard" league --- and i almost hesitate to use that term anymore, because the ways that people play rotisserie have become so diverse; everybody's developed their own set of rules --- but if you look at a group of leagues that have a similar format to yours, just look at the standings in a given category over time and see what it typically takes to pick up another point. for example, in the home run category you'll find that in a standard league it typically takes about 9 and a half homers to move up one spot in home runs --- to gain one marginal point in that category. a guy who hits 30 home runs is probably worth about 3 points in home runs in a typical standard roto league. if you don't play in a standard league, you can still apply the same kind of thinking, but instead of using that 9.5-homer figure, you're going to have to look at leagues similar to yours and estimate how many homers, or steals, or saves it takes to gain 1 SGP in your format.

probably the easiest and most successful way for anybody who's played in the same league for a few years is just to look at their own league's past standings.

so you figure out how many marginal SGPs each guy is worth in each category, add it all up, and arrive at a bottom-line SGP figure, which translates into a bottom-line dollar figure. and that's what you bid in your roto auction.
right --- that's how you value the player when you go into your auction. you try to buy the guys who you can get for less than the values you've set, and avoid the guys who are going for more. i should make it clear, there are many other things you can incorporate into that bottom-line calculation besides what you think a guy is going to do this year.

what other factors are you considering?
one is position scarcity; i think that's a pretty familiar concept to people; my methodology just gives you a way to quantify it. another has to do with the volatility of a player. you take a guy who's got a certain set of expected stats and you can be pretty highly confident that he's going to produce those expected stats or something close to them. and then you have another guy who you project, at the median, to produce the same stat line, but that second guy has tremendous uncertainty --- he could potentially be much better or much worse. how do you factor that into your pricing? a lot of people find this counterintuitive, but i would argue that the second player, the guy with the more volatile projection, is worth more.

that's totally counterintuitive. most people would discount a player who has a significant risk of being a complete failure. i'm inferring here that the upside is worth more than the downside.
the upside is worth more than the downside, and in a nutshell that's because you can replace a guy on the downside. you don't really have to take the downside.

and then one other major factor that you can incorporate if you're in a keeper league is future value. you can think two or three years ahead, and so the book talks about how you can try to quantify that.

how did you and ron shandler get hooked up?
i think i actually just sent him a copy of the book. i asked him for his feedback on it, and he seemed to really like it. in fact, he adopted the marginal SGP approach for the dollar values on his site, BaseballHQ. and i would write occasional columns for BaseballHQ; i did that from 97 or 98 to 2003.

what's the reason for the update?
the first edition went out of print in early 03, and we had found in the last couple of years that there actually continued to be demand for the book. and believe it or not, if you searched for the book online just a few months ago, the only copies available were being offered at prices of hundreds of dollars. so that sort of convinced ron shandler and me that there's still interest in this material, and we saw the opportunity to update all of the examples and statistics in the book, and we had some material that hadn't been in the first edition, in the form of some essays i had written over the years for BaseballHQ. so that's what the second edition contains; it's an updated and expanded version of the original.

just humor me here for a second. ron shandler, your publisher, has consulted for the cardinals as a stat consultant. and shandler's very familiar with your work --- he's even adopted it at his site and incorporated it into his evaluations at BaseballHQ. these are all facts. when i watch jocketty exercise the discipline that he does in the market for players --- determining a value and sticking to it --- i feel as though he's implementing a system that's very similar to the one you've outlined in your book. and the first edition of the book appeared early in jocketty's tenure, in his 2d or 3d year as a gm. you can probably see where i'm headed: do you believe, or suspect, that there's any cause-and-effect relationship between the way jocketty operates and the book you released 10 years ago?
there's none that i know of.

none that you know of. do you feel as if there ever could have been advice passed to walt --- that ron shandler could have given jocketty a marginal SGP primer?
[laughs] i'd love to think so.

it's certainly fun to think about, isn't it? i'm probably taking this idea a little too far, but i'm simply struck by the fact that the conversion you were making 10 years ago with respect to roto ball --- converting a player's projected stats into a bottom-line dollar figure --- is the same one that today's leading sabermetric experts are doing with respect to real-life baseball. for example, mgl, mitchel lichtman --- who used to consult for the cardinals --- and his partner, tom tango, have really zeroed in on a dollars-per-marginal-win formula, which they use to gauge whether a given free-agent contract represents a wise investment or a waste of money. the Hardball Times and Baseball Prospectus also have similar formulas --- all of which postdate the SGP formulas you introduced in 1997.
i'm comfortable saying that these ideas were being published and applied in rotisserie baseball before i ever saw them applied to real-life major-league baseball. i don't know if you could put an exact timeline on it. it did seem to me that one followed the other. what the actual cause and effect was, i have no idea.

one could theorize that rotisserie baseball --- and perhaps this only happened indirectly --- might in some sense have profoundly influenced the way real baseball is now practiced. it's not such a stretch to suggest that these concepts first got into circulation among rotisserie players and then bubbled up to the sabermetricians --- many of whom, after all, are hardcore roto players --- and they, in turn, are now advising front offices.
well, now you've got me really intrigued. until i refocused on doing this new edition, i had not been paying that much attention to the sabermetric literature for a few years. i don't know; maybe it really is true.

did you read moneyball?
i did.

what was your impression of moneyball?
entertaining. actually, the number cruncher in me was really hoping for even more detail and insight as to how the oakland a's did it. but i think the thing that actually struck me most about the book was how long these ideas had been circulating. these ideas of trying to apply economic thinking more rigorously to getting better baseball results had been circulating since bill james. they'd been built on in a lot of different directions since then. it was really eye-opening to me to see how slow major-league baseball as an institution has been to embrace some of these ideas. i don't know how much of that is just michael lewis's version of the tale, but these ideas have been lying around for 20 years, some of them, and a much richer set of ideas for at least 10 years, and nobody's really been acting on them. and here's a guy who finally had the initiative and confidence to go against the grain and do it. it was fun and affirming to read about it, but it also made me wonder: what took so long?

even now, there's still a strong push back against sabermetrics. there are a lot of people who have a disparaging view of it even if it's incorporated in a partial way, as one component of an organization's personnel decisions.
as much as i enjoy it, and as much as i believe in all the things you can do with the numbers, i would never argue that it should be the sole, or even the biggest, part of how a team makes its decisions. but i certainly think it can be a valuable tool. and i think that any team that doesn't use that tool is putting itself at a significant disadvantage.

what about fantasyland, by sam walker of the wall street journal. did you read that one?
i did read that.

what'd you think of it?
well, i participated in nl tout wars in 98 and 99, and then i had to give up my spot because my son was born on march 22d, 2000, and it was the weekend of the draft. so i had to tell them that year that i couldn't do it.

i hope you grounded that kid for a few years.
hey, he didn't choose when to be conceived.

that was really inconsiderate of him.
i did rejoin the league again in 2003 ---- that's how long it took for another spot to open up. but the year i got back into the league was the same year the book went out of print, and i decided i was going to stop doing the columns for BaseballHQ. but it's a lot of fun. as an avid rotisserie player, it's a lot of fun.