Happy Thanksgiving! Regularly Scheduled Articles Will Resume Monday, December 1
May 31, 2009
Prospectus Idol Entry
Strat-O-Matic For the People
Since you're reading this article at Baseball Prospectus, there's a good chance I know a little bit about you. You're passionate about baseball, to the point that some of your friends who don't share your enthusiasm may actually find it a little annoying (but enjoy your company nevertheless). You've invested a lot of time and energy learning more about the game and how it actually works, as opposed to how others seem to think it works, and consider that time well-spent. Thus you cringe whenever you're listening to a ballgame and the genial ex-player doing color commentary, who may have been a childhood idol of yours, describes some 8-year veteran with a career 726 OPS as a "professional hitter", or your favorite team's manager complains about his players clogging up the bases by taking too many walks. I've been there. Everyone here understands.
So what's brought you to this state of extreme fandom, where you dig through each morning's box scores with an addict's fervency and troll the interwebs in an endless search for ever stronger baseball insight? That much I can't claim to know about you, but for me and many others the gateway drug is easy to identify: Strat-O-Matic Baseball, or as it's known on the mean streets of Glen Head, Long Island, "Strat".
For those of you not familiar, Strat is a baseball simulation game utilizing the statistics of players and teams in past seasons. Each pitcher and hitter gets a "card" with three columns listing play outcomes. Games are played by creating a lineup for each team and then rolling three dice to determine the outcome of each plate appearance. Half of the time (depending on a dice roll) the results are read from the hitter's card; otherwise the pitcher's card is used. Obviously it's more complicated than that (speed, defensive prowess and positioning, ballpark factors, even weather can affect the outcome of plays), but it's the pure matchup of hitter vs. pitcher/defense-just as in real baseball-that functions as Strat's power plant and has seduced so many into testing their skill at being both manager and general manager of a fantasy ballclub.
Back in the day, when parents would buy a copy of Strat because little Jimmy had grown bored playing Payday and Stratego all winter, the cards and dice were always actual cards and dice and the games were meant to be played face-to-face. Nowadays Strat also has a computer version, which my tech-savvy children find comically simple but which does exactly what a lazy old poop like me who grew up playing other, more complicated simulation games wants a computer version to do: roll the dice and look up the results so I don't have to. Not only does this speed up game play, but it allows opponents to play online-making it easier to run leagues where geographically dispersed opponents can not only set up franchises to compete for seasonal titles but also play individual games against each other.
Unlike Strat, most forms of fantasy baseball don't involve playing individual games. Usually fantasy players draft a set of players and follow them through a given baseball season, hoping that the stats they accumulate are better than the stats accumulated by their opponents' players. Doing this definitely starts the process of casual fans digging deeper into the sport. When you're following your fantasy team players during the season, you tend to read every box score, not just the one for your favorite team. Keeper league owners start paying more attention to minor league prospects. Fans start watching, listening to and reading about more games and more players-which can only be a good thing. Moreover, fantasy baseball does a wonderful job of teaching player valuation and position scarcity-it becomes immediately apparent that an above average first baseman is usually less valuable than an above average middle infielder.
Strat (and other baseball simulation games), when played in a keeper league format, does this too. The same rules of supply and demand apply, as does the need to pay attention to minor league prospects. To Strat owners, today's morning box scores don't possess as much immediacy-unlike most other fantasy leagues in which yesterday's Curtis Granderson home run is essentially harvested and stored for later weighing, Strat players are more interested in the slower process of seeing what their players' cards might be like next year. Thus a Strat owner like me can spend a week at a remote fishing lodge in Western Ontario, without cell phone coverage or a daily newspaper, let alone internet access, and not feel as if they're at a huge competitive disadvantage.
But Strat goes further than other types of fantasy leagues in teaching about baseball, or perhaps more accurately in fueling the desire to learn more, due to: (a) the need to play actual games, which teaches things like lineup construction, bullpen management and the value of certain in-game strategies (something BP's own Joe Sheehan has occasionally trumpeted during his occasional jeremiads against bad managerial decision-making); and (b) the mathematical underpinnings of the Strat cards themselves which tempt owners into creating their own metrics to assign comparative values to cards based on both their offensive and defensive ratings.
So what are those "mathematical underpinnings"? A strat card, whether for a hitter or a pitcher, is divided into three columns-each one, based on a single die roll, with an equal chance of being used for a given plate appearance. Each column has eleven possible play results (e.g., a single or a flyout) numbered "2" through "12", which are determined by adding the pips from two rolled dice. If you're not a craps player and the odds of each two-die combination aren't intuitive to you, the chart below might help.
Die Results D1=1 D1=2 D1=3 D1=4 D1=5 D1=6 D2=1 2 3 4 5 6 7 D2=2 3 4 5 6 7 8 D2=3 4 5 6 7 8 9 D2=4 5 6 7 8 9 10 D2=5 6 7 8 9 10 11 D2=6 7 8 9 10 11 12
There are 36 possible combinations (or "chances" in Strat-speranto, the game's official language) per column, with six resulting in a "7", but only one resulting in a "12". Multiply those 36 chances per column by the 3 columns on a card, and you get 108 chances per card-of which 6 chances would be the "7" row in a given column. This can also be expressed as a rate -- 6 chances out of 108 is roughly .056.
The nice thing is that you can look at a given card, sum up the number of chances that result in a hit, walk or HBP, divide that by the 108 total chances on a card, and come up with a number that is analogous to that card's OBP. Perform similar work counting total base chances, and you come up with a Strat version of SLG. Add them together (or apply your favorite multiplier for the relative value of OBP to SLG) and presto! You've created a value statistic (I call mine Offensive Score) with which you can compare players.
But what about a player's defense? Strat accounts for this by placing "X-chance" results on the pitcher's card. These require another roll to determine the play's outcome. Each fielder is rated (as with defensive metrics in general, nothing sparks more controversy among Strat players than fielder ratings), and there's a separate chart that determines whether the play ends in an out, hit, error or possible double play, depending on the fielder's ratings and a dice roll. These results can also be expressed as "chances", and can be used to calculate the number of extra base runners (i.e., OBP) and total bases (i.e., SLG) that a given fielder might allow-resulting in what my personal Strat spreadsheet calls "Defensive OPS". Do some math to see how many times in an average game the player has their defense tested vs. how many times they come to bat, apply that factor to Defensive OPS and subtract it from the player's Offensive Score-and you now have a value stat that includes both offense and defense (I call mine Total Score).
Since Strat cards also have both offensive and defensive double-play chances, you can try to include them in your Total Score calculations by determining (1) how frequently the player will bat with a runner on first and less than two outs; and (2) how damaging the double play will be to your offense or beneficial to your defense. Trying to figure this one out is what first brought me to the Run Expectancy Matrix, among other advanced concepts. See how this might get a person to start doing their own research into advanced baseball metrics? Strat itself ranks numerous variables (e.g. pitcher hold ratings, hit-and-run ratings), and owners can include as many as they can reasonably model in their final value calculations.
This probably sounds like a lot of work, but that's what spreadsheets are for-and thankfully, the friendly folks at Strat-O-Matic will gladly sell you all the raw ingredients you need via a spreadsheet that lists every card's chances. For me, the highlight of each bleak midwestern January is the day my Strat spreadsheet for the upcoming season arrives. My family leaves me in quarantine while I toil away at the new season's data, happy as a pig in slop. Ain't nothin' like the taste of an equation made from scratch.
So if you're already a Strat-addict like me, I hope the game continues to stoke your interest in baseball metrics and that this article has given you some ideas. But if you haven't already tried our gateway drug, I highly recommend it-just be warned that it might only be the first step on a long, strange trip.
Many thanks to Margaret at Sturgeon Lake Lodge north of Silver Dollar, Ontario for making the submission of this article possible.