Baseball!

[The Brain]

Cool. I'm not into baseball but nice reading about different games.

[The Minsky Moment]

As a kid in the early 1980s, my knowledge of the game of baseball came from watching and listening to Phil Rizzuto call the Yankee games on channel 11 and what managers would say the press.  I knew the pitching was 90% of baseball, that winning was about making productive outs and executing the little things like bunts and hit and runs.  Hitters should be aggressive and swing early the count. As for pitchers, well Bert Blyleven might strike out lots of guys but his won-loss records were crappy - the best pitchers might allow some runs here or there but they could figure out a way to win.  Pete Vuckovich was the best pitcher in 82 because he had the best won-loss record.  These things were self-evidently true and every astute baseball man knew them from long experience.

In 1982, Bill James Baseball Abstract was published to the mass market for the first time; the next year Pete Palmer published the Hidden Game of Baseball.  These books were not like any baseball book I had ever read. The numbers and methods they used were totally unfamiliar.  To be specific, they *had* methods.  They didn't seem to care about the authority of experienced "baseball men".  James in particular seemed perfectly content to state flatly that these conventional stories were wrong. 

(although Vuckovich remains the greatest Serbian-American pitcher, edging out Steve Sundra).

These two books started a trickle and then a flood of analytical work in baseball.  It took a while for there to be any effect on the game; for over a decade there was no effect at all.  Sadly progress often does happen one retirement (or death) at a time.  40 years later, however, the impact is huge.  MLB teams now spend tens of millions of dollars a year on "analytics" - and all that work traces back to James and Palmer.

The big dice and card games - APBA and Strat - were designed pre-analytics.  In many ways they were ahead of their times - in making such a game, you have to think analytically about how hitting, defense and pitching interact.  Strat cards didn't report on base percentages, but the cards themselves showed walks as a positive offensive event.  Strat also made some adjustments to the new analytical work when it came out, incorporating new generation fielding stats in their fielding ratings - a move that caused some waves when they downgraded the rating of Derek Jeter in one of his gold glove award years.

That said, newer game designs have been able to take more creative advantage of the analytics revolution.  Two aspects in particular have influenced game design:

1) Retrosheet - an open source project to record play-by-play accounts of every game in major league baseball history; started in the late 80s. There is now data for the majority of games going back to the late deadball era.  Retrosheet data allows more precise information about things like batted ball locations, defensive performance, pitcher ability to control runners, etc.  T Retrosheet supplies a lot of the source data for sites like baseball-reference or fangraphs.

2) Defensive Independent Pitching Statistics "DIPS" - arising out of articles by Voros McCracken around 1999-2002.  It is now accepted that most (though not all) pitchers have little influence on controlling what happens to batted balls into the field of play.  Statistically there is little or no correlation year-to-year on pitcher's ability to suppress base hits on balls put into play.  Thus for most pitchers, their success depends on controlling walks, strikeouts, and homeruns, and hoping they have a strong defense behind them and a good dose of "hit luck."

There are several new game designs that use these concepts.  I've played 3 of them and will discuss later.

[The Minsky Moment]

My favorite card and dice game at the moment is Inside Pitch Baseball, which is based on the DIPS theory described in the post above.  Each play always starts with the pitcher card.  The card contains a 6-by-6 matrix and uses 2 six sided dice (i.e. like APBA but with the results plotted on a matrix).  The main results can be: "K" (chance for a strike out); "W" (chance for a walk), "HR" (chance for homerun), or a blank.  If there is a chance for an event, you roll a d20 check against the batter tendency for that event; each batter is rated 1-20 for each event depending on their tendency to strike out, walk, or hit homers, with 10 being the average.  Elite strikeout pitchers can get "K+" results that add 10 to the batter check - an automatic strikeout for most hitters. 

If the result is blank then the batter rolls on his 6x6 matrix which reflects the batters performance on balls put into play. Every hit and out is given a location based on the batter's tendency - this comes data comes from retrosheet.  In addition, most pitchers have a least one automatic out, to reflect their own out tendencies - e.g. groundball pitchers vs flyball.  Finally, each pitcher card has 6 out of 36 chances to generate a fielding play check and 3 chances for a ballpark check.  If it a ball park check then instead of going to a batter card you go to the unique card for that ballpark, which reflects its tendencies to affect different kinds of hits and outs.

Here's how it looks for 2 effective but very different pitchers:


Both pitchers had good years but Koufax's card is filled with more than a dozen strikeouts, and most of them are K+s.  Hawkins only has 5 "Ks" (plus one on righties) and none of them are pluses. Hawkins has far more blanks which means he will generate far more balls in play but with a good defense behind him an playing in a pitchers park he should be able to scatter small hits without much damage most games.

The system handles left-right platoon advantage elegantly - some cells on pitcher and batter cards have "/" results with the result on the left side for left handed opponents and on the right side for right handed opponents.  That makes it very easy to identify the impact of a player's left-right split.

More to come . . .

[The Minsky Moment]

Inside Pitch, continued

Although it may seem odd at first, the elegance of the interaction between batter, pitcher and ballpark in Inside Pitch is really impressive.  As an example take these two cards from the 1977 American League:



Red Sox fans are gritting their teeth looking at this one.  Rice, who hit 39 homers in the season has a homerun ("HR") rating of 19 against right-handers and only 9 against lefties.  Zisk, who hit 30 homeruns in real life, is even better than Rice  against right-handers (20) and much better against lefties.  What gives?  It doesn't seem to add up.

When you look at the real life season split data, however, it comes into focus.  First, Rice hit 34 homers against righties but only 5 against lefties (albeit in only 147 ABs).  Zisk OTOH had a strong home run rate against both sides.  More importantly for this example, Rice hit 27 homers at home in Fenway Park but only 12 away; whereas for Zisk it is reversed: 18 road HRs, 12 at home.  The way Inside Pitch handles this is that the Fenway Park card generates a ton of homerun chances for Rice to take advantage of in home games - 10 to be exact -  whereas the Comiskey Park card generates only 3 such chances for Zisk at home.  Play out a full season with both players at home for half their games and Rice should outpace Zisk despite Zisk's higher ratings.

For a concrete example of how the game handles park effects, here is this example from the 1962 season:



The old Polo Grounds is a home run heaven, with 9+ HR chances, but 1962 Chavez Ravine is where power goes to die.  In addition to having no HR chances, the stadium modifies all batter homerun check rolls universally by -1.

One virtue of this system is that it effectively handles extreme tendencies that cause 50-50 systems like Strat-o-matic fits. Nellie Fox, for example, has strikeout ratings of 1/2 - on a "K" result he will only strike out 5/10% of the time. Even on a "K+" result, he will wriggle out almost half the time.  Similarly, ultra control pitchers will simply generate very few walk checks.

Another complaint often raised with 50/50 games like strat is the situation where a player brings in a shut down closer at a key moment, but then rolls 3 straight times on the batter card and gets hit hard. That may be "realistic" in the sense that even elite closers sometimes get lit up by elite hitters - think Ortiz vs Rivera or Bonds vs Gagne. But even if it passes the simulation test, it fails as a *game*: the scenario makes it seems like the player's decision had no in-game effect.

Inside Pitch doesn't have that problem: if you bring in your elite closer, he will get a roll on his card.  It doesn't guarantee a good outcome and the batter's check may counteract it, but perception of efficacy is present.

OK  - what are the cons:
1) There are usually at least 2 die rolls per play and sometimes more. On the other hand, the checks are usually pretty simple and can be found right on the cards; the use of outside chart checks is limited.
2) Although use of outside chart checks is limited, the exception is a rather convoluted set of baserunner advancement rules.
3) Although ballparks are meticulously rated for offensive and defensive events, they don't have left-right splits - an issue for parks like Fenway, Yankee Stadium or the old Polo Grounds with very distinctive differences between right and left dimensions.  Such splits are implicit in the ratings of home players - see the Rice example above - but that doesn't work for visiting players.
4) The game doesn't calculate left-right splits for players unless they had at least 40 plate appearances each side.  Although well-intentioned, that has the perverse effect of not accounting for the platoon splits of full dedicated platoon players like John Lowenstein, who would play full time against righties but never against lefties. In game, he will thus be allow to crush it from both sides.

These are mostly quibbles but a bigger issue is that the game isn't that compelling as a head-to-head experience.  The game system handles strategy events like base stealing and bunting in a fully automated fashion; although its possible to run manually, it doesn't have the same feel as other games more optimized for head-to-head play.  And although results are stated on card, the D8 results don't have the same visual impact as the DOUBLE (cf) in Strat.  From the developer's point of view, I suspect this shortcoming isn't a big issue; from the game forums it is clear that his target audience is heavily oriented towards season replayers who replay parts or entire past seasons solitaire. 

In that regard, the game is also available in computer form.  Unlike many other such ports, which i usually find less than inspiring, the Inside Pitch computer version genuinely re-creates the entire board game experience in electronic form: you have the game cards in front of you on screen, roll electronic dice on screen, and then you locate and click the result on the cards. You are interacting with all the board game parts in the same way as you would on paper.  The computer just takes care of the bookkeeping and the pesky runner advancement rules for you.  So paradoxically, even though I really like ISG as a board game design, I find myself playing it in computer format rather than with cards and dice.

[jimmy olsen]

I had a baseball game on the Sega Saturn that I loved.

I think I played whole seasons with the 96 (I think) red sox and giants. Clemens won like 25 games and struck out way over 300 batters. Bonds must have hit like 80 home runs. 

[The Minsky Moment]

Those old console games could be a lot of fun, but precise statistical realism was not really the point.

[The Minsky Moment]

Payoff Pitch

Payoff Pitch is like a lighter version of Inside Pitch.  Both are based on DiPS, both start with rolls on a pitcher card with results reflecting the pitcher's ability to generate strike outs, bases on balls, and homeruns.  But Payoff Pitch does it in a way that is simpler and more evocative.  In the place of the 6X6 matrix of Inside Pitch containing specific check results, Inside Pitch uses a 2d6 roll on a simple 2-12 table with descriptions of the "payoff pitch" that concludes the plate appearance.  TOUGH - corresponds to strikeout ability; PATIENCE - the pitcher's tendency to yield walks; WHEELHOUSE - tendency to allow homeruns; IN PLAY - tendency to yield balls is play.  In addition, there can be checks for fielders ("DEFENSE") and the ballpark's tendency to boost homerun power ("BALLPARK"). For example:



The row of "Tough" results in Mike Scott's miracle 1986 year reflects his very high strikeout rate that year,.  Bob Lemon, on the other hand, tended to yield a lot more balls hit into the field of play, but since those end up as outs about 70% of the time (and mostly singles the rest of time) and since he also didn't give up that many walks or homeruns, he should perform well despite the fewer punchouts.  Neither yields a lot of homeruns as shown by the fact that Lemon has only one low probability WHEELHOUSE result at 12 and Scott has none (Note that BALLPARK checks have some probability of ending up as a WHEELHOUSE).

Similar to Inside Pitch, play then proceeds to a check vs. the batter's card using percentile dice.  Each batter has a separate set of results for each category of the pitcher's payoff.  "TOUGH" checks against the batter tendency to strike out, but even if the K is avoided it very rarely results in a hit.  PATIENCE - checks against the batter tendency to walk; WHEELHOUSE - reflects the batter's home run power (but will often result in another kind of hit for a slap-style hitter); IN PLAY represents the hitter's batting average on balls hit in play.  For some seasons left-right splits are available:



Here Reese's lack of power is evident by the fact that he has only a 19% chance to hit one out even on a Wheelhouse result.  On the other hand he draws lots of walks (more than 50% on a patience check), has a good average on balls in play (.320) and on a "Tough" result only strikes out 1/3 of the time and even has a 20% chance of eking out a hit.  Platoon slugger Lowenstein OTOH has nearly a 50-50 chance of a homerun on Wheelhouse vs righties and also walks a lot, but hits only .250 on balls in play and does a lot worse than Reese on TOUGH results (51% strikeout chance and only 9% chance of a hit).  The card also clearly shows his uselessness against left-handed pitching. Unlike Inside Pitch, which only shows platoon splits for players with at least 40 at bats from each side, Payoff Pitch does splits for all players for the seasons it does splits for, which is good for dealing with full platoon players like Lowenstein and teammate Pat Kelley.

Overall, Payoff Pitch doesn't have all the bells and whistles of Inside Pitch: it is less precise in terms of pitcher results, lacks the more accurate hit and out location data, deals with park effects more crudely, and simplifies fielding, among other things.  But it is accurate enough for most purposes and the tradeoff is a game that is usually more quick and easy to play live.  The baserunning rules are simpler then IP and (like Inside Pitch) outside chart lookups are minimized. For that reason and because I really like the look and feel of the pitcher cards, I find it usually more preferable for head-to-head play.

It is also well priced.  Payoff Pitch, like Inside Pitch and most of the newer generation games, sells player seasons in PDF as well as printed form and their PDF pricing for complete seasons is very competitive (usually $9 a season).  If you choose the printed option, you pay a lot more ($34-49) but you do get nice professional printed color cards on good quality stock.

[Syt]

Thanks again for this, Minsky.

I'm unlikely to try these games myself, but it made me buy a subscription to MLB.tv again, and I may dabble with OOTP again

I'm sure the Bucs won't let me down this year.

[Valmy]

Well it depends on what you expect the Bucs to do 

[Syt]

Well it depends on what you expect the Bucs to do 

As long as they do better than the O's

(ok, that was mean, I'm sorry. )

[Valmy]

Quite a high bar there 

[The Minsky Moment]

I may dabble with OOTP again

I've got my "Perfect Team" up and running but I'm firmly in the second division ATM.

[Syt]

I may dabble with OOTP again

I've got my "Perfect Team" up and running but I'm firmly in the second division ATM.

I did that a bit when it was new, but I'm not a fan of blind pack based competitive games.

[The Minsky Moment]

I did that a bit when it was new, but I'm not a fan of blind pack based competitive games.

I've done it on and off, the key I think is not caring that much how you do but just having fun with it.  If I stink it up in the rookie leagues, who really cares?  Also the packs only matter at the very beginning; after that I just interact with the auction economy.

[The Minsky Moment]

The real problem with PT is what's happening today - I have Bo Bichette Live on my perfect team but he is playing the Yankees in real life.