THE BOOK--Playing The Percentages In Baseball

As an example, I’m going to shuffle cards, and this is how it would go:
“Careful”, pick 9 -> borderline
“Swing”, pick J -> Single

“Careful”, pick 9 -> borderline
“Swing”, pick A -> out (1st out)

“Careful”, pick 8 -> borderline
“Swing”, pick Q -> Single (runners 1B, 2b)

“Careful”, pick 4 -> borderline
“Swing”, pick 6 -> out (2nd out)

“Wild”, pick 4 -> borderline
“Take”, pick A - ball (1-0)

“Careful”, pick 9 -> borderline
“Take”, pick 7 -> strike (1-1)

“Wild”, pick 6 -> borderline
“Take”, pick 8 -> strike (1-2)

“Wild”, pick K -> down-the-middle swing
“Swing”, pick 2 -> out (3rd out)

We have a game theory class at the school and I emailed this to the teacher.  I’ll let you know if they give it a try.

The pitch-by-pitch was a little tedious for me.  Maybe once you have the rules all memorized it wouldn’t be so bad.  I’d probably prefer the PA to be determined in a card flip or two (at most).  Maybe with a special flip for attempting to take an extra base on a hit or steal a base.
vr, Xei

I’d recommend a scheme of some sort to physically move the cards to signify the count as the PA progresses - pull strikes towards the pitcher, balls towards the hitter, or something

Xei: right, that’s the choice to make, whether we want it a PA-by-PA, or a pitch-by-pitch.  Of course, nothing is stopping us from having a basic version (PA by PA) and an advanced version (pitch by pitch).

Yes, there’ll be other card flipping for taking the extra base, stealing, DP etc.  Right now, I’m mostly interested in the playability of pitch-by-pitch.

Greg: good idea about splitting the cards into ball/strike that way.

***

Also, at the end of every half-inning, the deck is reshuffled.  A typical half-inning has 16 pitches, so that would mean 32 flips of cards.  Presumably, a gamer is going to want to swing alot more than take, so, it might be more like 10 pitches in a half-inning, or 20 flips.

52 flips, 26 pitches, would mean something like 7 to 9 batters, so it should happen often enough that you have to reshuffle in the middle of the inning, especially when you need to have extra picks for baserunning.

***

Thanks for the comments…

Where is the strategy?

The cards are dealt randomly, and the opponent is the same each at bat.  No new information is added to prompt a strategy change.  Why not call out the same thing each time?

Consider this.  Both players draw six cards each at bat.  They look at their hands before they make their call.  The pitcher then draws his card from the batter’s hand, and the batter then draws his card from the pitcher’s hand.  The pitcher knows what is in his hand but not the batter’s and vice versa.  Rules are otherwise the same.

Now there is an element of strategy since the what each player call’s will be affected by which cards are in his hand.

At a glance, it seems that the best hitting strategy is to swing at borderline pitches with two strikes, but take with less than two strikes. 

Pitchers, on the other hand, should stay careful until the count’s 3-0, 3-1 or 3-2.  At that point, it depends on the situation whether to “go after” the batter and trade the possibility of an XBH for the increased likelihood of an out.

Ed/6: the focus should be STRICTLY on playability.  You and a friend have a deck of cards, and want to play say three innings of pitch-by-pitch baseball.  Is the pitch-by-pitch aspect good or not good?

Strategy will be a separate issue that can be addressed later.

Book/7: right, no need to worry about arbitrage at this point.  The idea is that I might then change the outcomes based on whether the batter is ahead or behind in the count, so that will go toward Ed/6’s objections.

Right now, just focus on playability.  Can you see yourself playing three innings of this, as opposed to playing say blackjack or hearts?

Personally, I think it’s too repetitive, even without dominant strategies taking away interesting choices.  The big thing here is that every pitcher/batter matchup is the same.  (Also, it’s impossible to strike out swinging.)

Ouch, yes, swing-and-miss I overlooked.  I don’t know how that happened.  I need to add that in, as well as 2-strike fouls.

Ok, then you seem to be saying that as a basic version, you will lose interest fast?  That in order for it to have value at the pitch-by-pitch level, then we need to have it tailored based on count.  So, for example, I need to have a different outcome set if the batter is ahead or behind the count.

That correct?

If so, if I add that level of complexity, does it become too complex for a card game?

That is, suppose that we want to create the best game we can out of a deck of cards.  Does this mean we’re going to be precluded from doing it at a pitch-by-pitch level because of the complexity?

And basically, we’re now forced to make it a PA-by-PA game?

This is a bit off-topic, but has anyone reading this had any experience with “Card Baseball,” invented by Sid Sackson in his “A Gamut of Games”?  I tried it a few times and wound up in serious pitchers’ duels, which might just be that I didn’t come up with the right strategies.

Sorry to hijack this thread.  Offline replies preferred.  Now, back to our regularly scheduled programming ...

Sounds very much on-topic.  I’d like to hear more.

OK, looking for links, because I don’t own the book ... here’s one:

http://www.boardgamegeek.com/boardgame/21682/card-baseball

There’s a link to a .pdf with the rules, transcribed from the book.

When I was in high school, I wrote a program for the TRS-80 to play it, but the computer played mostly randomly so it was easy to beat.  I played a few games in university, but neither of us got a lot of hits off each other.

It was still pretty good, and I’d love to try it again sometime ...

My strategy tip, as I remember it, is pretty obvious: keep an assortment of numbers and colors in your hand.

Phil: ooooh… I like that game.  I like the idea behind it, which has strategy implications of regular card games, as opposed to what I’m doing, which is really just some random-number generator.

I like it.  Seems a bit complicated, but, I like the whole thinking behind it.

First off, some people are making too much out of the details, e.g. “there’s no strategy” or “all the players are the same.” This sort of stuff can be added later, by having batters with different abilities, situational pitching adjustment etc.

I think what Tom is looking for is comment on the play mechanics. Which is really what you need to start with when designing a game. How long does it take? What does it feel like to play?

I dont think a pitch by pitch game is insurmountable. for one thing this sort of stuff could be programmed into a computer, god knows they are ubiquitous nowadays. So presumably a pc implementation would speed things up. Or perhaps Tom is sold on the idea of a board game? Well what do you say, Tom?

There have been some other attempts at baseball games like this where there is a guessing game pitcher vs batter. I think Avalon Hill did one like that, but in any event that can readily looked up. I am pretty certain there is at least one baseball game like that.

Also have you seen, I think it’s called finger baseball or something. It is sort of like rock ppaer scissors. You and I both “shoot” our fingers. 1 is a single, 2 a double, 3 a triple, 4 a HR and 5 a walk. If both the batter and pitcher match then the outcome is whatever the match was E.g. we both shoot 3s it’s a triple. If no match it’s an out.

Its based on psychology of course, but apparently there’s a real art to playing this. One of the mathematician/writers wrote about this. Here is a youtube of slightly more elaborate version:

http://www.youtube.com/watch?v=iSRmo5Nd22U

Again seems simple but this sort of game could be detailed by having different outcomes for various mismatched fingers e.g. 1 vs 2 becomes a stolenbase or whatever.

Again I would not try this as a board game, but with a computer you could do a lot.

I’m sure I have that Sid Sackson book around somewhere.  I’ve even played that game.  It’s addicting.

Tom, are there rules for what happens with men on base?  Baseball isn’t baseball without a double play, the ability to take (or not succeed at) the extra base, etc.

Bread:

Yes, there’ll be other card flipping for taking the extra base, stealing, DP etc.  Right now, I’m mostly interested in the playability of pitch-by-pitch.

I have uploaded rules in the link in Phil/13.  See the top of this thread.

Sunday/15 was marked for moderation and is now open.

Yes, Sunday has it right that I’m only focused on playability to start with, and the rest is details.

I don’t want to do a computer game, considering that Tippett has Diamond-Mind and there’s SBS Baseball, among MANY others.  There’s going to be little value-added.

The constraints are you have a deck of 52 cards, and you want to do this pitch-by-pitch.  Everything above that is a feature.

***

Now, I like the ideas presented so far, Phil’s link where each player is holding 6 cards, and he plays one of those cards.  It simply gets me completely out of the mindset I was in, which is adapting a computer game to a card game (Random number generator). 

As others have pointed out, this looks like it could be pretty tedious.  I agree.  I think it was a bad job on my part to have such a myopic view.

And, I like Sunday’s view, of both players playing a finger (or card) at the same time, and the value of the card will depend on the opponent.

This was exactly what I was getting at here:

a. If down-the-middle swing…
HR: Ace
2B: K
1B: Q, J, 10, 9
out: 2-8

b. If borderline and swing…
2B: K
1B: Q, J, 10
out: 2-9, Ace

Where the Ace is either a HR or out, depending on the strategy selection of the pitcher and batter.

***

These are exactly the kind of comments I am interested in, those that focus on the playability aspect.

It sounds like the best I can do is to try to improve on what’s already out there, in terms of verifying the percentages.

As Phil said, he seemed to play in alot of low-run games, and it may simply be because the percentages were off somewhere.

I like the pitch-by-pitch nature of the game, but I’d be more interested if at least some decisions had to be made concurrently (without having knowledge of some decision the other player has made).

Obviously the hitter can decide swing/take based on pitch location, but if you introduced something like “sitting dead red” that had to be decided before you knew if the pitcher was choosing “wild” or “careful”, that’d be neat.  “Sitting dead red” would give better results on pitches “down the middle”, but maybe you’d have to take more called strikes in exchange as the tradeoff.

Game play situations like this would then involve some game theory/reading of the opponent’s tendencies, instead of having complete information for every decision.

I’m not sure what the game mechanism would be for keep decisions secret until the appropriate time, though.  And more decisions makes things more complicated.  But if you have a good (combination of accurate and intuitive) model of the hitter/pitch matchup, “complicated” could still equal “fun”.

MLB Showdown is where it’s at. My friends and I played a 50 game season a couple years ago and we had a blast.

Obviously the hitter can decide swing/take based on pitch location, but if you introduced something like “sitting dead red” that had to be decided before you knew if the pitcher was choosing “wild” or “careful”, that’d be neat.

In one version I wrote, then discarded, I had the batter have a red and black wildcard, which he would select before the pitcher make his play (red=swing, black=take), facedown.  After the pitcher makes his call, he flips over his wild card. 

***

MLB Showdown: can you explain the rules?  Wikipedia is blocked at the office.

Well one way to speed it up would be to have players use their fingers for the pitches.

So pitcher throws his “2” which is a curve, the batter signals “3” which is “make contact” swing and we get a miss.....

Then the cards could come into play ONLY when contact is made. So if contact was made on fingers 3-2 this would be a HR if an ace, but its fly out if a K.

So you would have several inputs here: the finger combo as well as the card, and you cross refernence it.

I dont know if Tom is wedded to the idea that the idea of having to play cards for the pitches, this will make the game take longer, whereas fingers can go real fast. Then you only flip the cards a few times each inning

What do you think?

I like the idea of using fingers.  You can even try use say rock/paper/scissors instead of 1/2/3.

All good ideas.

I agree that a simultaneous decision is needed by batter and pitcher to make it interesting. Otherwise the correct strategy will be worked out in a few weeks.  Of course there is a game strategy optimum ratio of swing/take wild/safe but with 12 counts and possible effects of base-out state it would be less easy to work out.

I would play a game something like this

pitcher declares wild or careful
batter declares swing or take
(you could use different poker chips held in your hand like in some types of hi-lo poker)

then you draw a card to see where the pitch ended up
safe -
jqka is on the edge of strike zone
3-10 is down the middle
2 is unhittable ball

wild
jqka is on the edge of strike zone
3-10 is unhittable ball
2 is down the middle
(the proportions are just an illustration)

if the batter takes -
down the middle and on the edge are strikes rest balls

if the batter swings - draw another card to see what happens

unhittable ball
ace= checked swing = ball
2-k swing and miss

down the middle
range of outcomes which overall favours the batter lots oh hr 2b and singles but with afew swing and misses and also field outs foul balls

on the edge
range of outcomes which overall favours the pitcher e.g swing and miss pop up flies ground outs but a few hits.

Can pitch fx tell us what happens to balls hit form the edge of the strike zone vs those hit in th emiddle (I am assuming it is a big difference)

Obviously baserunners complicate/make the game more interseting. YOu could add pitchout to the pitcher’s options which would reduce the success of base stealing etc. based on the game theory section of the book.

You could also have three options wild neutral conservative and three batter options swing for the fences swing normal or take pitch.

I think though the future of this is online including team play where players on a team take turns at bat and pitch one inning each in rotation.

I play ootpx and FOF and while online leagues are available they are all based on games being simulated there is no head to head in-game strategy game available (although I have heard people mention this on their wish list). Online would also allow more options and more closely model real life if not constrained to the 13 denominations in a deck.

James

I like the idea of throwing the fingers out: 1=FB, 2=Curve / Breaking, 3=Changeup

And the batter tries to guess as well.

I want to set it up so that the proportion of the pitches matches what really happens.  So, something like 60/30/10.

If you throw changeup and batter guesses changeup, there’s a huge payoff.  If you throw FB and batter guess FB, the payoff is not that big.

If you throw FB and batter guesses changeup, then pitcher gets a huge advantage.

So, for example, if let’s say there’s a match, then it’s an automatic swing.  But if let’s say batter guesses change and pitcher throws FB, it’s an automatic take.  And if you take on a FB, there’s a huge proportion that it will be strike.  So, that’s the incentive for the batter trying to guess fastball.

If you need numbers to think about, presume that half pitches are takes, half are swings.

I tried a follow up post a couple of weeks ago but I dunno what happened, this inteface is not the best.

Anyhow, I was wondering if tom wanted to make this more of a baseball simulation or just something that is more abstract? If the former then I was thinking the outcome matrix could be based on that recent discovery, that BABIP is constant or at least approx. constant for all batters

You could use this information to constuct a matrix where everytime contact is made the ball is an out .70 of the tiem and a hit .30. This would certainly seem realistic based on what we know statistically. Then from there one could adjust it for various batters/pitchers. e.g. Cecil Fielders .30 would be distributed more for power and Mario Mendoza’s would be mostly singles. Same for pitchers with Nolan Ryan giving up more HR balls, etc.

What direction were you planning to go with this Tom?

ANyhow if this doesnt post I give up.

I played a very cool simulated baseball game that was quite advanced that was, at one point, based on playing with a deck of cards.  It wasn’t just the rank, but also at times the suit, so all 52 cards could come into play.  It also had different player cards, a la strat-o-matic complete with aging curves and what not.

The league was also populated with a bunch of stat freaks (back in the late 90s and early 2000s) so we created EQA, WPA, SNVA, etc. for all of our fictional players.

Unfortunately the league stopped and the site got taken down but you can see the information using the ever useful internet archive wayback machine.  The history link talks about the cards at web.archive.org/web/20020811175525/www.bantha.org/~develin/baseball/history.html and some of the basic rules and franchaise histories (sort of like a mini-baseball reference) are linked from web.archive.org/web/20020929085557/www.bantha.org/~develin/baseball/index.html.

A bunch of us talk about restarting the league every so often, but the main problem is we need an AI or else the demand on the commish to play the games is too great.

At the bottom of the main blog post, I added a link to a card game that reader James created that was based on the game we’re discussing.

Thoughts?