keyboard_arrow_uptop

There are records that are very important and records that aren't important at all. Within each category are records that get lots of attention and records that get very little attention. Thusly:

Records that… Got lots of attention Get no attention
Matter Barry Bonds hit 73 home runs in a one season! That's the most ever! More than Babe Ruth! Mike Trout has more bWAR through age 22 than any player in history.
Don't matter The Padres have never had a player hit for the cycle! Somebody turned an unassisted triple play! Edwin Jackson is the only pitcher to produce a 4.2/9/6/6/4/4 pitching line in a start

And within the space in between the columns are the records we could argue about. Webb/Albers, for instance, used to exist firmly in the bottom-right quadrant, but now it lives in all four quadrants simultaneously, you might say. So this is a look at one record, with the attempt at deciding which quadrant it goes in.

The record was set—tied, technically—on August 5th, 2007. The Mets were facing the Cubs, who were starting Jason Marquis for a sold-out Sunday afternoon game. The Mets managed to put runners on base in every possible base/out state: Runner on first with none out, runner on first with one out, runner on first and second with none out, and so on. There are 24 base/out states, and the Mets hit them all.

How common is this? Less common than you'd think, or more common. I have no idea how common you think it is? I'd have guessed it happens literally never or twice a year, which is how often the vast majority of possible events occur, so the actual frequency—about once every 15 years—puts it in that rarity sweetspot where its occurrence and non-occurrence are both unpredictable.

The Mets were the most recent team to do it. How'd it happen? Here are the innings in which they first achieved each base/out state:

000 100 020 003 120 103 023 123
0 outs 1 5 6 9 8 8 8 8
1 out 1 6 1 6 6 9 6 6
2 outs 2 2 2 1 2 1 6 6

It didn't need to be this way, but as it turns out that table aligns nicely with this next one, which shows how frequently each base/out state usually happens:

000 100 020 003 120 103 023 123
0 outs 24.27% 5.49% 1.84% 0.27% 1.41% 0.51% 0.35% 0.39%
1 out 17.35% 6.35% 3.04% 1.04% 2.51% 1.11% 0.87% 0.96%
2 out 13.77% 6.19% 3.89% 1.59% 3.16% 1.45% 1.00% 1.16%

The Mets, in other words, got all the easy ones out of the way early, then knocked off the more difficult states as they got more chances. The most common state–aside from the one that we start each inning with–is roughly 65 times more likely to happen than the least likely, which most often occurs by way of leadoff triple. Here's the same table as above, except frequency (by percentage) is changed to frequency (by rank):

000 100 020 003 120 103 023 123
0 outs 1 6 11 24 14 21 23 22
1 out 2 4 9 17 10 16 20 19
2 outs 3 5 7 12 8 13 18 15

The 10 rarest events all involve runners on third base, and the four rarest involve runners on third base with nobody out. What makes it so hard to hit all 24 base/out states in a game is that, while some of these events are likely to cluster up with other events–"runner on second with none out" is likely to cluster with "runner on third with one out"; and so on–some are very unlikely to cluster with other events. Having a runner on third with nobody out presents a nearly impossible chance of next having runners on second and third with none out, for instance. Once you get one of the four rarest states–runner on third, nobody out–it's actually pretty difficult to fill up the rest of the runner on third, nobody out states.

Yet, as we see in the first chart, the rare events–the ones angling down the top/right corner–happened in clusters. How'd they manage this?

Take the eighth. The Mets put their first two batters on, first and second with nobody out–a semi-uncommon state, but easy enough. Then the lead runner stole third, while the trailing runner stayed put. This is not an expected outcome. Major-league teams were in this starting point about 2,500 times last year, and the steal-of-third, no-steal-of-second play happened in the neighborhood of eight times.

The next batter doubled, scoring the runner on third but, because he had stayed at first two pitches earlier, not the trailing runner. That punched 023/none out on our card. Next batter walked intentionally–about 13 percent of batters will walk or get hit with runners on second and third with nobody out, and we had our bases loaded, none out scenario.

Similarly, the sixth turned on one unexpected advancement. The Mets put two on with one out, then both runners stole. Second and third, one out (punch); which led to an intentional walk, loading the bases with one out (punch); followed by a groundball to second base. Now, there were 1,500 bases loaded, one out situations last year. About 1,000 led to balls put in play, about 500 of which were groundballs, about 350 of which were fielded by infielders. But in the large majority of such cases, the infielders had time to turn a double play or get a force at second or home. Only 48 of those plays led to just one out, at first base. That's what happened here, leaving runners on second and third with two outs. Punch. Another intentional walk, and another punch.

The ninth was uninteresting. Leadoff triple, groundout to the pitcher, walk.

So, does this record mean anything? It's not as though the Mets had to be an offensive juggernaut to do it. They scored only eight runs, and the previous three teams that pulled this off scored nine, 12, and 26. (1975 Astros; 1976 Astros; 1985 Phillies. Twenty-five teams managed 23 base/out states in a game, though, surprisingly, only three took extra innings to do it, and only one took longer than 10 innings.) It's not really something tied to the character of their team, other than having rabbits at the top of the order stealing third and having lefty/righty mashers in the middle drawing intentional walks. It's apparent from the game logs that it was a flukey thing, though it's not so flukey that you gasp in awe at the sequence of events. They were relatively mundane. And absolutely nobody in the ballpark would have noticed. Or, probably, cared.

So, resolved: This is best filed under Rare Event Trivia. It is not a record that means anything.

Thanks to podcast listener Dana for asking the question that led to this inquiry.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
dianagramr
2/23
I love stuff like this ... thanks for doing the legwork!
BrettLarter
2/23
So was this a play index search? How does one go about doing this?
whatzitmather
2/25
I believe Sam said on the podcast that he asked one of their research people to pull the data - not sure play index could handle this one.

It would be neat to have some sort of way to know in real time how many base/out states were remaining...short of having a bingo card...