The bases are loaded and I wish I were — Skip Caray

If you’re boycotting the All Star Game, here’s some stuff to read. Or you can read it while you’re watching the ASG. Or whatever. I’ve heard I’m supposed to keep providing “content.”

Is Batting Average with RISP a Thing?

I’ve spent the last week or so looking at BA with RISP, and I’m coming to a few conclusions which I’ll try and convince you are correct.

  • Batting average with RISP is extremely important to scoring runs. Over half, and up to 2/3, of every team’s runs result from hits with runners in scoring position.
  • But the biggest component of scoring runs with RISP is getting to RISP. And teams which get a lot of runners into scoring position are already good-hitting teams, so that even if they hit somewhat worse with RISP than without RISP, they will still score a lot of runs because their inferior BA with RISP is off a higher base and they have so many opportunities to start with.
  • Most teams have higher BAs with RISP than without. There are several reasons for this, but the main one is that teams which have managed to get runners in scoring position are facing less competent pitching at that moment — that’s how the runners got in scoring position in the first place.
  • There’s a lot of variation around expected BA with RISP (adjusting for team BA without RISP). There is so much variation that I’m doubtful that Batting Average with RISP is a team skill. Like other things that appear to be skills (e.g. clutchness or playoff success) a lot of what looks like competence and incompetence is just luck. And BA with RISP has an unusually large compenent of luck for a lot of reasons.
  • As far as the Braves are concerned, we’ve got 99 Problems, but RISP BA isn’t one of them, except insofar as it is derivative of other batting problems.

RISP Facts

Let me start with the table I put together last week, which covers the 2024 season.

Non-RISP RunsNon-RISPNon-AB with RISPAB with RISPBAAll Runs
HRNon-HRABHitsBAPlaysRunsPlaysHitsRunsw/RISPTotal
Team
ARI17630408310420.255238210114394105790.2849886
LAN22919411010480.25503767714123755170.2656842
NYA22524418410210.24403218712663314790.2615815
BAL25022426910650.24952436412983264500.2512786
PHI19228416810590.25413246513663644990.2665784
MIL1483240019650.24123516714713945300.2678777
NYN19021416910020.24033006913413554880.2647768
SDN20227423811090.26172885112883474800.2694760
BOS18228418110590.25333396113963454800.2471751
MIN16730412810210.24732867313623314720.2430742
HOU17824416510840.26032857113653644670.2667740
CHN1622341059920.24173498313363264680.2440736
KCA1493041759920.23763067412463514820.2817735
CLE1782441079490.23113157612033144300.2610708
ATL20819423610260.24222556212453074150.2466704
CIN1552141139200.22373577412123104490.2558699
SFN17739417110010.24002536512893024120.2343693
TEX1701742459700.22852806012273324360.2706683
DET1452542419600.22642366112013134510.2606682
COL17321418610050.24012726012683144280.2476682
SEA1911240838980.21993216312472974100.2382676
SLN18126421410670.25322977912932963860.2289672
TOR1591541099730.23682896613013334310.2560671
PIT1582042099910.23543237412682924130.2303665
WAS1052040289590.23813537413463474610.2578660
OAK2012242469970.23482516211862703580.2277643
MIA13626422210120.23972745313003354220.2577637
ANA1572041669490.22783237111912783870.2334635
TBA1392540959670.23613036012942743800.2117604
CHA1252342179320.22101984011662553190.2187507

Two teams had over 1400 AB with RISP, and those two teams were the two highest-scoring teams in MLB (the Dodgers and DBacks) and were two of the three teams to score over 500 runs by at bats with runners in scoring position. (It should be noted that these weren’t all hits. Many of these scored via outs.) These teams scored more runs by at-bats with RISP than the White Sox scored overall.

This also translates to more runs scored from non-at-bats (walks, wild pitches, steals, etc.)

For those who like graphs (and who doesn’t?) we can succintly summarize the distinction bewteen BA with and without RISP with this chart:

Here I have graphed the BA without RISP on the x-axis and the corresponding BA with WISP on the Y axis and placed the team at the intersection. I have drawn in the 45 degree line (which is not at 45 degrees on this diagram because the x scale and y scale are different.) Any team above the line had a higher batting average with RISP than without. 22 teams are above line and only 8 are below it.

Why is this? Well, first there are events which would be AB outs which are potentially non-ABs with RISP. The most obvious example is outfield fly balls with men on third and less than two outs which magically transform from “out” to “sac fly.” But quantitatively, this can’t explain the difference. A second reason is that defenses often play in, or at double play depth, with RISP. This may be optimal at preventing runs (net) but it does so at the expense of more hits And the final reason is probably the biggest one. A pitcher who has yielded RISP isn’t pitching as well as one who doesn’t. Under that circumstance, it is unsurprising that subsequent batters do better.

There are other effects that might go in either direction. Batters may focus more with RISP, but so may pitchers. RISP situations are likely to see pitching replacements and pinch hitters which should be expected to work in opposite directions. It would obviously require a lot of intense analysis to separate all these effects, and we really don’t have enough plate appearances to really try.

What we can do is run a simple analysis to see how BA with RISP is correlated with BA witthout RISP. Running this regression on the data above we get:

BA with RISP = 0.114 + 0.574 BA without RISP

What this equation means is that you take about 60 percent of BA without RISP and then add 114 points. That line is shown on the graph as well: 16 teams are above it and 14 are below it, but there is a very large discrepancy around the line: the standard error of this regression is about .016, so results plus or minus .032 are unsurprising. Thus, a team with a BA without RISP of 0.250 has an expected BA with RISP of 0.114 + 0.574 x 0.250 = .257, but we shouldn’t be surprised to see a number anywhere between .225 and .289, which is essentially the whole range of the chart.

Why Does Everyone Talk About BA with RISP?

The first reason is simple: RISP-BA is the first cousin of “clutch.” While not all RISP situations are high leverage, almost all high leverage situations have RISP. When a batter fails with RISP, or when a team fails to get a run home with 2nd and 3rd and no one out and ends up losing by one, it is easy to look backwards at that moment.

The second reason is simple as well: because it is a common calculated number. The more relevant number would be conversion percentages. If you are expected to score 1.5 runs in a given situation, how many runs do you score over 100 of them? If it’s way under 150, you’re doing something wrong. A further problem with conversion percentages is that there is a different one for all 24 base-state, out situations… that’s 24 numbers, and since most runs score with RISP, a single RISP number is easier to discuss. It should be noted that this measure gives short shrift to walks (in the same way that Batting Average continues to overshadow On Base Percentage) and simple things like groundouts that drive in runs. Jurickson Profar’s game-winning 9th inning plate appearance on Saturday was a groundout: he worsened the Braves’ BA-RISP while simultaneously winning the game.

Third, like “clutch” (or really coterminously with “clutch”) it is thought to be a skill. It is thought that some players can rise to the occasion and others can’t, and this skill shows up as clutch and BA-RISP. Mere mortals would indeed be challenged to perform is such tense high-leverage situations. The vast majority of MLB players aren’t, for the simple reason that those who were intimidated in this fashion never made it to the major leagues. At the margins, there are probably managerial strategies that slightly improve run production in these situations: judicious use of pinch hitters, hit-and-runs, squeeze bunfs, etc. But essentially, there is little difference that we can see between hitting with RISP and hitting with the bases enpty other than the fact that hitting with RISP requires some pitcher to have put runners on — by definition he isn’t pitching at an elite level at that moment, on average.

What About the 2025 Braves?

Here are the Braves stats through the All Star break:

SplitPAABRH2B3BHRRBIBBSOBAOBPSLGOPS
RISP89076227318437221261962110.2410.3240.3780.702
—200018106244676762621724460.2460.3180.3990.717
1–716655531542302252521620.2350.2940.3710.665
Non-RISP27162465115600997841142246080.2430.3120.3910.703

So the Braves indeed have slightly underperformed in BA and OPS with RISP. But the difference is quite slight. Even if we were use the regression adjustment to say what the BA with RISP ought to be, we would get .114 + 0.574 x 0.243 = 0.253. The difference over the 762 opportunities is 762 x (0.253-0.231) or 17 hits. This translates to about 20 extra runs. Now 20 runs is nothing to sneeze at, and if they had been in exactly the right spots it might have been amazing. But it takes roughly an extra 10 runs to win one more game, so the Braves hitting with RISP has cost them about 2 games. (There’s no point in being much more precise here given the crudeness of the regression calculation.)

In other words, we got 99 problems and RISP BA is a small one.

Comments welcome.