“In this game you never know enough.”—Dale Mitchell

The hit-and-run play is not highly regarded by the analytical crowd. It is considered a one-run play and, like the sacrifice bunt attempt, it garners derision from people who hate small-ball tactics. No lesser authority than Earl Weaver preached against the use of the hit-and-run play in his book, Weaver on Strategy:

I don’t have a hit-and-run sign, and I believe it’s the worst play in baseball… you often give the opposition an out on the hit-and-run play. That’s because you can’t trust the pitcher to throw a strike, so the hitter is often waving weakly at a ball that’s off the plate. That usually results in a weak grounder that gets the runner to second, but the hitter is easily retired at first. Hell, you may as well bunt! Over the course of the season, only a few guys actually get hits on the hit-and-run play, because everything must go right for it to work. About the only thing you can say for the hit-and-run is that it prevents the double-play grounder. But when you add up the caught stealing, the weak grounders, and the line-drive double plays, that advantage vanishes. I’ll take my chances with a normal swing anytime.

As far as I can tell, no one has ever published a study on the value of the hit-and-run play. Thanks to data collected by MLB Advanced Media and provided by Retrosheet, it is now possible to research this question. The pitch sequence data from Retrosheet includes a notation of runners going with the pitch back to the 2003 season. The nine seasons of data from 2003-2011 should be enough to get a good idea of the value of the hit-and-run play.

The first problem to overcome is identifying the hit-and-run plays. Except on those rare occasions where the batter or the runner missed the sign, we know that the runner should be going and the batter should be swinging. However, not all such plays will be a hit-and-run.

First, how accurate are the stringers at identifying pitches on which the runners were going? The best benchmark I could find was that the runner was noted as going on about 90 percent of pitches resulting in a stolen base or caught stealing. So we are probably missing at least 10 percent of the hit-and-run plays. However, since I did not see a big difference in the quality of stringer data between parks, the data we have is hopefully still of good quality for addressing questions about the hit-and-run.

It is helpful to begin with a definition of the hit-and-run play so that we know what we are looking for. In the course of researching this topic, I read several good definitions, but by far the best exposition of the hit-and-run and the strategies involved appears in Keith Hernandez’s Pure Baseball, pp. 20-29.

The riskiest option, the hit-and-run, is one of the more strategically complex plays in the game. I don’t know any way to get into the subject other than at length because, as I mentioned, experience has taught me that some big leaguers don’t fully understand what the play is all about, when it’s a good idea, when it’s a terrible idea. So here goes. Be patient—the in’s and out’s of the hit-and-run go to the heart of the baseball strategy, and you have to understand them to understand the game.

Hernandez identifies the following characteristics of a hit-and-run play:

(1) a runner on first base, who by running forces a middle infielder to cover second in anticipation of a throw from the catcher, thereby opening a hole in the infield defense for the batter

(2) in a ball-strike count where the pitcher needs to throw a strike or something close

(3) not a two-strike count, because you cannot force the batter to strike out while attempting to protect the runner

(4) not a three-ball count, because the batter walks on a bad pitch and should take power swings at a good pitch

(5) usually with no outs, occasionally with one out, but never with two outs, because a defensive swing yielding a ground ball is likely the third out

(6) not with a power hitter at the plate.

He discussed additional nuances and strategies around the hit-and-run play that are well worth reading, but this summary hits the main points. I ended up with a definition very similar to the one outlined by Hernandez.

Let’s look at the situations where there was a runner at first base who ran with the pitch. The runner should be more likely to get caught stealing on a hit-and-run play where the batter swung and missed than if it were a straight steal and the batter took the pitch. Of course, it also may have been a straight steal if the batter swung and missed of his own volition.

I do not know any way of identifying or quantifying exactly which plays were hit-and-runs. However, the situations where the stolen base success rate was much lower when the batter swung than if the batter took are more likely to be hit-and-run situations. These likely situations, in combination with the fact that the batter swung and the runner ran with the pitch, should help us to identify a high proportion of the hit-and-run plays mixed with only a few straight steals.

Hernandez argued that the hit-and-run play was most likely to be used with no outs, where the following batters had the most opportunities to bring the baserunner home and the cost of a defensive swing leading to a groundball out was the lowest. Let’s see if we can pick that up in the stolen base success rate based upon the number of outs. The following table shows the data for situations with less than two strikes.

Outs Stolen Base Success Rate Batter Takes Batter Swings Difference 0 79.0% 60.4% -18.6% 1 77.3% 64.3% -13.0% 2 79.0% 81.1% +2.2%

When the batter swung and missed, the runner was caught stealing much more often with none out or one out than with two outs. This data is a good indicator that hit-and-run plays almost never happen with two outs. It also suggests that the hit-and-run is used a little more often with none out than it is with one out, though without having a good estimate yet for the stolen base success rate on pure hit-and-run plays, it is hard to calculate exactly how much more often.

Let’s look at the same data, but this time for ball-strike counts. I will include all of the ball-strike counts for comparison, even though hit-and-run plays are very unlikely to be executed with two strikes or three balls. We can check if the data backs up our assumption. Because I determined that two-out situations do not contain many hit-and-run plays, I will restrict the data in the following table to situations with less than two outs.

Count Stolen Base Success Rate Batter Takes Batter Swings Difference 0-0 80.5% 58.3% -22.2% 1-0 76.8% 60.7% -16.1% 2-0 84.3% 72.4% -11.9% 3-0 90.2% n/a n/a 0-1 77.7% 72.8% -4.8% 1-1 75.6% 59.3% -16.2% 2-1 72.6% 62.3% -10.3% 3-1 68.9% 63.9% -5.0% 0-2 82.5% 84.1% +1.6% 1-2 79.5% 84.7% +5.2% 2-2 78.4% 72.3% -6.1% 3-2 57.0% 54.0% -3.1%

The five ball-strike counts with a difference of more than 10 percent in stolen base success rate were 0-0, 1-0, 2-0, 1-1, and 2-1. There was some difference on the 0-1 and 3-1 counts, but it was small enough that we probably do not lose many actual hit-and-run plays if I exclude those counts from our list of likely hit-and-run situations.

The score of the game is another possible factor. Teams are unlikely to pursue strategies that put baserunners at risk when they are trailing by a large number of runs. However, teams are more liberal with the hit-and-run play relative to the score than I expected. The following graph shows the change in stolen base success rate when the batter swings and misses, given various score differentials for the batting team. The data in the graph is for situations with less than two outs and in the five likely hit-and-run ball-strike counts I identified.

The stolen base success rate decreased by about 19 percent when the score was within three runs. It decreased by about nine percent when the score of the two teams was exactly four runs apart. Though the runner-going situations with a four-run differential seem to include some straight steals, I will include them in our definition of a hit-and-run situation.

Finally, we arrive at the definition used for a hit-and-run situation in this study: (1) runner on first, bases not loaded, (2) none or one out, (3) a ball-strike count of 0-0, 1-0, 2-0, 1-1, or 2-1, and (4) the team leading or trailing by four runs or less. If the runner went on the pitch and the batter swung in such a situation, I will consider it a likely hit-and-run play.

The batter’s position in the lineup affects the likelihood of seeing a hit-and-run play in any given hit-and-run situation.

The second hitter is the most likely position in the lineup to be involved in a hit-and-run play, particularly on the first time through the order. The fourth and fifth-place hitters are relatively unlikely to be involved in hit-and-run plays. The probability of a hit-and-run play increases again at the bottom third of the order.

I could further screen our likely hit-and-run situations by removing the third through sixth spots in the lineup. However, I suspect the variation in hit-and-run rate is more an indication of a decision based upon hitter quality. It might vary more widely based upon specific personnel and managerial preference than some of the criteria that are more closely tied to the game situation. Thus, I chose not to use lineup position as part of the definition for a hit-and-run situation.

Here is how the hit-and-run plays break down by team from 2003-2011, as best I was able to count them using the Retrosheet data and the above definition of a hit-and-run situation.

Batting Team Hit-and-Run Plays Opportunities Hit-and-Run Pct. LA Angels 536 8319 6.4% St. Louis 535 8652 6.2% Detroit 416 8542 4.9% Minnesota 395 8404 4.7% Chicago Sox 387 8272 4.7% Tampa Bay 354 7828 4.5% LA Dodgers 378 8551 4.4% Toronto 368 8471 4.3% Seattle 359 8350 4.3% Kansas City 342 8222 4.2% San Francisco 347 8441 4.1% Houston 334 8136 4.1% Montreal/Washington 335 8239 4.1% Cincinnati 289 8137 3.6% Pittsburgh 286 8072 3.5% Florida 288 8252 3.5% San Diego 292 8381 3.5% Texas 275 8150 3.4% Colorado 284 8421 3.4% NY Yankees 303 9109 3.3% Milwaukee 273 8220 3.3% NY Mets 271 8332 3.3% Baltimore 271 8406 3.2% Chicago Cubs 250 8370 3.0% Arizona 240 8101 3.0% Cleveland 248 8449 2.9% Oakland 202 8793 2.3% Philadelphia 188 8608 2.2% Atlanta 178 8695 2.0% Boston 163 9045 1.8%

It is worth noting that American League teams used the hit-and-run slightly more often than National League teams during this period.

Here are the totals year by year for each team.

Team 2003 2004 2005 2006 2007 2008 2009 2010 2011 ANA 45 69 60 67 64 55 37 67 72 ARI 26 41 36 16 14 22 15 31 39 ATL 20 29 31 14 13 15 14 20 22 BAL 20 47 31 26 25 44 27 26 25 BOS 11 12 19 30 26 21 12 11 21 CHA 32 35 50 31 37 37 52 65 48 CHN 32 19 42 43 23 25 20 20 26 CIN 16 22 24 19 24 38 36 54 56 CLE 37 31 20 22 25 27 19 33 34 COL 21 24 37 26 22 23 36 36 59 DET 47 39 57 54 46 44 51 49 29 FLO 41 42 24 48 28 18 30 29 28 HOU 25 37 56 40 39 35 37 37 28 KCA 28 44 35 33 40 34 40 42 46 LAN 27 17 49 53 44 58 45 48 37 MIL 35 42 29 30 26 41 19 21 30 MIN 41 39 46 56 57 42 50 17 47 MON 45 50 41 52 15 26 35 36 35 NYA 36 23 25 39 47 60 28 27 18 NYN 37 24 40 23 34 30 34 28 21 OAK 25 14 23 18 29 15 21 37 20 PHI 17 24 23 13 30 25 13 23 20 PIT 51 49 38 22 31 15 29 16 35 SDN 31 35 38 28 25 18 28 39 50 SEA 30 41 38 36 60 53 26 47 28 SFN 26 44 46 28 38 57 33 35 40 SLN 56 61 73 69 58 50 44 61 63 TBA 24 28 41 59 36 32 24 38 72 TEX 37 22 15 16 27 24 30 53 51 TOR 34 48 59 61 40 39 18 14 55

Mike Scioscia’s teams led the major leagues four times in hit-and-run plays, and Tony La Russa’s teams led three times. For managers who spent the bulk of a season with a team and managed multiple years, here are those with most and the fewest hit-and-run plays. (I did not have managerial data split out at the game level in order to divide season data on teams with multiple managers in one season.)

Subject to the limitations of what I can discern from the data, we now have a pretty good idea of when hit-and-run plays occurred and which teams used them. But, is the hit-and-run play a good idea? That turns out to be a question fraught with selection bias.

Teams that attempted the hit-and-run play scored 0.11 runs on the play and 0.69 runs in the remainder of the inning on average, compared to 0.17 runs scored on the play and 0.70 runs in the remainder of the inning by teams that did not attempt the hit-and-run play in hit-and-run situations. On its face, it seems that the hit-and-run play cost teams about .06 runs per play. Before I wade through some of the factors that bias this simple answer, let’s examine how the results of hit-and-run batters differ from those of the batters who do not hit-and-run in the same situations.

The batters who attempted the hit-and-run play ended that plate appearance with an average batting line of .294/.323/.403, as compared to the batting line of .267/.324/.419 for the batters who did not attempt the hit-and-run play. At first take, the hit-and-run appears to confer about 26 more points of batting average, at the cost of some walks and extra base hits. Since the two pools of batters have very similar batting average skills, as we will discuss in more detail, this difference is largely attributable to the hit-and-run play. Creating the hole in the infield defense really is a valuable advantage.

The hit-and-run attempt also significantly cut the rate of walks, strikeouts, and the double play. It is no surprise that swinging at pitches early in the count, perhaps even bad pitches, cut the walk rate, dropping it from seven percent of plate appearances by batters who did not hit-and-run in these situations to only four percent of plate appearances by batters who attempted the hit-and-run play. The strikeout rate also dropped for hit-and-run batters, to 13 percent of plate appearances, compared to 15 percent for batters who did not execute the hit-and-run. Presumably, the increase in balls in play early in the count had a greater effect than did being forced to swing at bad pitches to protect the runner. The double play rate dropped from 13 percent of plate appearances when the hit-and-run was not attempted to nine percent when the hit-and-run was attempted. That rate includes both groundball and line drive double plays.

Baserunner advancement is another obvious consequence of the hit-and-run play. Extra runners caught stealing roughly balanced out the reduction in double plays, leaving about the same number of outs and runners on base on average whether or not a hit-and-run was attempted. However, the runner from first base had a better chance of moving up to second or third base on a hit-and-run play. The drop in extra-base hits that accompanied the hit-and-run slightly decreased the runners’ chance of scoring from first base as a result of that plate appearance.

Runner Result Hit-and-Run Play No Hit-and-run Difference Out 19% 19% 0% 1st Base 21% 37% -16% 2nd Base 41% 32% +9% 3rd Base 15% 8% +7% Scored 4% 5% -1%

Now that we have some idea of the typical results of the hit-and-run play, let’s examine some of the potential biases in our data, particularly in the average runs scored in the inning during and after the hit-and-run play.

One obvious consideration is that the number of outs affects the chances of scoring. The overall run scoring following the hit-and-run play looks better than it really was because the hit-and-run play was used more frequently with no outs, when run scoring expectations were highest.

Runs with No Outs Run with One Out On Play Rest of Inn. Total On Play Rest of Inn. Total H-R Play .09 .90 1.00 .13 .50 .62 No H-R .14 .93 1.07 .18 .52 .70 Difference -.05 -.02 -.08 -.05 -.02 -.08

I included as potential hit-and-run situations those with runners on first and second or runners on first and third. However, teams are much less likely to hit and run with multiple runners on base. If I restrict the analysis to the hit-and-run situations with only a runner on first base and the other bases empty, here are the run-scoring results.

Runs with No Outs Run with One Out On Play Rest of Inn. Total On Play Rest of Inn. Total H-R Play .08 .89 .97 .07 .49 .55 No H-R .07 .82 .89 .08 .45 .53 Difference .00 +.07 +.08 -.01 +.03 +.02

Once I have adjusted for the number of outs and made sure the baserunner state is the same, the hit-and-run play is a net gain for run scoring in the inning, particularly when it is attempted with no outs.

The quality of the batters and pitchers involved is another important concern. I calculated the average performance of batters and pitchers during the period 2003-2011 and weighted the result by the number of times they appeared in a hit-and-run situation. I compared it to the average performance of the batters and pitchers during 2003-2011, weighted by the number of times they attempted a hit-and-run play.

The pitchers were of similar quality in both pools, with an average of .004 fewer runs allowed than average, as calculated from their linear weights runs allowed. The batters who attempted hit-and-run plays, however, were noticeably worse hitters than those who did not attempt hit-and-run plays in hit-and-run situations. The batters who attempted a hit-and-run had an average batting line of .274/.342/.419, or +0.003 runs above average per plate appearance, as compared to the batters who did not attempt a hit-and-run, who had an average batting line of .275/.349/.448, or +.014 runs above average per plate appearance.

Before I add the difference in batter quality to the ledger in favor of hit-and-run plays, however, we need to consider the effect of the lineup order. Because more hit-and-run plays occur before the middle of the order comes up, the likelihood of a runner subsequently being driven in by the big bats is higher. Conversely, fewer hit-and-run plays occur ahead of the bottom of the lineup, where the weaker hitters are less likely to drive the runners home.

We can compare the number of runs scored in the remainder of the inning for attempted hit-and-run plays to the runs scored when a hit-and-run play was not attempted in a hit-and-run situation for each spot in the lineup. Here again, I restrict the data to those situations where there was only a single runner on base, in order to compare on a level playing field, or at least as level as I can make it.

The hit-and-run play is most advantageous in the second, third, and ninth spots in the lineup, and it is a net negative only for the fifth-place hitter, although I am not controlling here for the quality of batter involved in the hit-and-run plays, other than as represented by the lineup position.

The preceding graph shows the weighted average of the no-out and one-out situations. We can also look at them separately.

The no-out situation shows a bigger advantage for the hit-and-run attempt over the one-out situation at every lineup spot except for the second and sixth-place hitters.

In order to combine the remainder-of-inning scoring advantage with the batter quality advantage for a total hit-and-run advantage, I need to adjust both for lineup position. After adjusting for the lineup, teams scored .057 more runs when they attempted a hit-and-run play as compared to when they did not. After adjusting for the lineup spot, the quality of the batters who did not hit-and-run was .005 runs per plate appearance better than that of the batters who did hit-and-run. Thus, the advantage for attempting a hit-and-run play during 2003-2011 appears to be about .061 runs on average.

That is a much larger number than I expected when I embarked on this research. I have attempted to remove as much of the selection bias as I could reasonably identify. It is possible that I have overlooked some bias or used a mistaken assumption, but every direction from which I came at the analysis pointed to the hit-and-run being a positive offensive play in most circumstances in which it was attempted.

We have already considered how the number of outs and the lineup position affect the advantage to be gained via a hit-and-run play. The ball-strike count also plays a role. The more favorable the count is to the hitter, the less likely that the batter will be forced to swing at a pitch he does not like. On the other hand, the same is true if the batter is not protecting the runner, and in that case, he may be more selective and take more powerful swings.

Ball-Strike Count Hit-and-Run Attempt Pct. Hit-and-Run Advantage (R/PA) 0-0 1.2% +.041 1-0 3.2% +.050 2-0 1.9% +.031 1-1 3.2% +.026 2-1 6.6% -.026

The biggest advantage for the hit-and-run play comes early in the count, on the first pitch or at 1-0. However, 2-1 is the most popular count on which to attempt the hit-and-run play, even though the average result of the play in that count is a net negative.

With this in mind, let’s see which teams did the best job of turning the hit-and-run play into extra runs over the course of 2003-2011. This time I am considering only the plays with a single runner on base in order to compare apples to apples, as explained earlier, so the “plays” totals will be slightly lower than those listed in the previous team table.

Batting Team H-R Plays Oppor. H-R% Avg. Run Advantage Adj. for Batter Quality Total Net Runs NY Yankees 289 6512 4.4% 0.288 0.294 85 San Diego 263 6143 4.3% 0.175 0.180 47 LA Dodgers 335 6302 5.3% 0.123 0.132 44 Houston 316 6011 5.3% 0.092 0.111 35 Florida 264 6040 4.4% 0.127 0.133 35 LA Angels 492 6158 8.0% 0.060 0.068 34 Milwaukee 245 6015 4.1% 0.129 0.136 33 Detroit 390 6193 6.3% 0.072 0.083 32 Boston 151 6379 2.4% 0.176 0.189 28 St. Louis 495 6149 8.1% 0.028 0.052 26 Chicago Sox 359 6136 5.9% 0.054 0.069 25 NY Mets 248 6163 4.0% 0.085 0.095 24 Philadelphia 172 6232 2.8% 0.138 0.137 24 Seattle 329 6176 5.3% 0.060 0.068 22 Pittsburgh 265 5923 4.5% 0.073 0.084 22 Toronto 324 6060 5.3% 0.060 0.068 22 Baltimore 256 6163 4.2% 0.080 0.086 22 Minnesota 371 6177 6.0% 0.049 0.058 22 Mont./Washington 301 6068 5.0% 0.061 0.067 20 Cincinnati 267 6005 4.4% 0.058 0.063 17 Texas 255 6001 4.2% 0.051 0.062 16 Kansas City 324 6052 5.4% 0.010 0.016 5 Colorado 259 6151 4.2% 0.006 0.015 4 Chicago Cubs 241 6111 3.9% -0.017 0.000 0 Atlanta 165 6149 2.7% -0.014 -0.006 -1 Tampa Bay 323 5980 5.4% -0.010 -0.006 -2 San Francisco 301 6073 5.0% -0.027 -0.018 -5 Oakland 195 6282 3.1% -0.052 -0.042 -8 Arizona 216 5966 3.6% -0.053 -0.048 -10 Cleveland 217 6085 3.6% -0.111 -0.101 -22

Across nine seasons, a handful of teams have managed to add a total of 30 runs or more by using the hit-and-run play. The average team has added about two runs per season on hit-and-run plays.

Which batters hit and run most often? Here are the leaders from 2003-2011. I have also listed each batter’s success percentage in hit-and-run situations with a runner at first base only, where a successful hit-and-run play is defined as one that advanced the baserunner. League average was a 60 percent success rate.

Among others of interest, Jonathan Herrera attempted a hit-and-run play in 27 percent of his opportunities, Skip Schumaker had an 86 percent success rate at advancing the runner from first, and Albert Pujols attempted a hit-and-run play in three percent of his opportunities (with a 49 percent success rate). No word on how many of those plays he called himself.

Here are the baserunners who were on first base for the most hit-and-run attempts, along with their stolen base and caught stealing totals when the batter swung and missed in such situations. Since most of these players are regular basestealers, some of the plays I have classified here as hit-and-runs were probably actually straight steals.

The hit-and-run is far from the worst play in baseball. For a small-ball tactic, it has been quite successful over the past nine seasons, increasing scoring by .06 runs per attempt on average. The value of the hole in the infield defense is real, adding about 27 points to the batting average of the hitter. The double plays avoided by executing the hit-and-run offset the runners caught stealing on the play, and the extra bases gained by the runner when the ball is put in play are enough to move the play into the plus column overall.

However, there are some situations where the hit-and-run attempt made less sense and was a barely positive or even a net negative play—with the fourth and fifth hitters in the lineup up, with one out, or in the popular ball-strike count of 2-1.

Additional research could be done to fine-tune the definition of hit-and-run situations. In particular, I suspect that some of the situations with runners on first and second or runners on first and third could be removed. It would also be interesting to examine the detailed PITCHf/x data for the hit-and-run plays in the seasons for which that data is available.

Furthermore, future studies on the stolen base breakeven point or on the value of catcher defense in controlling the running game should probably incorporate what we have learned about the hit-and-run play.



The information used here was obtained free of charge from and is copyrighted by Retrosheet. Interested parties may contact Retrosheet at 20 Sunset Rd., Newark, DE 19711.