Having followed Longhorn baseball for most of our lives, nearly all of us are familiar with the traditional Texas approach to scoring runs. There seems to exist a segment of the fanbase that frowns upon power in a baseball lineup and will always yearn for smallball as their style of choice. Some of that is certainly based on a tendency to want to stay in your comfort zone, some of it is probably based on the ridiculous brand of 1990s "baseball" known as gorilla ball perfected by the LSU Tigers.
Regardless of each individual's motivation, it's hard to argue against the notion that the Texas fanbase overall is very hostile to the idea of Earl Weaver baseball at Disch-Falk. What I take issue with, though, is the idea that just because Disch-Falk is a pitcher's park (although not as extreme as most Texas fans believe) that necessarily means bunting is generally a good idea. It's no secret that I'm not a big fan of bunting in nearly any situation, although I'm very much in favor of the safety squeeze in situations where a single run is called for as long as the players are competent at its execution. Outs are an offense's most precious commodity in the game of baseball and giving them away should therefore cause a pain in your gut. It's true that college defenses are not nearly as precise as their professional counterparts, but I have always found it hard to believe that the difference there is enough to offset the proven wrongness, on the major league level, of bunting in most situations.
The phenomenal Boyd Nation has put together an expected runs table for college baseball based on the outs/runners situation for the years 2005-2008. It's obviously important to keep certain things in mind when reviewing the table, such as park factors, team lineup, etc. However, that shouldn't stop us from being able to draw general conclusions related to college baseball. Take a moment to review the table and then we'll look at possible bunting situations, both likely and unlikely, and what the numbers say.
(Information below is formatted as follows: Situation - Expected runs in the inning, % of innings scoring at least one run, Expected runs after successful bunt, % of innings scoring at least one run after successful bunt)
Runner on 1st, 0 outs - 1.10, 52, 0.84, 48
Runner on 1st, 1 out - 0.65, 34, 0.39, 27
Runner on 2nd, 0 outs - 1.39, 69, 1.08, 70
Runner on 2nd, 1 out - 0.84, 48, 0.44, 31
Runners on 1st and 2nd, 0 outs - 1.83, 72, 1.59, 75
Runners on 1st and 2nd, 1 out - 1.16, 51, 0.73, 35
Runners on 1st and 3rd, 0 outs (bunt to 2nd and 3rd) - 2.07, 89, 1.59, 75
Runners on 1st and 3rd, 1 outs (bunt to 2nd and 3rd) - 1.40, 71, 0.73, 35
So, based on the view from 30,000 feet, the only situations in which the numbers say bunting makes sense are those with a runner on second base and nobody out - but only if the game situation suggests that you need to do everything you can to score at least one run because even then your expected value in runs for the full inning goes down. Essentially, the only time the numbers say you should bunt if you're an average team in an average park is when you are tied or down by one run and you have a runner on second base (whether or not someone is on first) and nobody out late in the ballgame. And if you don't have a pitching staff you trust a whole lot, it should probably be the bottom of the ninth inning. Even in these situations, the numbers aren't very convincing as the percentage goes up from 69% to 70% in one and from 72% to 75% in the other. Another thing to keep in mind about these numbers is that the expected runs values are slightly depressed by game-ending scores, although that effect is likely negligible.
The data also tells us is that even in the best-case scenario for bunting from the EV(runs) perspective, bunting a runner or two over will cost your team 1/4 of a run on average. And that's only if the sacrifice is successful. Finally, this analysis slightly depresses the disparities in any particular situational perspective. This is because, for example, the runner on 1st with no outs number includes the runner on 2nd with 1 out situations caused by sacrifice bunts, which we know score at a lower rate than the former situation.
The other aspect of offense that these numbers allow us to review is the stolen base. Using the same table and accounting for both successful and failed stolen base attempts, we can determine the break-even point for each situation. This point represents the success rate that a runner must have in order to make the attempt worthwhile. Note that this assumes no impact on the current at-bat (i.e., assumed to occur early enough in the count not to affect the pitcher/batter matchup).
(Situation - Required success rate to increase EV(runs), Required success rate to increase % chance of scoring at least one run)
Runner on 1st, 0 outs - 72.9%, 66.7%
Runner on 1st, 1 out - 74.0%, 65.9%
Runner on 1st, 2 out - 71.8%, 63.0%
Runner on 2nd, 0 outs - 82.3%, 76.1%
Runner on 2nd, 1 out - 75.3%, 65.1%
Runner on 2nd, 2 out - 88.6%, 87.1%
Runners on 1st and 2nd, 0 outs (double steal - assume lead runner caught if unsuccessful) - 66.9%, 57.1%
Runners on 1st and 2nd, 1 out (double steal - assume lead runner caught if unsuccessful) - 64.2%, 50.0%
Runners on 1st and 2nd, 2 out (double steal) - 78.1%, 85.7%
Runners on 1st and 2nd, 0 outs (lead runner only) - 83.1%, 69.1%
Runners on 1st and 2nd, 1 out (lead runner only) - 78.6%, 63.0%
Runners on 1st and 2nd, 2 out (lead runner only) - 87.7%, 83.3%
Runners on 1st and 3rd, 0 outs (steal 2nd) - 83.1%, 97.6%
Runners on 1st and 3rd, 1 outs (steal 2nd) - 84.2%, 89.0%
Runners on 1st and 3rd, 2 outs (steal 2nd) - 89.0%, 102.9%
For reference, Texas had a stolen base success rate of 75.3% last year.
Here we can see one particular piece of conventional wisdom confirmed - never make the first or third out at third base. The only time it makes sense for even a decent base-stealer to attempt to swipe third base is with one out. No surprise there. Of course, we can also spot an anomaly in the data where the table says that at least one run is scored more often in a 1st and 3rd with 2 outs situation than in a 2nd and 3rd with 2 outs situation. I'm not sure why that would be, but there's a lot of data behind it. Perhaps not enough to make it more than a fluke, though. The only thing I can come up with is that opposing coaches are more willing to issue the intentional walk in a 2nd and 3rd situation that creates force plays at 3rd and home, enabling more escapes without any runs scored but increasing the average number of runs allowed for the remainder of the inning.
The other thing I noticed immediately is that the double steal is the best stolen base strategy available. The ideal stolen base attempt is a double steal with 1 out in the inning.
2008 Texas Longhorn Statistics
But what about our particular team and park? Well, I'm not as good to you as Boyd Nation is to us all, so I only looked at the 2008 numbers for Texas from the box scores still available on texassports.com as of today. Yes, I went through each game's play-by-play. The data:
|1st and 2nd||2.10||1.68||0.73|
|1st and 3rd||1.38||1.60||0.85|
|2nd and 3rd||1.70||3.38||1.35|
|1st and 2nd||77%||67%||39%|
|1st and 3rd||75%||70%||36%|
|2nd and 3rd||100%||88%||48%|
I'm not going to repeat the full analyses above, but the basic lesson is the same. A couple of notes about our particular results last year are that our team struggled in situations with a runner on third and two outs and performed far better than the norm in situations with runners on 2nd and 3rd and one out. Nevertheless, the important thing is that our team, just like the national average, scored more runs per inning and scored at least one run more often in runner on 1st and no outs situations than they did in runner on second and one out situations. Also like the national average, we gave up about 1/4 of a run each time we bunted that runner over and decreased our likelihood of scoring even a single run by 3-4%. The number of data points for some situations isn't sufficient for final conclusions, but the trends indicate that the national averages above can serve as a good measuring stick for situational decision-making. Our numbers were, of course, higher than the national average in most scenarios, but the relationship between the numbers is similar.
In summary, STOP BUNTING AUGIE!