Do Teams Play for the Tie at Home?
Posted by Raphy on November 23, 2010
There is an old axiom in sports that you play for the tie at home, while on the road you play for the win. I was curious if this strategy was actually implemented in baseball, so I decided to use the "Batting Event Finder" to select one opportunity to play for a tie and compare the results at home and on the road. Teams which trail by one run with no outs in the ninth and a runner on first will only sacrifice the runner over if they are "playing for the tie." Here are the percentages of such plate appearances which result in sacrifice bunts at home and on the road. (Keeping in mind that the years prior to 1973 may not be 100% complete.)
| Years | Home PA | Home SH | Home % | Road PA | Road SH | Road % |
|---|---|---|---|---|---|---|
| 2006-2010 | 267 | 53 | 19.85% | 251 | 26 | 10.36% |
| 2001-2005 | 259 | 80 | 30.89% | 278 | 30 | 10.79% |
| 1996-2000 | 269 | 85 | 31.60% | 288 | 32 | 11.11% |
| 1991-1995 | 259 | 86 | 33.20% | 238 | 23 | 9.66% |
| 1986-1990 | 259 | 84 | 32.43% | 247 | 34 | 13.77% |
| 1981-1985 | 254 | 92 | 36.22% | 241 | 30 | 12.45% |
| 1976-1980 | 288 | 105 | 36.46% | 248 | 37 | 14.92% |
| 1971-1975 | 268 | 99 | 36.94% | 254 | 30 | 11.81% |
| 1966-1970 | 245 | 88 | 35.92% | 247 | 35 | 14.17% |
| 1961-1965 | 221 | 71 | 32.13% | 236 | 24 | 10.17% |
| 1956-1960 | 171 | 73 | 42.69% | 169 | 23 | 13.61% |
| 1951-1955 | 131 | 37 | 28.24% | 144 | 22 | 15.28% |
| Total | 2891 | 953 | 32.96% | 2841 | 346 | 12.18% |
The key here is to compare the percentage of plate appearances at home that are sac-bunts vs. those on the road. A quick look at the table shows that there is a tremendous difference.
Here is the same information in chart form. As you can see the difference is quite striking. ( The years recorded are the statistics for the previous five years. So when 2010 is marked in the chart, it really is 2006-2010.)
Since 1950 a home team is more than 2.7 times as likely as a road team to bunt a runner over from first when trailing by 1 with no outs in the ninth. While the degree has varied over time, and individual single season anomalies exist (if you're curious, check out 2008), this has essentially been true in baseball for the last 60 years. Judging from this one scenario, it would be logical to conclude that the axiom is true. Home teams play for the tie; road teams play for the win.
November 23rd, 2010 at 8:45 pm
This was really interesting. Rather than looking at whether or not teams SHOULD utilize this approach, you took a different approached and first determined whether or not they even do. Interesting stuff!
November 23rd, 2010 at 9:03 pm
Great stuff.
That AL drop in '06-'10 is startling and seems statistically significant. Took a long time for teams to catch up with the higher scoring?
Now the question is, is bunting so much more at home the correct strategy?
I am using this handy little WPA calculator I just found at Hardball Times:
http://www.hardballtimes.com/thtstats/other/wpa_inquirer.php
Using a 4.5 RPG environment:
In the top of the 9th, the visiting team's chance of winning is 27.7%. If the bunt is successful, their chance falls to 22.5%.
In the bottom of the 9th, the home team's chance of winning is 33.4%. If the bunt is successful, it falls to 28.1%.
Both drop by about the same number of percentage points, though it is a bigger percentage drop for the visiting team. So on average, it's not the right play to bunt there. But, if depending on the batter, pitcher, position of the fielders, the bunt seems to be the right play, it doesn't appear from this analysis that the home team should be bunting so much more often than the visitors. It looks like an equally good/bad play for both.
Now, what if the run does score? If the visiting team gets to the bottom of the 9th tied, they have a 36.3% chance of winning. If the home team gets to the 10th inning, they have a 50.0% chance of winning (actually, I'd say it's slightly higher than that, but we'll use it). So the visitors chance of winning improved by 31%, and the home chance improved by 50%. The value of one run is worth 61% more to the home team, so by that analysis it does make more sense to bunt (but probably not as many times more than the visitors as they actually do).
Someone please tell me how I've mangled my math. Or look at it another way.
November 23rd, 2010 at 9:14 pm
Johnny - I was going to mention the drop-off, but I didn't want to confuse the issue. The data for individual years is limited, but FWIW, the numbers seemed to have returned to normal after a 4 year drop off from 2006-2009. Here are the year by year numbers since 2004:
November 23rd, 2010 at 9:21 pm
@2. While the WPAs you ran may show that bunting is a bad play, remember that the closer is pitching in these scenarios, so the actual scoring environment is much lower.
November 23rd, 2010 at 9:24 pm
I too was looking at the sharp drop for the past five years and thinking about the cause. I would think it would have more to do with "moneyball" (for lack of a better term) strategies proliferating baseball that discourage the use of the bunt in general, but that is bald speculation.
The next logical question is, as has been commented above, whether utilizing the bunt less frequently is a better, worse or neutral strategy. If it is a better strategy we should expect to see an increase in the home field advantage enjoyed by teams over the past 5 years as home teams have relied less upon the bunt in these situations (I don't have the time to look up home winning percentages over the years, but I do recall reading somewhere that home teams were winning a lot more of late, along with speculation that this was attributable to the crackdown on amphetamines).
I like JT's @2 approach of examining in terms of WPA. My initial thought was to look at winning percentage for times when the bunt was utilized vs. not utilized, but I realized that this is a biased split because having your best hitter up in the situation analyzed would dramatically reduce your chances of bunting (vs. having a weak hitter up), but should also increase your chances of winning because of his hitting ability, as well as that of the batters who would typically follow your best hitter in the lineup.
November 23rd, 2010 at 10:50 pm
remember that the closer is pitching in these scenarios, so the actual scoring environment is much lower
Very true. I mentioned that it may not be a bad play dependent on the pitcher, but I should have just changed the run environment.
So let's suppose with a good closer the run environment is just 3.0 RPG. Even there, the visiting team drops from 23.4% to 19.7% with a successful bunt, and the home team from 26.9% to 23.1%. This somewhat correlates to the findings in The Book, where at 3.2 RPG the chance of scoring after the successful bunt drops from 37.4% to 36.2%.
November 23rd, 2010 at 10:56 pm
However, The Book also says explicitly: "Late in a close game, in a low run-scoring environment, it is correct to often sacrifice bunt with a runner on first and no outs." [emphasis mine]
I'd have to re-read the chapter to see how they come to that conclusion which is somewhat contrary to the run-expectancy findings. (Win expectancy is the key instead of run expectancy, but in the 9th inning they become quite similar.)
November 24th, 2010 at 8:49 am
I have heard this axiom many times in reference to baseball, but I've never heard it in reference to any other sport... and I can't think of a reason why it would make sense in any other sport besides baseball (the home team bats last, so as long as you tie the game in the last inning, you at least guarantee yourself another chance if you don't win in this inning). Have any of you guys heard people say that home teams play for the tie in other sports? If so, is there a way for it to make sense?
November 24th, 2010 at 9:38 am
Maybe in football, choosing whether to go for 2 after a late TD? I think I've probably heard it there. I guess one could assume the home team will have a slight advantage in OT, though I'm sure there aren't enough OT games for the actual results to have much meaning.
November 24th, 2010 at 10:44 am
Since 1950 a home team is more than 2.7 times as likely as a road team to bunt a runner over from first when trailing by one with no outs in the ninth.
I won't bore you with how I came up with the search parameters, but here's an extreme example supporting Raphy's conclusion: L.A. Dodgers' opponents did not successfully sacrifice at the Coliseum or Dodger Stadium when down by one to the Dodgers in the ninth inning, no outs and runner on first, from the arrival of big-league ball in Los Angeles until May 1965, a span of seven years. During that time Dodgers' opponents successfully sacrificed in their home parks in that situation 11 times.
November 24th, 2010 at 11:05 am
JT, The biggest difference between your numbers and those from The Book is that you are assuming that 100% of the time someone attempts to bunt there is a successful sacrifice. This is not true, it is much closer to 75%. Sometimes the results are worse, force of the lead runner, DP, pop-up, K. And sometimes they are better, BB, hit, RoE, FC no out. All of these possibilities must be considered, (and correctly weighed), to get an accurate RE/WE. And of course it should be adjusted to the specific batter and pitcher. Maybe I-Rod should bunt more to avoid DPs.
November 24th, 2010 at 12:50 pm
Of course, Kds, thank you. I was aware of that as I did the searches but didn't bother to consider it when I wrote my conclusions, such as they were.
November 25th, 2010 at 1:45 am
Adding on to Kds's point @11 --
The chance of a sac attempt failing is, I think, widely underestimated by MLB managers, which warps their judgment of when a sac attempt is appropriate.
November 25th, 2010 at 1:57 am
Complete tangent: Bunting for a double.
According to the B-R split data, Cliff Pennington had a bunt double this year. Did anyone hear of this? I googled but couldn't find a thing. Other recent bunt doubles found in the B-R split data have been confirmed by other sources (Rafael Furcal in 2009, Juan Pierre 2007 and 2004).
November 25th, 2010 at 7:20 am
John - I once wrote a post on the topic:
http://www.baseball-reference.com/blog/archives/2099
November 25th, 2010 at 10:16 am
"Have any of you guys heard people say that home teams play for the tie in other sports? If so, is there a way for it to make sense?"
Is there any other sport where the home team always gets the last chance on offense?
November 25th, 2010 at 10:37 am
"Have any of you guys heard people say that home teams play for the tie in other sports? If so, is there a way for it to make sense?"
I don't know if it is true or not, but I did see it in regards to basketball with the explanation being that home teams have a large advantage in OT.
I did a quick check with the NBA PI and it seems that home teams win OT games about 52% of the time. (This may be incorrect as it was the first time I used this tool.) It would seem very rare that such a small disadvantage would warrant adding the risk of taking a 3 pointer on the road
November 26th, 2010 at 6:16 pm
@15, Raphy -- Thanks for the directions. Your blog post noted that you used the Team Batting Event Finder to compile your list of bunt doubles. Could you explain exactly how that helped you? The only bunt-related events I see there are "Sacrifice Bunts" and "Sacrifice Bunts Att.," neither of which is broad enough to encompass all bunts.
Also, do you have any line on Pennington's (alleged) bunt double?
November 29th, 2010 at 5:58 pm
@18 John -
Here's what I did.
For each season I checked the hit trajectory split on the league batting splits page.
If there was a 2B listed under "bunts", I clicked on bunts and found out which team had the bunt double.
I then used the Batting event finder and searched for all doubles for that team that season.
I was then able to use the web browser find feature to search for the word bunt within the web page.
I don't remember if some teams had multiple pages of doubles or not.
November 29th, 2010 at 6:10 pm
@18 cont.
Re: Pennington:
His double occurred on April 8th. It is the second one listed here:
http://bbref.com/pi/shareit/m51RR
Its not in their highlights, but if you have access to mlb.tv (I don't) you can watch it in the 8th inning of that game.