Baseball Toaster Mike's Baseball Rants
Help
This is my site with my opinions, but I hope that, like Irish Spring, you like it, too.
Frozen Toast
Search
Google Search
Web
Toaster
Mike's Baseball Rants
Archives

2009
01 

2008
10  09  07 
06  05  04  03 

2007
12  11  10  09  08  07 
06  05  04  03  02  01 

2006
12  11  10  09  08  07 
06  05  04  03  02  01 

2005
12  11  10  09  08  07 
06  05  04  03  02  01 

2004
12  11  10  09  08  07 
06  05  04  03  02  01 

2003
12  11  10  09  08  07 
06  05  04  03  02  01 

2002
12  11  10  09  08  07 
Links to MBBR
Base Relief
2006-05-25 22:24
by Mike Carminati
For this relief much thanks.
—William "Author" Shakespeare, "Hamlet"

There is a certain relief in change, even though it be from bad to worse!
— Washington "Julius" Irving

Relief pitching has been evolving ever since Harry Wright first inserted himself to spell a faltering starter back in the early days of the National Association, the stepfather of today's National League. Today, team's have closers, left-handed setup men, right-handed setup men, long relievers, situational lefties…we've gone from 10- to 12-man staffs since five-man rotations became standard. Teams are approaching, on average, four pitchers per game (3.71 last year up from 3.55 at the start of the decade).

But how do we know that all this effort has made any difference?

Are teams better off now? Do they hold leads any better than they did in the past? And if they do, does anyone know it?

Well, ever since I started delving into Retrosheet's game log data, I have been pondering how to answer those questions. I think I have the answer.

Taking the linescore data available, I was able to determine the games in which teams gained leads and how and if they retained them throughout baseball history. I did this for teams leading at the end of the sixth, seventh, and eighth innings in various situations. What follows are the results mostly in tables as is my wont.

Decade% Leads Retained after 6% Tied after 6
1900s79.81%14.91%
1910s78.07%14.91%
1920s76.70%12.76%
1930s76.14%12.97%
1940s76.75%14.17%
1950s76.88%14.06%
1960s76.49%14.81%
1970s76.49%13.87%
1980s77.06%13.44%
1990s77.51%12.85%
2000s78.09%12.40%
Total77.08%13.60%

It looks like the state of relief pitching is as strong as it's been since the dead ball era when complete games were prevalent and relief pitching was waiting for John McGraw to rediscover and reinvent it. Teams in the first half of the 2000s have been retaining leads at a 78% clip even as ties after six innings have become rarer and rarer (12.40% in the 2000s is the lowest ever) and scoring is still at record highs.

However, given the high scoring a 10-0 lead after six that ends up a 12-9 win would constitute holding a lead. Let's look at the lead-retention percentages, instead, based on the situation at the end of the sixth inning:

Decade% 1-R lead Retained% 2-R lead Retained% 3-R lead Retained% >3-R lead Retained
1900s58.83%73.65%87.79%96.72%
1910s57.10%72.87%86.78%95.65%
1920s52.36%69.73%84.26%94.80%
1930s50.88%69.07%82.02%94.99%
1940s52.79%72.14%83.08%95.76%
1950s52.46%71.60%84.98%95.80%
1960s53.01%72.48%85.63%96.24%
1970s52.74%72.67%86.42%95.54%
1980s52.20%73.87%85.33%96.00%
1990s53.52%72.66%84.41%95.52%
2000s51.87%72.68%85.75%95.74%
Total52.89%72.30%85.01%95.68%

So retaining one-run leads has not improved since the dead-ball era, and the results for the others have stayed at about the same level for decades. We have just seen bigger leads at the end of the sixth, which are easier to hold.

It seems that all the specialization in the bullpen has not really improved a team's ability to hold a lead. Though there are of course other benefits like saving a starter's arm and creating more opportunities to second guess a manager.

Next we will look at those situations that more closely examine closers and key setup men, and we'll look at which teams have been the most effective at holding leads. But that's for manana.

Comments
2006-05-26 02:29:53
1.   Underbruin
While interesting, I'm not so sure the statistics really accurately tell the whole story - yes, if my team is ahead by 2 runs and I've got a guy out there who has thrown 25 complete games this season pitching into the 9th inning, there's probably a better chance a fresh pitcher can hold the lead than he can. It just feels intuitively "right," to a degree.

Why is it possible this doesn't show in the stats? Well, because the -other- team ALSO has their guy, pitching into the 9th inning following 25 complete games so far this season. It's a balancing act. I guess a question, then, would be: are your tables indicating if a 1-run lead was "retained" or if the final score was simply that the team that was ahead by 1 run won by that amount (e.g. like a pitcher earning a BS in the top of the 9th even if his team comes back in the bottom of the 9th to retake the lead)?

2006-05-26 07:17:44
2.   studes
Mike, I don't understand the second table. Since 1979, home teams have won 77% of the games in which they had a one-run lead after six innings.

What is your table saying?

2006-05-26 11:58:33
3.   Mike Carminati
Studes,

The second table lists the percentage of games in which the team, whether home or visting, that led after 6 innings retain that lead for the rest of the game. That is, that there was no point in the game from the sevnth inning on in which the team lost the lead. The 77% that you quote, if I understand correctly, pertains to teams that won but that did not necessarily hold their lead for the rest of the game. I can check my data and get a subset of games in which a) the home team lead by one run after 6 and b) won the game. That should match the stats you mention.

2006-05-26 12:49:01
4.   studes
OK, so you would expect 50% at one-run leads, every thing else being equal because each team has the theoretical same chance of scoring. Except that you might assume that the team that already has the lead is the better team anyway and more likely to keep the lead.

Question: does the home team 9th inning difference matter in this analysis?

2006-05-26 12:50:38
5.   Mike Carminati
Studes,

Do you mean does homefield help? I can do home/road splits on the data.

2006-05-26 15:20:21
6.   studes
No, I was wondering out loud about the issue that home teams don't bat in the ninth if they're ahead, so they won't score as many runs after the bottom of the sixth as the visiting team. I'm not sure it matters for what you're doing -- just thought I'd bring it up.

Comment status: comments have been closed. Baseball Toaster is now out of business.