## Splits Tease 2: Pitchers by Run Support, Blyleven and Morris

Posted by Sean Forman on March 16, 2010

I'll soon be adding a split that shows how pitchers do when their teams score 0-2, 3-5, or 6+ runs in a game.

Here is a little taste.

Pitcher | 0-2 runs | 3-5 Runs | 6+ runs |
---|---|---|---|

Jack Morris | 143 G, 17-110 (.134), 6 SHO, 4.00 ERA | 216 G, 98-66 (.598), 16 SHO, 3.56 ERA | 190 G, 139-10 (.933), 6 SHO, 4.21 ERA |

Bert Blyleven | 232 G, 39-163 (.193), 22 SHO, 3.35 ERA | 273 G, 119-81 (.595), 22 SHO, 3.27 ERA | 187 G, 129-6 (.956), 16 SHO, 3.33 ERA |

This is almost the perfect example of Simpson's Paradox. Blyleven was (very nearly) had the better winning percentage in each category of run support, but has the worse overall winning percentage due to the weighting of the various categories.

Rich Lederer feel free to use this.

March 16th, 2010 at 11:04 am

Great splits idea!

March 16th, 2010 at 11:21 am

how about when they allow those splits in a game?

March 16th, 2010 at 5:07 pm

How about when they allow those splits in a game?I have that information for Blyleven's teams.

0 runs: Blyleven 0-40 with 2 NDs

1 run: Blyleven 15-65 with 15 ShOs and 5 NDs

2 runs: Blyleven 24-58 with 7 ShOs and 23 NDs

3 runs: Blyleven 29-49 with 5 ShOs and 20 NDs

4 runs: Blyleven 38-19 with 6 ShOs and 25 NDs

5 runs: Blyleven 52-13 with 11 ShOs and 25 NDs

6 runs: Blyleven 28-2 with 6 ShOs and 16 NDs

7 runs: Blyleven 30-1 with 4 ShOs and 13 NDs

8 runs: Blyleven 25-2 with 1 ShO and 8 NDs

9 runs: Blyleven 14-1 with 1 ShO and 5 NDs

10 runs: Blyleven 12-0 with 2 ShOs and 4 NDs

11 runs: Blyleven 4-0 with 2 NDs

12 runs: Blyleven 6-0 with 2 NDs

13 runs: Blyleven 5-0 with 1 ShO and 1 ND

14 runs: Blyleven 4-0 with 1 ShO

15 runs: Blyleven 1-0

Overall: Blyleven 287-250 with 60 ShOs and 151 NDs

Just to lay to rest one myth about Blyleven's career: He was 15 and 10 in 1-0 games. The only final score that I've found for which his record is much worse than might be expected is 3-2 — in those games Blyleven won only 8 and lost 22. Blyleven was

in 3-2 games during the five-year stretch 1976-80, when his overall W-L record was 61-56, with 52 NDs, 17 ShOs and a 3.18 ERA.0-13March 16th, 2010 at 5:18 pm

I wonder whether you could get an example of Simpson by changing the boundaries, e.g., use 0-2, 3-4, and 5+, or maybe 0-3, 4-5, and 6+. This is not to be confused with Suitcase Harry Simpson's paradox, http://www.baseball-reference.com/players/s/simpsha01.shtml, which is the one where you can never tell what team you'll be playing for next week.

March 16th, 2010 at 6:48 pm

Sorry, Dave — I realized after posting that you were asking about runs allowed in a game. My post #3 addressed the runs scored by Blyleven's teams. Here's that info.

Team allowed 0 runs: Blyleven 67-0 with 60 ShOs and 1 ND

Team allowed 1 run: Blyleven 79-10 with 11 NDs

Team allowed 2 runs: Blyleven 51-21 with 15 NDs

Team allowed 3 runs: Blyleven 43-34 with 20 NDs

Team allowed 4 runs: Blyleven 32-47 with 27 NDs

Team allowed 5 runs: Blyleven 10-45 with 26 NDs

Team allowed 6 runs: Blyleven 2-33 with 21 NDs

Team allowed 7 runs: Blyleven 3-19 with 11 NDs

Team allowed 8 runs: Blyleven 0-15 with 7 NDs

Team allowed 9 runs: Blyleven 0-10 with 2 NDs

Team allowed 10 runs: Blyleven 0-4 with 2 NDs

Team allowed 11 runs: Blyleven 0-5 with 3 NDs

Team allowed 12 runs: Blyleven 0-1 with 3 NDs

Team allowed 13 runs: Blyleven 0-2 with 1 ND

Team allowed 14 runs: Blyleven 0-2

Team allowed 15 runs: Blyleven 0-1

Team allowed 18 runs: Blyleven 0-1

Team allowed 20 runs: Blyleven 0-0 with 1 ND

Overall: Blyleven 287-250 with 60 ShOs and 151 NDs

March 17th, 2010 at 3:46 pm

Please Mr. Tuna tell us bout that 20 run game? Assuming he started - how did he got off the hook for the L? Did he get hurt early?

March 17th, 2010 at 5:05 pm

See here: http://www.baseball-reference.com/boxes/MIL/MIL199007080.shtml

Blyleven didn't pitch well but came out in the 4th with a lead. The Angels' bullpen then got lit up in the 5th. Oddly (or maybe not?) it was three pretty good relievers who combined to give up 13 runs that inning (Minton, Fetters, and Eichhorn -- combined for 43 major league seasons).

March 17th, 2010 at 5:09 pm

The 20-run game is here: http://www.baseball-reference.com/boxes/MIL/MIL199007080.shtml

Blyleven started, pitched 3-and-a-third, gave up 6 runs - but his team, the Angels, had already scored 7, so he couldn't get the loss. The Brewers scored 13 in the 5th.

March 17th, 2010 at 6:42 pm

Blyleven has considerably more shut-outs in games where his team scored fewer runs. Now, I'm sure a large part of this is due to the fact he was pulled from games in which he was shutting a team out but his team had scored a lot of runs. But, even on the lower end, in games where his team scored 2, 3, or 4 runs, there is a drop. Is this just coincidental? Does this lend any credence to the argument that, staked to a lead, a pitcher pitches differently and might give up runs he wouldn't otherwise? I struggle to think the inverse is true (his team isn't scoring so somehow he is trying harder), but could imagine pitcher's relaxing, even with leads of just 2, 3, or 4 runs. Can this be parsed out from the data? Can you figure out how many times he left a game with a shut-out in tact?

March 18th, 2010 at 12:11 am

If Blyleven (or any other pitcher) has more shutouts in games where his team didn't score much, it might just be that conditions weren't conducive to scoring that day - maybe the wind was blowing in, or the game was played in Shea Stadium, or scoring around the league was down that year, or the other team arranged its rotation so its ace would go against Blyleven, or....

March 18th, 2010 at 6:43 am

Thanks guys for the quick reply. Here's another one for the memory banks... This was the Sunday before ASG... why did the AL have all these one game 'series' the Wednesday, the day after the ASG?

March 18th, 2010 at 6:25 pm

I vaguely remember something about the reasons for the one-game series, but I don't remember what it was. My first thought was that there were rainouts earlier in the season that were being made up, but two of the games were in domed stadiums. Also, they were all American League games. (And two pitchers who would be with the Phillies in 1993 got decisions that day. I had almost forgotten that Curt Schilling was once an Oriole.)

March 18th, 2010 at 6:27 pm

Gerry-

Thanks. I hadn't thought about that. I do wonder sometimes why we see 1-0 games or 20-18 games as often as we SEEM, too. Then, I realize it probably just SEEMS this way and confirmation bias is trickling in. Though, perhaps on a give day or in a given season, some of these scenarios are more likely, and for reasons outside of "one team/guy/whatever responding to the extreme effort of the other."