The sidebar on the left now includes the aggregate Leveraged Win Values of each batter, including all games between 21st April and 8th May. I shall be adding subsequent games and updating the list regularly. It ought to be pointed out that Adam LaRoche's position at the top of the list was not affected by his home run in the 9th inning in the game of 8th May. He only about doubled his total for the season during the period!
The numbers basically capture what percentage of a win each batter has contributed given what he did in a plate appearance in the particular context of the game at that exact moment. I've taken to calling it 'clutch' in the short summaries heading my individual game comments. Thus, LaRoche's accumulated plate appearances during the period covered were worth an entire win, and then some, in themselves.
The most glaring matter requiring comment is Danny Espinosa's situation. Espinosa is clearly well adrift of the impact of most regulars. I confess to not understanding Davey Johnson's persistence in putting him any higher than seventh in the lineup. If it's a case of not having too many left-handed hitters in succession, placing Wilson Ramos behind Chad Tracy (where Espinosa batted on Tuesday night) seems to make far more sense, based on these numbers.
I'd also just like to note that Tracy has been a pleasant surprise so far.
Showing posts with label Leverage. Show all posts
Showing posts with label Leverage. Show all posts
Thursday, 10 May 2012
Saturday, 18 July 2009
Zambrano, Ramírez, Fontenot, Bradley
Messing around with the leverage index during yesterday's game and comparing it with the win probability reveals a disagreement. My method of multiplying the 'win value' of an event by the leverage does not have the same results as the win probability effect. Does that mean I'm doing it wrong? Maybe it does.
I was a surprised at how the Ramirez homer scored so low in my method, not even as high as the Bradley double play. Curiously, though, it was the point at which the balance of the game shifted towards the Cubs permanently. It may be,in this case, that I'm rating double plays too highly. Are they the same as two outs, or the same as an out plus a caught stealing?
______________
* I'm not sure I call this a double play. I've got two events, the Zimmerman strike out and Morgan being thrown out at third.
Player/Event/Inning WPA WV*LI
Zambrano/double/2nd 18% 13.9%
Ramirez/homer/3rd 11 4.9
Zimmerman/DP/1st 10 11.9*
Harris/BB/1st 9 7.6
Fontenot/double/2nd 9 8.6
Bradley/GDP/8th 5 7.6
I was a surprised at how the Ramirez homer scored so low in my method, not even as high as the Bradley double play. Curiously, though, it was the point at which the balance of the game shifted towards the Cubs permanently. It may be,in this case, that I'm rating double plays too highly. Are they the same as two outs, or the same as an out plus a caught stealing?
______________
* I'm not sure I call this a double play. I've got two events, the Zimmerman strike out and Morgan being thrown out at third.
Sunday, 5 July 2009
Adventures in Independence Day Leverage
Yesterday's game is what baseball is all about, for me. The score wasn't too high, the game was dominated by pitching, and neither of the two contending teams gave up on the game.
During the game, I performed an exercise of calculating Leverage x Win Value in order to create a number that would represent the value of each event to winning the game. Instead of using the run value of each event, I used the win value as detailed in Tango, Litchtman, Dolphin's The Book. (Run values are used in the famous linear weights formula devised by Thorn and Palmer in The Hidden Game of Baseball.)
It was an interesting exercise, because it highlighted how a well-timed single can be far more valuable than a home run, and why on-base percentage is more valuable than slugging.
Dunn's home run in the bottom of the 7th, leading off, was worth 14.76 per cent of a win. But Zimmerman's single in the bottom of the 8th was worth 24.78 per cent of a win. Even Dunn's single in the next plate appearance was worth more than Dunn's homer, at 15.96 per cent of a win. By the time we get to Willingham's single, the cumulative effects of the runs scored up to that point have substantially reduced the leverage, and he only gets 4.62 per cent of a win.
The reason is the multiplier effect of leverage. The base runners and the differential in score added up to making those 8th-inning situations of greater significance. A typical linear-weights formula wouldn't have captured this, and just awarded Dunn a 1.44 runs for his home run and 0.44 runs for each of the three singles. Furthermore, it also accounts better for the effect of piling up baserunners, each additional runner pushes those already on base closer to home plate. Bard would not have scored from first on any of the Zimmerman-Dunn-Willingham hits; he had to get to second.
On another matter, anyone following this game would think bunting was a brilliant strategy. A bunt made the Braves' first run possible, and a bunt arguably resulted in the Nationals' being in a position to tie the game quickly, although in the subsequent walk meant the bunt really had no effect. I'm not trying to argue the point, just observing that in this game bunting worked.
Finally, I'm awarding a Hero of the Day to MacDougal, for defending the lead in the top of the 9th.
During the game, I performed an exercise of calculating Leverage x Win Value in order to create a number that would represent the value of each event to winning the game. Instead of using the run value of each event, I used the win value as detailed in Tango, Litchtman, Dolphin's The Book. (Run values are used in the famous linear weights formula devised by Thorn and Palmer in The Hidden Game of Baseball.)
It was an interesting exercise, because it highlighted how a well-timed single can be far more valuable than a home run, and why on-base percentage is more valuable than slugging.
Dunn's home run in the bottom of the 7th, leading off, was worth 14.76 per cent of a win. But Zimmerman's single in the bottom of the 8th was worth 24.78 per cent of a win. Even Dunn's single in the next plate appearance was worth more than Dunn's homer, at 15.96 per cent of a win. By the time we get to Willingham's single, the cumulative effects of the runs scored up to that point have substantially reduced the leverage, and he only gets 4.62 per cent of a win.
The reason is the multiplier effect of leverage. The base runners and the differential in score added up to making those 8th-inning situations of greater significance. A typical linear-weights formula wouldn't have captured this, and just awarded Dunn a 1.44 runs for his home run and 0.44 runs for each of the three singles. Furthermore, it also accounts better for the effect of piling up baserunners, each additional runner pushes those already on base closer to home plate. Bard would not have scored from first on any of the Zimmerman-Dunn-Willingham hits; he had to get to second.
On another matter, anyone following this game would think bunting was a brilliant strategy. A bunt made the Braves' first run possible, and a bunt arguably resulted in the Nationals' being in a position to tie the game quickly, although in the subsequent walk meant the bunt really had no effect. I'm not trying to argue the point, just observing that in this game bunting worked.
Finally, I'm awarding a Hero of the Day to MacDougal, for defending the lead in the top of the 9th.
Subscribe to:
Comments (Atom)