Last season two-thirds of our games were close (within three runs). Our overall percentage of close games in 2011 stands at 57%. This season has been the tale of two boroughs. In Manhattan (Heckscher, Great Lawn and 54th street) our record is 22 close games and 8 games decided by four or more runs - 73% close games. We played 15 games in Riverdale. Four were close and eleven were not. That's just 26% close games.
To add to the mystery, several of the Riverdale games were shared victory blow-outs. Team A would beat team B 7-1 and then, in the second game, team B would beat team A 11-3. My current theory is two-fold.. in a park with a reachable fence multiple run innnings are more likely, giving a team a greater chance of racking up a big lead. Secondly, if total runs scored is higher, the case in Riverdale, then a 5-2 game in Heckscher (considered "close" because it is within 3 runs) is not considered close if twice as many runs are scored and each team has the same percentage of runs in Riverdale, say, 10-4 (considered a blow-out).
Since close games are our goal, we need to give some thought as to whether we should ever play in a reachable-fence ballpark.
4 comments:
meh..........
I'm very sympathetic to the point that close low-scoring games are more fun.
But as Havelock admits in the post, the metric used for "close games" in the post is not very good, and it has a bias against high-scoring games that still weren't necessarily blowouts, they could have had several lead changes, and the losing team could have been just one big inning away.
A better metric that would still be simple to use and explain would be to declare a game "close" if it was decided by 3 or fewer runs -OR- if the losing team scored at least one-third of the total runs. (By this metric 10-4 wouldn't be close enough, but 8-4 or 10-5 would be.)
A really good metric would look at a generic SFLOI's teams chances of scoring a certain number of runs per inning in a certain park. So you'd count all the innings played, sort them by number of runs, and come up with a probability curve for each number of runs per inning, i.e. for Hecksher this might be:
0 runs - 40%
1 - 20%
2 - 15%
3 - 15%
4 - 5%
5 - 5%
Then if you a given game has a 3-run differential, you can measure what the chances are that the losing time could tie or win given one additional inning (for this example, chance would be about 25%, though I'm simplifying and overestimating slightly because really the other team gets a projected inning too - exact answer left as an exercise to the student, as good textbooks always say).
This percentage chance that one additional inning would change the outcome could be compared across parks, seasons, pitchers, whatever. Ian just might need a better database to compute it ;)
Bill
Havelock replies: Bill's point is a good one. It is so much more likely that a team will come back from a five run deficit in Riverdale than at Heckscher, so a five run lead at Riverdale is like a three run lead at Heckscher. One might further argue that having a team make a big comeback (as Garcia's team did last December - I think it was 13 runs) makes for an even more compelling game than a low-scoring contest. Still, the hope that a team could win a game in one inning by loading the bases and clearing them once, rather than batting around the line-up, is so much more palpable. Also, larger differences give the impression that the sides were unfair. As the sides-make,r I confess to a little vanity about this.
Surprisingly, we were 13 of 20 (65%) close games at East 10th Street. I assumed the terrible fielding surface, and the absence of some of our best fielders by that time of year, would make things worse.
Post a Comment