Playoff Predictions

jtur88 · August 16th, 2008, 08:43 AM

In case you missed it, ESPN is running multiple simulations of all remaining games, and setting odds on which teams will make the playoffs. It in the last three columns in their standings chart:

ESPN - MLB Standings, Pro Baseball Standings, Major League Baseball Standings and Team Records

Acording to the simulations, only the LAD/Ariz and the CWS/Min race are not yet settled. Two other teams (Phil and StL) have a 1/3 to 1/4 chance of making the post season by going on a tear. The simulations, of course, are based on current rosters, and cannot take into account waiver acquisitions or injuries.

Grandstander · August 17th, 2008, 06:59 PM

Well, I wonder about the simulations that they are running. For example, either the Twins or the White Sox will obviously be in the post season, and the one which isn't is currently just a half game behind Boston for the wild card. However, Boston is listed as having an 84.3 % chance of reaching the post season, while the Twins are at 59.5 % and the White Sox are at 55.9 %. The Cardinals are but two games behind the Brewers for the wild card slot in the NL, but the Brewers are given a 71.9 % chance of reaching the post season while St. Louis is mired at 24.2 %. Somebody got some 'splain' to do.

Both Bay Area teams are at 0.1 %, even though the Giants trail Arizona by 11.5 games while the A's are 20.5 games behind the Angels. Of course that hardly matters since the actual chances of either team reaching the post season before the season started were 0.0 %.

jtur88 · August 17th, 2008, 10:57 PM

I suppose the simulations must be programmed to reflect the individual players "expected" perrormances in the remaining games. I would predict that the Cardinals bullpen is going to keep on blowing about half their save opportunities. The Cards have a slightly easier schedule---each has 15 remaining games against contenders, with the Cards an edge in home games.

Boston is quite a bit ahead of Chi/Min in the WC race, so the simulation showed neither Chi nor Min winning enough games to overtake Boston, so only one of those two teams is likely to be in the playoffs. Their combined chances are 117, which means one of them is sure to win the AL Cental, and a 17% chance that the other will obertake Boston.

If the A's are 0.06 and the Giants 0.14, they would both round off to 0.01. Maybe its relevant that Oakland is in tenth place for the wild card, a much better chance of making it than the 12-th place Giants.

Grandstander · August 18th, 2008, 06:13 AM

jtur88

Quote:

I suppose the simulations must be programmed to reflect the individual players "expected" perrormances in the remaining games. I would predict that the Cardinals bullpen is going to keep on blowing about half their save opportunities. The Cards have a slightly easier schedule---each has 15 remaining games against contenders, with the Cards an edge in home games.

If such subjective factors are going to be introduced, then this isn't really a math prediction model, it's merely one more opinion.

Quote:

Boston is quite a bit ahead of Chi/Min in the WC race

When did a half game become "quite a bit?"

jtur88 · August 18th, 2008, 11:20 AM

The White Sox and the Twins still have three games against each other. This would give the advantage to Boston, who are likely to win 2 games on those 3 dates, while the Central team demoted to the wild card race will likely win none or one. Not necessarily the case, but probably occurring in most sim trials. When I wrote that, the last day I had looked, the Red Sox had a 2.5 game lead.

The simulation has to be subjective, because the terms have to be pre-entered into the simulation, and their future cannot be objectively known. The Cardinals keep blowing saves because the only bullpen pitchers they have keep giving up runs in late innings.. The simulator will, obviously, predict that they will keep on doing that, unless it is subjectively tampered with by somebody who is convinced that they will suddently find the strike zone.

Simulations have limitations, which are obvious. Walking Pujols with first base open constitutes the manager running a bunch of simulations in his own mind, using the same data that mathemeticians use, and determining that one strategy will paly out better than the other in a majority of cases. Computer simulations merely allow one to do a lot more runs a lot faster, but with the same imperfect data.

I'm not vouching for the results. They are only as good as the data that has been fed into the simulators, which cannot be better than fairly good. But the device is useful for getting a rough idea about how a complex formula could play out.

jtur88 · September 27th, 2008, 07:28 AM

With two days left,. ESPN.s milliln-season simulator now lists Brewers chances over Mets at 77-23, and Twins chances over White Sox at 66-33.

August 16th, 2008, 08:43 AM	#1 (permalink)
jtur88 Hall of Famer Join Date: Nov 2006 Location: South Texas Posts: 7,857	Playoff Predictions In case you missed it, ESPN is running multiple simulations of all remaining games, and setting odds on which teams will make the playoffs. It in the last three columns in their standings chart: ESPN - MLB Standings, Pro Baseball Standings, Major League Baseball Standings and Team Records Acording to the simulations, only the LAD/Ariz and the CWS/Min race are not yet settled. Two other teams (Phil and StL) have a 1/3 to 1/4 chance of making the post season by going on a tear. The simulations, of course, are based on current rosters, and cannot take into account waiver acquisitions or injuries. __________________ ------------------ When people ask what I hope to see before I die, I answer that I've already seen too much. Last edited by jtur88; August 17th, 2008 at 12:31 PM.

August 17th, 2008, 10:57 PM	#3 (permalink)
jtur88 Hall of Famer Join Date: Nov 2006 Location: South Texas Posts: 7,857	I suppose the simulations must be programmed to reflect the individual players "expected" perrormances in the remaining games. I would predict that the Cardinals bullpen is going to keep on blowing about half their save opportunities. The Cards have a slightly easier schedule---each has 15 remaining games against contenders, with the Cards an edge in home games. Boston is quite a bit ahead of Chi/Min in the WC race, so the simulation showed neither Chi nor Min winning enough games to overtake Boston, so only one of those two teams is likely to be in the playoffs. Their combined chances are 117, which means one of them is sure to win the AL Cental, and a 17% chance that the other will obertake Boston. If the A's are 0.06 and the Giants 0.14, they would both round off to 0.01. Maybe its relevant that Oakland is in tenth place for the wild card, a much better chance of making it than the 12-th place Giants. __________________ ------------------ When people ask what I hope to see before I die, I answer that I've already seen too much. Last edited by jtur88; August 17th, 2008 at 11:08 PM.

August 18th, 2008, 11:20 AM	#5 (permalink)
jtur88 Hall of Famer Join Date: Nov 2006 Location: South Texas Posts: 7,857	The White Sox and the Twins still have three games against each other. This would give the advantage to Boston, who are likely to win 2 games on those 3 dates, while the Central team demoted to the wild card race will likely win none or one. Not necessarily the case, but probably occurring in most sim trials. When I wrote that, the last day I had looked, the Red Sox had a 2.5 game lead. The simulation has to be subjective, because the terms have to be pre-entered into the simulation, and their future cannot be objectively known. The Cardinals keep blowing saves because the only bullpen pitchers they have keep giving up runs in late innings.. The simulator will, obviously, predict that they will keep on doing that, unless it is subjectively tampered with by somebody who is convinced that they will suddently find the strike zone. Simulations have limitations, which are obvious. Walking Pujols with first base open constitutes the manager running a bunch of simulations in his own mind, using the same data that mathemeticians use, and determining that one strategy will paly out better than the other in a majority of cases. Computer simulations merely allow one to do a lot more runs a lot faster, but with the same imperfect data. I'm not vouching for the results. They are only as good as the data that has been fed into the simulators, which cannot be better than fairly good. But the device is useful for getting a rough idea about how a complex formula could play out. __________________ ------------------ When people ask what I hope to see before I die, I answer that I've already seen too much. Last edited by jtur88; August 18th, 2008 at 11:38 AM.

September 27th, 2008, 07:28 AM	#6 (permalink)
jtur88 Hall of Famer Join Date: Nov 2006 Location: South Texas Posts: 7,857	With two days left,. ESPN.s milliln-season simulator now lists Brewers chances over Mets at 77-23, and Twins chances over White Sox at 66-33. __________________ ------------------ When people ask what I hope to see before I die, I answer that I've already seen too much.

August 17th, 2008, 06:59 PM	#2 (permalink)
Grandstander Hall of Famer Join Date: Nov 2006 Location: Parts Unknown Northern California Posts: 6,611	Well, I wonder about the simulations that they are running. For example, either the Twins or the White Sox will obviously be in the post season, and the one which isn't is currently just a half game behind Boston for the wild card. However, Boston is listed as having an 84.3 % chance of reaching the post season, while the Twins are at 59.5 % and the White Sox are at 55.9 %. The Cardinals are but two games behind the Brewers for the wild card slot in the NL, but the Brewers are given a 71.9 % chance of reaching the post season while St. Louis is mired at 24.2 %. Somebody got some 'splain' to do. Both Bay Area teams are at 0.1 %, even though the Giants trail Arizona by 11.5 games while the A's are 20.5 games behind the Angels. Of course that hardly matters since the actual chances of either team reaching the post season before the season started were 0.0 %.