coolstandings.com

[cruiserx43]

For those of you who are interested and I didn't see it posted yet so i apologize if I missed it but there is a link through ESPN that sims the remainder of the season and indicates the Brewers actually have a better chance to reach the playoffs than the Cards.

 

http://www.coolstandings.com/baseball_standings.asp?i=1

[1992casey]

I fixed your link. http://forum.brewerfan.net/images/smilies/smile.gif

 

I first looked at this last night because I was curious about the validity of ESPN's playoff odds. According to its FAQ, it's looked like these people have done their homework. One of our competent statheads could probably scrutinize it better than I can, of course.

 

The only thing I noticed last night was there was a discrepancy between the odds listed on ESPN vs. coolstandings.com. They're the same this morning. I'd assume that ESPN simply updates the odds on their site later.

[OOOverbay]

Sooo, did they just rip off BP's constantly updated playoff odds?

[1992casey]

No, it doesn't look like a rip-off. Numbers are different. However, I'd bet that when designing their system, coolstandings.com would have been aware of any public information about Baseball Prospectus' procedure.

 

Baseball Prospectus explains what they do on each of their playoff odds pages:

• Playoff Odds Report
• Playoff Odds Report (ELO adjusted)
• Playoff Odds Report (PECOTA adjusted)

The good news would be that at a glance, I don't see remarkable dissimilarities between the results reported on the two sites.

[cruiserx43]

Thanks for fixing the link, if nothing else it's something new to look at and ponder.

[1992casey]

Interesting stuff in the playoff odds links listed above. I don't have the exact numbers, but I remember that before yesterday's game, the Brewers' odds varied between about 48-52%. Today, they range from 54-59%. This has to account for the addition of Sabathia. I don't think one win and a minor standings shift in early July would cause such a dramatic change.

[zzzmanwitz]

I'm pretty sure none of the BP expectations account for Sabathia's signing.

[Brawndo the Thirst Mutilat]

None of any of these sims are accounting for the addition of Sabathia, from what I can discern. The BP versions are to be preferred, at least in my view, because they are using third-order winning % instead of just pythagorean expectations.

Long story short, coolstandings is only concerned with how many runs a team has scored and how many runs a team has allowed, and whether the team is the home team or the visitor. BP is actually looking at the constituent building blocks of run scoring and run prevention, so they should be getting a more accurate picture of how strong the teams actually are when simulating. Now, they are using their own black-box EQA equations to do that, and even though they make such a big secret out of those, they aren't quite as good as Base Runs and Linear Weights (which is what you wish someone was using for this type of deal), but it's still a darn sight better than just assuming that RS and RA tell you exactly how strong a team is.

Why? Briefly, because how your hits and hits allowed are distributed (i.e. between innings) affects how many runs you score, and this is no less subject to random variation than how the runs you score and allow are distributed between your games. This is why it's better to use individual offensive events and defensive events as BP does on their adjusted standings.

The different versions BP offers appear to simply regress this season's results so far to different targets -- the mean (.500 winning %) vs. pre-season PECOTA standings (Gallardo's injury wouldn't affect this, nor would the addition of Sabathia) vs. something called ELO.

(Not sure what's different about ELO -- the substantive explanation isn't free, and I'll be darned if I'm going to pay them for their mostly inferior-to-what-is-freely-available content).

Not sure how helpful I've been, post follow-up questions if you have any. This is hard stuff to figure out, in part because BP is pretty bad about inventing new language and not offering you the cipher, at least in an easy-to-find, you-don't-have-to-pay kind of way.

[rluzinski]

BP is not going to adjust for the acquisition of Sabathia. They also didn't adjust their pre-season projections for Gallardo's loss. I also feel that they don't take into consideration, perfomrances from prior years enough (they don't regress enough towards their pre-season projection.

 

The right way to do it is to estimate playing time for every player going forward and to update every player's projections to include their performance for this year. Run a monte carlo simulation for the rest of the year and there you have it.

 

MGL at insidethebook.com did just that about 6 weeks ago (when the Brewers were under .500) but he hasn't updated it since.

[Brawndo the Thirst Mutilat]

Mr. Kalk, over at Brewcrewball, does the math on how much the addition of Sabathia affects the Brewers' playoff chances using BP's methodology.

[monty57]

Mr. Kalk, over at Brewcrewball, does the math on how much the addition of Sabathia affects the Brewers' playoff chances using BP's methodology.

Since I don't have a login over there, I'll ask the question here: "Does this take into account the remaining schedule?"

It looks like they're using home vs. away, but what seems to be more important is that the Brewers, Cubs and Cardinals play against each other a lot over the remainder of the season. Two additional wins against the Cubs / Cardinals is more important than two wins against a non-playoff contender. Plus, the Cubs and Cardinals beating each other up is very important to our playoff picture.

Edit: After reading the entire post, it looks like Russ beat me to that question