Master Index of Archived Threads


FanGraphs Playoff Odds

Benjamin Grimm Jun 15 2021 10:40 AM

I'll update this from time to time, if I remember to! As of today, FanGraphs has the Mets projected to 90.6 wins (this number is not likely to be exactly correct) and winning the division by nine games over the Phillies, who will win 81.6 games. The Mets are at 81.4% to win the NL East, 4.3% to win the Wild Card (which adds up to 85.7% to make the postseason) and the odds of them winning the World Series is 12.7%. The Phillies are at 8.6% to win the division. The Braves are at 8.3%, Nationals at 1.6%, and the Marlins at 0.1%. The only team with better World Series chances than the Mets are the Dodgers, at 17.1%. The AL teams with the best chances are the Astros (10.5%) and the White Sox (10.4%). If things turn out as FanGraphs predicts, the Mets and Brewers would meet in the NLDS. The two wild card teams would be the Padres and Giants, who would play for the right to face the Dodgers. In the American League, the Wild Card game would be between the Red Sox and the A's, with the winner playing the Rays in the ALDS. The other ALDS matchup would feature the Astros and White Sox.

Edgy MD Jun 15 2021 11:32 AM Re: FanGraphs Playoff Odds

It's weird how different sites have such varying numbers. BB-R.com has them at 55.6% to make postseason, 4.4% to win World Series . I generally respect FanGraph's math more, but seeing them more optimistic to the tune of 30.1 and 8.3 points, respectively, tells me that there are at least a few arbitrary factors folks are working with. They have nearly triple the World Championship optimism that BB-R has!

Benjamin Grimm Jun 15 2021 11:45 AM Re: FanGraphs Playoff Odds

I saw a disclaimer on BB-Ref's site saying that Baseball Prospectus and FanGraphs have more sophisticated algorithms. I don't know if BP is still doing playoff odds. The last few times I visited the site I couldn't find it anywhere. A Google search points to a "legacy" subdomain with the odds from 2019. https://legacy.baseballprospectus.com/odds/

dinosaur jesus Jun 15 2021 12:22 PM Re: FanGraphs Playoff Odds

I've been comparing the Baseball Reference and FanGraphs projections for a while now. It is interesting how different they are. I think what's going on with Baseball Reference is that to get a good sample size they're looking at a team's performance over the last 100 games. As of today, the Mets are two games over .500 for that span. And Baseball Reference projects them to be two games over .500 for the rest of the season. That led to some screwy results earlier in the season, though. There was a point when the Mets had actually performed better than BR's original projection, but their playoff odds had gone down. And I think that's because the more games they played, the more games from the end of 2019, when they were really hot, dropped out of the sample. So if they won, say, three out of five, but a four out of five stretch from 2019 dropped out, the sample, and thus the projection, got worse In any case, if they keep playing well, and more and more of that lousy 2020 season drops out of the sample, you're probably going to see those projections jump.

Edgy MD Jun 15 2021 12:28 PM Re: FanGraphs Playoff Odds

It's good to know Aaron Altherr is still contributing.

Benjamin Grimm Jun 15 2021 12:39 PM Re: FanGraphs Playoff Odds

It seems weird that they would base this on anything to do with the 2019 Mets, or even the 2020 Mets. The Mets playoff odds are being hurt because of guys like Steven Matz, Michael Wacha, and Rick Porcello?

batmagadanleadoff Jun 15 2021 12:51 PM Re: FanGraphs Playoff Odds

Benjamin Grimm wrote: It seems weird that they would base this on anything to do with the 2019 Mets, or even the 2020 Mets. The Mets playoff odds are being hurt because of guys like Steven Matz, Michael Wacha, and Rick Porcello? That's the first thing I thought of as soon as I read DJ's post. I can't come up with a single reason for using that methodology.

batmagadanleadoff Jun 15 2021 12:54 PM Re: FanGraphs Playoff Odds

Benjamin Grimm wrote: It seems weird that they would base this on anything to do with the 2019 Mets, or even the 2020 Mets. The Mets playoff odds are being hurt because of guys like Steven Matz, Michael Wacha, and Rick Porcello? That's the first thing I thought of as soon as I read DJ's post. I can't come up with a single reason for using that methodology. I figured that what the most sophisticated prognostications were doing was to project each individual player's season before the season started and base their playoff odds on that data.

Ceetar Jun 15 2021 01:19 PM Re: FanGraphs Playoff Odds

Benjamin Grimm wrote: It seems weird that they would base this on anything to do with the 2019 Mets, or even the 2020 Mets. The Mets playoff odds are being hurt because of guys like Steven Matz, Michael Wacha, and Rick Porcello? That's the first thing I thought of as soon as I read DJ's post. I can't come up with a single reason for using that methodology. ..playoff odds are a lot more meaningful and interesting with 60 games to go in the season, and using the first part of the season as a proxy for how good they've been, and adjusting based on that is certainly..something. It's not really a projections system, and isn't, for example, factoring in expected IP From Noah Syndergaard. While some teams experience significant turnover from year to year, a roster is going to still be somewhat consistent. This certainly means early early numbers are worse for this particular system, but that's fine, after game 10 you're better off just looking at projections.

dinosaur jesus Jun 15 2021 02:41 PM Re: FanGraphs Playoff Odds

No, it really doesn't make much sense to include information from 2019 and that weird year of 2020. But in Baseball Reference's defense, looking at the last 100 games would make more a lot more sense at a more normal time. There was probably no good way to project this season, no matter what approach you took. Even if you do it starting with player projections, what are you basing those on? Stats from two years ago? From last year, when they only played 60 games and a lot of them seemed to be pretty checked out? Just to make sure I wasn't making up what I said about Baseball Reference, I checked their projections page again. Here's their explanation: "MLB playoff odds are based on 1000 simulations of the rest of the season and playoffs. The team's estimated quality is determined by their performance over their last 100 regular season games (even if it spans seasons) and includes a regression to the mean factor." The obvious flaw there is that we don't really know what the mean is that we're supposed to be regressing to.

batmagadanleadoff Jun 15 2021 03:30 PM Re: FanGraphs Playoff Odds

Are they saying how they're using 2019 data? Maybe for the Mets, they're using Lindor's 2019 data, but not Matz's?

dinosaur jesus Jun 15 2021 03:47 PM Re: FanGraphs Playoff Odds

What I quoted is all I know. They've got an algorithm, but algorithm is just an anagram for logarithm to me, and I don't understand logarithms either.

Frayed Knot Jun 15 2021 03:51 PM Re: FanGraphs Playoff Odds

Talk to me about playoff odds in September

Lefty Specialist Jun 15 2021 04:35 PM Re: FanGraphs Playoff Odds

Frayed Knot wrote: Talk to me about playoff odds in September Yes. This is utterly meaningless right now.

Edgy MD Jun 15 2021 08:22 PM Re: FanGraphs Playoff Odds

Says the guy without a hardcore gambling problem.

Centerfield Jun 15 2021 08:45 PM Re: FanGraphs Playoff Odds

Frayed Knot wrote: Talk to me about playoff odds in September And only with less than 17 games left in the season.

Benjamin Grimm Aug 05 2021 07:22 AM Re: FanGraphs Playoff Odds

Mets division odds are down to 49.4%, plus 1.9% for the wild card, and 4.4% to win the World Series. Phillies are at 29.9% to win the division, Braves at 20.6%. The remaining 0.1% goes to the Nationals.