Our Playoff Odds Have a New Look
The Playoff Odds page looks different! The playoff-odds data and simulation method remains the same; however, we have revamped our reporting page to make it easier to understand and more powerful.
The most noticeable changes are the table layout and the mobile layout. We’ve tried to make it easier to understand what the columns mean for users who are new to the site. The goal of the mobile layout is to allow users to reach the most important information more quickly. Every column on the desktop page is viewable on the mobile layout by clicking the “Full” button.
We have also added a few new display features. The detailed round-by-round postseason odds (winning the ALDS or the NLCS, etc.) are hidden by default during the regular season. They remain present, but are collapsed in a Postseason column. The default view will switch once the Postseason begins, and most of the regular-season projections will be initially hidden. The updated calendar widget allows one to view playoff odds on any day dating back to 2014. We have also added new quick dates in the drop-down menu next to the widget, in order get to the current day’s playoff odds or preseason odds.
Strength of Schedule
Another addition is the Projected Strength of Schedule (SOS). This isn’t the backwards-looking strength of schedule with which you might be familiar from NFL tiebreakers or the really old college football BCS ranking calculations. This looks ahead at the team’s remaining schedule to determine how difficult it will be. Technically, this is the Average Projected Opponent Winning Percentage for Remaining Games, but that’s a mouthful of words.
We are calculating the opponent’s winning percentage using the same engine we use to calculate the Playoff Odds. Changing the projection mode to Season-to-Date or Coin Flip will affect the SOS metric. SOS will center around .500, and early in the season there won’t be too much deviation from it since there are many games remaining against various teams.
One reason we decided to include SOS is that it can be useful to explain why certain teams are projected to do better in the future, especially in the last few weeks of the season.
I would caution against using this for determining how balanced a schedule is. If you are comparing a team projected to do very well (let’s say the Indians) to a team projected to do poorly (like maybe the White Sox), we aren’t removing the common games between the two teams, so the White Sox appear to have a tougher schedule because they have to play the Indians (having a high opponent winning percent), while the Indians in those games have to play the White Sox (low opponent winning percent).
Displaying Changes in Playoff Odds Over Time
The playoff-odds graphs do a good job of showing evolution of the baseball season and, in particular, the change in playoff odds over time. We thought we’d include a way to display those changes in our tables. If you choose “Changes” as a “Display Option”, the tables will show the difference in the playoff odds or projected wins between two dates. It’s also possible to compare any two dates for which we have playoff odds.
As usual, please let us know if you have any problems or find any bugs.
I build things here.
Why did the Mets’ probability of making the playoffs go down over the past few days?
Sean just does the code (to my understanding) and can’t really answer questions about why the projections have the Mets’ odds lower.
Moreover, they’re only down 3% from the other day, which happens to coincide with the fact that they sent a player with a 215 wRC+ down to the minors around the same time.
Yeah, you’d have to look at all the factors that go into the playoff odds to figure out specific reasons as to why a team went up or down. There are always roster changes, how other teams are doing, random variation, etc.
Might have been the demotion. Would be interesting to gauge the sensitivity of sending down such a player (if that is the reason why the odds went down). If so – he’s worth 3%? That implies a large WAR.
My guess is it’s the d’Arnarud injury – he was projected for a decent WAR and now they prob have Nido in the projections which is a big change.
Much improved!
Is ROS w% based on the SOS, or what the team would be expected to do against a .500 schedule?
Short answer, SOS, and it’s the same as it’s always been. The ROS W% is basically the average winning percentage that the simulation ran, so switching the projection mode will affect that.
Why do the Twins have 11% chance of winning the division and a 19% chance of a wild card, but only a 0.7% chance of winning the World Series? That implies they only have ~30% chance of winning each best-of-seven series. That seems pretty harsh.
Seems about right to me
Even if its 50% for each game/series = 6.25%
Mulitply that by the 19% WC chance = 1.2%
You have to also add in the 11.9% chance of winning the division times 0.5*3. That gives an answer of 2.7% = (.119 * 0.5^3 + .197 * 0.5^4).
My previous numbers above were from April 12th. Now that they’ve upgraded the Twins WS odds to 1.1% on April 13th, it means they’d have a 38.5% chance to win each best-of-seven series. That I can believe 🙂
So the Dodgers are going to play 25 games over 500 the rest of the way and the Snakes just 2 games over..??
This is a fantastic upgrade. Much added and nothing taken away. Keep up the good work!
This is an awesome update!
Does your model take into account the fact that if a team outperforms and makes the playoffs, it is likely a better true talent team than was originally projected?
It uses player projections which are updated daily based on performance.
I know. I’m asking if the current post-season success odds uses a model which accounts for the fact that if a team outperforms to make the playoffs, the current projections for that team are likely too pessimistic.
We can estimate this ahead of time and make adjustments so that we do not underestimate the chances of post-season success for currently-projected “bad” teams.