Update Post Game: Astros won this game and play the Yankees in ALCS. A data flow error has been discovered in Lineup tier calculations for all playoff games this year. HOU below has a maxed out tier 4.00 lineup at +27 tonight and Ray’s lineup was around -0.5 tier number, not -1.59 shown in below table.
As games completed standard deviations were calculated using past playoff lineups when they should have been using the end of September regular season set of lineups. The error got worse as more playoff games polluted the lineup dataset, by raising league averages significantly. This depressed both home and away L tier numbers. The resultant simulation % estimation didn’t change much in the below game however. League championship tier numbers will be calculated the way they should.
When this season is in the books, when retrosheet.org releases event data, all this gets recalculated anyway for the historical dataset. Right now we’re working off current year data which uses a different set of scripts to keep track of everything.
End of update
ALDS Game 5 tonight. Washington beat Dodgers last night which makes NLCS interesting.
TBA HOU 10_10_7:07_PM
| WAA | Vegas | TC Sim | EV | L | S | R | |
|---|---|---|---|---|---|---|---|
| TBA | 30 | 0.294 | 0.340 | X | -1.59 | 1.61 | 1.72 |
| HOU | 52 | 0.737 | 0.660 | X | 2.75 | 4.00 | 3.46 |
Vegas has Houston almost 3-1 favorites, TC Sim almost 2-1. Lineujp value taken from last game. HOU usually fields a max 4.00 lineup so that and home field advantage should push things up for them with TC Sim.
| Starters | WAA | WinPct | IP | Tier |
|---|---|---|---|---|
| Tyler_Glasnow_TBA | 3.74 | 0.777 | 60.7 | 1.61 |
| Gerrit_Cole_HOU | 9.64 | 0.704 | 212.3 | 4.00 |
This model has Cole ranked #2 behind Verlander. WAR has him ranked #10. Glasnow has a very high Win Percentage based on only 60 innings pitched. Tiering is done on raw WAA not rates like Win Percentage.
That’s all for now. Next up some historical playoff numbers to put these two series into perspective. Until then ….