Back to the future we travel to the year 2020 when MLB starts its season tomorrow July 23 with two night games and then the next day the rest of the league starts their season. The season ends September 27 with only 60 games scheduled for each of the 30 teams.
Normally in April we do a playoff horse race using 3 year splits based on opening day rosters of each team when all rosters are known. Rosters are divided into Starters, Relief, Hitters, and then added to make a Total which gets sorted from highest to lowest.
Although most teams in the top third of these lists either make the playoffs or are competitive throughout the season and the bottom third, not so much, there are always exceptions. Cubs and Houston had the two lowest totals at the beginning of 2015 but both teams had young talent with little to no accrued WAA value. The next few years both these teams rose to the top when their young guys started racking up value in three year splits while their teams dominated the league.
This model needs a month of current year data in order to properly rank players. It needs at least 45 days, or 25% of a season, in order to simulate and handicap games. Think of a baseball season like a marathon or the Indy 500. At the start runners and cars are bunched together and as time goes the field spreads out to reveal front runners, middle of the road, and bringing up the rear participants. Even win/loss records of most MLB teams are still bunched up closely by mid May.
Vegas however starts handicapping on day one of a season and probably even bets meaningless pre-season games. In order to handicap on day one the simulator for this model would have to depend upon player data from previous seasons. The simulator currently doesn’t work that way but it is something that should be looked into and perhaps a shortened season like this will provide some insight into how that can be done.
Simulation draws data from complete seasons in the past from teams limited to 25 player rosters. This season will have 30 player rosters and other roster shenanigans that may invalidate comparison with team valuations in games from the past. Currently expanded rosters is a problem for the simulator in Septembers for many reasons that we won’t get into here but is still a work in progress to figure out had this been a normal season. This entire shortened season is similar to a normal September expanded rosters.
A normal baseball season is a 6 month marathon. This 2 month sprint for each team requires different managing strategies. It’s like a marathon race requiring runners to only run 10 miles. Runners will run differently if they know they don’t have to run the entire distance and baseball players will know they can rest a few days off without worrying about IL/DL rules.
In April this model fixes broken scripts from data sources that change their APIs. This will be the first year this model estimates current year rosters which will be a lot more accurate than past years when they were being downloaded from Wikipedia. Wikipedia was a good source but sometimes lagged a few days.
We’ll be getting detailed box scores from the mlb.com api which will allow for a much more accurate valuation of relief squads than in past seasons. MLB.com switched their API from XML to JSON this season. The code to decipher JSON was being worked on in March when everything in the world went haywire.
Like in past Aprils, after analyzing all opening day rosters there really isn’t much to discuss for the next few weeks when we get a handle on what players and teams are breaking out this year . Right now I don’t even know what teams are in the Cubs’ division or even how playoff matchups get chosen.
Speaking of Cubs. While testing the JSON API from mlb.com using a recent Cubs White Sox game it appears the Cubs are 8-14 and the White Sox are 12-7 before Cubs lost that game. When this season starts and we have full rosters we’ll cover both White Sox and Cubs. We knew White Sox would be contenders this season but not sure about Cubs. Can’t really tell without knowing exactly who is on each team however which we’ll know in a few days.
The 1984 Cubs season will continue unabated until the very tragic end late in October. Meanwhile, Cubs spilt a 4 game set with the last place Giants losing money for Cubs bettors and it’s on to Philadelphia for a three game series with Rick Sutcliffe on the mound tomorrow. Until then …