Applying xwOBA* to Virtually Real-Time Data | The Process Report

Applying xwOBA* to Virtually Real-Time Data

In my last piece I showed how I was calculating the productive value of a ball in play based on the exit velocity and launch angle for said ball, which I am calling Park-Adjusted Expected wOBA on Balls in Play or xwOBA*. I delved into a couple applications, but after fine-tuning a little more I think I’ve found a nice presentation style to showcase the players that look best and worst by this metric. Additionally, I have also incorporated non-ball in play events, such as strikeouts or walks or hit batsmen, into an additional metric I’m calling Total wOBA or twOBA* (tuba). Lastly, we can also adjust for volume by bringing plate appearances or batters faced into the equation. Weighted Runs Above Average, or wRAA, compares the player’s twOBA* to league average, and then multiplies by the number of plate appearances to more heavily reward/punish those that have been good/bad for longer. Being really good is great. Doing it everyday is better.

All data is current through April 15, 2017. First I want to start with just the players for the Rays, but then will expand out to look at each team, and lastly the best and worst players league-wide.

Starting with the ground that was trod in the last piece you can see how each player’s park-adjusted expected wOBA on balls in play matches up with their actual results to date. The +- column calculates how much a player has under or over performed his expectations so far. If guys like Xavier Cedeno or Austin Pruitt or Derek Norris or Tim Beckham continue to give up or hit balls at the velocity and launch angles that they have so far then you should expect them to post better results going forward. In the case of Cedeno that means a little better than average for a pitcher, while someone like Kiermaier might expect positive regression he would still find himself below the average. Context matters here so try to visualize how it all comes together. Matt Andriese has slightly over-performed, but a) it’s close, and b) he’s still giving up some of the hardest contact in the game. Jesus Sucre, on the other hand, has seen incredibly fortunate results in the very limited number of balls he has put in play compared to what should be expected. Blake Snell might have a ton of regression to go, however, he would still be a well better than average pitcher even once he gets there. I think you get it, by now.

As previously mentioned, twOBA* (tuba) takes the above xwOBA*, but also incorporates how often a batter walks, gets hit or strikes out. While many balls in play have a low expected production every strikeout is worth zero. These zeroes add up quickly, and you can see how they help someone like Archer go from a better than average contact manager to an absolute stud of a pitcher that strikes out plenty, but still faces a high number of batters. Jumbo Diaz has been a revelation as a no-cost signing during the time of Spring Training when most are sleeping one off in a hammock somewhere.

In fact, the Rays have a nice group of bullpen arms in Colome, Diaz and Hunter, but the aforementioned Cedeno, stalwart Erasmo Ramirez and newcomer Austin Pruitt have combined to form a very shaky bridge on the way to those very good backend arms. Archer has pitched like the ace that he is, and Odorizzi has mostly been fine as a second banana, though the hamstring is concerning. That leaves Snell floating around league average at the moment despite throwing a total of seven pitches to a left-handed batter this year. It will get better. Cobb was getting hit pretty hard even prior to today’s shellacking in Fenway, though the strikeout to walk looks very strong. Nominal fifth starter Matt Andriese has been about what you should expect from putting a swingman into a rotation spot.

Switching over to the batters you can see that Logan Morrison might just be the Most Valuable Player for the Rays through the first two weeks even if he isn’t quite MVP of the league. Dickerson has been a rock, and again, this doesn’t include the homer he hit today. Longoria continues to hit the ball hard when he makes contact, but his 18 strikeouts have him virtually pacing the league. The walks are nice to see, and I have no problem with him trading contact for power, but at some point he’s going to need to actually make contact. Rickie Weeks has hit the ball hard, but has also struckout out in half of his plate appearances leaving him right around the average, overall.

Tim Beckham and Kevin Kiermaier bring up the rear as two of the worst hitters in the entire league. Neither is hitting the ball hard and/or at ideal angles, while striking out by the boatload. It’s nice to see Kiermaier walking a little more, but the ultra patience isn’t doing him any favors when it comes to picking out a good pitch to take a whack at. Derek Norris is in a similar boat, and Brad Miller mirrors Longo’s non-ball in play results, but with more good than great expected results on his balls in play like you see from the Franchise. Add it all up and you can see why the team has lost five of their last six:

The pitchers have collectively allowed a twOBA* of .370, which places them 23rd in the league alongside such juggernauts as the Twins and Orioles. The hitters only have one neighbor, the Padres, because they sit in dead last. They’ve hit the ball well when they can get it in play as their twOBA* ranks around 20th, not great, but not last place. The problem has been the most strikeouts as a team in the league. They have walked around 10% of the time, which is astounding as a team, but they’ve also struck out over a quarter of the time. That’s no bueno.

Add up both sides of this and you’ll find the Rays with a second worst in the league wRAA of -17.3, which beats out only the Diamondbacks. However, the number one team so far is the Reds, so it should be said again that it is extremely early. It won’t stay that way for long though, if the team continues to play like this. In an ultra competitive division they cannot afford to fall behind, and should be incredibly fortunate for the banked wins over the first week-plus of the season. Moving on I also want to show who have been the best and worst players so far this year. You have seen the Rays figures above, so this should give you some context toward where they fit on this spectrum. Some of the players even show up on the lists, which can be both good, and bad depending on the list. Without further comment, let’s start with the good and bad pitchers, respectively, before doing the same with the batters:




  1. Shane wrote:

    Jason can you share this xml?

    • Jason Hanselman wrote:

      Nah, unfortunately I’ve worked really hard on this stuff and want to keep it proprietary. Normally I don’t have a problem sharing stuff, but this and the matchup tool need to stay in house. Thanks for your interest, and keep coming back for more updates. Especially my twitter @sandykazmir as I post the highlights after each update.

Leave a Reply

#layout { padding-left:20px; }