Game 7 — Probabilities — What We Know and Don’t Know
Hi folks. Harry Doyle here — with my pre-game analysis for the final match up the 2016 MLB World Series. Lots of graphs and maths. A bit long, but so is the season.
Note: You can do this yourself by hacking on Google Cloud, BigQuery and Datalab. Thanks to my friends at Sportradar for the help.
Note: My name is Eric Schmidt, not Harry Doyle. I am a developer advocate at Google. No, I am not that Eric Schmidt at Google. Reach me @notthateric. Yes, I get to analyze baseball data for work.
- Cubs are favored with a 53.49% win probability, Indians are the underdog at 48.78% win probability.
- Total runs are floating between 6.5 and 7.
- It will be between 69–70F at game time, with a 20% chance of rain. Wind will be 9–10MPH out of SSW.
- Progressive field sits dead north. It’s 410 ft. to center, 325 ft. to left, 325 ft. to right. The wind will help right handed opposite field hitters and left handed hitters who pull the ball.
- For some reason Google Maps thinks there is food on the pitcher’s mound. Perhaps this is the source of Aroldis Chapman’s power.
- The mark in the outfield is where Indians’ CF Naquin failed to communicate and make a routine catch in the 1st inning of game 6 costing the Indians two runs. This “error” came two batters after Tomlin hung a meat ball on 0–2 to Kris Byrant, which Bryant jacked for a home run. From there, Tomlin reduced his curve ball usage, getting into a major jam in the 3rd which led to a grand salami. 7–0 Cubs after 3.
Probabilities (Errors v. Mistakes)
The 0–2 homer given up by Tomlin was his fist of the season. It was Bryant’s second career home run conversion on 0–2.
Bryant saw 2,832 pitches in 2016. Bryant had only seen 1 other curve ball on 0–2 in in the same zone in 2016 (May 23 vs. Bumgarner). He struck out swinging.
The home run was Bryant’s first hit off of Tomlin in the series after seeing 11 previous pitches. The observed probability of this 0–2 outcome is very low; however, Tomlin made a huge mistake and paid the price. Was this mistake caused by too many free tacos? Perhaps a poor night of sleep? We don’t know. Baseball is unfair to players and fans sometimes.
The communication “error” by Naquin was not an official error as he nor the RF made an attempt to catch ball. Naquin played 821 innings in 2016, with 191 chances, and 2 errors — or a 0.01% error rate. However, if you look at other factors like his range factor (ability to track down balls) you can see his a weaker outfielder in comparison to the league, mainly because he takes poor routes to the ball.
Why was a sub-par 24 year old rookie outfielder playing in game 6 of the World Series on a hot night with the wind blowing dead over his head? He did hit .296 in the regular season and was hitting ok in the series — but if Indians’ coach Francona knew that Naquin would commit a mistake like that — would he have started him? We don’t know. Baseball is unfair to players and fans sometimes.
Whatever happens tonight, keep in mind that even when armed with tons of data, fancy tools, amazing talent, etc. that baseball is a game played by humans. Even if they don’t commit errors they make mistakes, in this final game mistakes will be costly. Let’s look at the humans for game 7.
“Klu-bot”
Tonight we will see the return of Indians starting pitcher Corey Kluber aka Klu-bot (Twitter and team meme for Corey’s performances). I analyzed his solid state performances in game 1 and game 4 AND talked about his ability to morph pitching strategies between games and within games.
For the Indians to win they need Corey Kluber to pitch like he has been, he needs to perform, and make few mistakes along the way. He needs to prevent runs, because we will see the return of Aroldis Chapman. If the Indians are losing when Chapman enters the game — game over.
Kluber has allowed 3 runs in the post-season. Zero in World Series game 1 at at home and 1 in World Series game 4 off of 5 hits at Chicago. He has not allowed a run at home in 19+ innings.
Kluber changed hit pitching strategy between game 1 and game 4. He shifted to more curveballs and featured his sinker more. This frustrated Cubs hitters and helped him stay ahead in the count more.
The Indians need Corey Kluber to show up and be Corey Kluber, specifically control the count, locate and produce outs. Let’s look at his competition Kyle Hendricks.
“The Professor”
Tonight the Cubs start Kyle Hendricks, who pitched an outstanding game 3 allowing zero runs on 6 hits. He did fall behind in the count at uncomfortable moments; however, he performed really well under the pressure making only three material mistakes (see graph left/below), but those mistakes did not result in any runs for the Indians.
You can read more about the Harry Doyle (pressure to control/execution) score here.
Hendricks is a well balanced and frustrating pitcher to face. He can induce high ground out percentages. He has strong control of his fastball and the ability to work the count with his curveball and get hitters to swing on junk.
If both Hendricks and Kluber show up we will have an awesome game to watch. If either of them falter you can be sure that the bull-pen for the leader will be activated early and with intention. The Indians will push Miller no matter the score and Chicago will most certainly bring Lester and Chapman
Miller
Miller has allowed 3 hits and 1 run in 5.1 innings pitched in the World Series. He produced 2 walks and 8 strike outs. He stays in control of the count and can perform well under pressure. He is 31. He has been around the league playing in Detroit, Boston, New York, Florida, and Baltimore and appeared in the post season in 2014 and 2015. He’s got grit.
Lester->Chapman
Jon Lester is sitting on enough rest after his game 5 start that he could come in for relief and with that bring some strong count control and a focus on strike outs.
Lester is a seasoned franchise starter, brining a ton of experience under pressure with 14 post season series under his belt. We could see him pitch 1–2 innings and then hand off to Chapman for an early close. Note: Arrieta said he is also ready to go after 5.2IP in his game 6 win last night.
Chapman
If nothing else is certain about game 7, we can be sure to see Aroldis Chapman come in for relief and or the close. He threw 20 pitches last night and tweaked his landing ankle on a play at first base.
In the World Series he has faced 23 batters, allowing… 2 hits and 1 run. No fancy graphs here. Chapman is just flat out intimidating. And his confidence level is off the charts. Last night when he gave up a hit to Ramirez in the 8th, you could see the look on his face deeply processing the “mistake” even though he was under very little pressure. He is a bull with tremendous power.
He hit 105MPH this season, read about that performance here.
What’s scary is this. Here is his Chapman's pitch speed distribution from the graph above.count 1053.000000
mean 99.737892
std 3.055205
min 86.000000
25% 99.000000
50% 100.000000
75% 102.000000
max 105.000000
Name: pitchSpeed, dtype: float64Scary --> When tailing his last 200 pitches, his mean is about .5 higher AND he has not yet hit back into the 105 mark AND he still has decent deviation.count 200.000000
mean 100.270000
std 2.541534
min 86.000000
25% 99.000000
50% 101.000000
75% 102.000000
max 104.000000
Name: pitchSpeed, dtype: float64This is just intimidating, especially if he is locating. If the Cubs are ahead and Chapman has the ball - game over.
Ok. I could go on and on about stats, probabilities, grit, and tacos — but as I said above baseball is unfair to players and fans sometimes. Tonight, someone is going to make a mistake, someone is going to break, someone is not going to be happy.
The Cubs are favored with a 53.49% win probability, Indians are the underdog at 48.78% win probability. What’s the margin of “mistake” on those lines???
Tomorrow, we can look at all of the data and tell you exactly why it happened — we are “just so smart with all of this data” <grin>. For now, let’s have some fun, bask in the excellence and be prepared for a great show.