It is October, the Indians have once again made the postseason but failed to bring home the coveted World Series title and the ‘curse’ moves on another year. Many of the Indians players didn’t play to their potential in the 3 games of the ALDS, which is always possible in the small samples sizes of the postseason. But when your 3 to 9 hitters average .101 (7-69) and you ‘top’ 3 relievers combine for an ERA of 24 and a WHIP of 4.67 you know you are never going to win any series.
There were warning signs before the postseason even started with the bullpen, but the complete disappearance of the bats was majorly disappointing to Indians fans. The focal point of this disappointment was Jose Ramirez, he went 0-11 with 1 walk across the 3 games. If you haven’t been watching the Indians this season — I don’t blame you the AL Central was poor — and just saw the postseason games; then you looked at his stats for the season (1.33 BB/K, 147 wRC+, 8.1fWAR, 7.9 bWAR) you might have said “it isn’t unreasonable for him to have a couple of bad games Russell, stop with the hyperbole”. But was just those 3 games he had a poor late August and September that stretched to nearly 40 games.
The graph above shows Jose Ramirez’s wRC+ (Weighted Runs Created Plus) for the last 3 seasons, it is a rolling average over the last 36 games which is equivalent to the last 120 plate appearances. wRC+ is a very useful stat for tracking a players offensive output, it works with 100 being league average. So over the ‘16 & ‘17 seasons Jose spent most of his time as an above average player but he had hot and cold streaks which became more pronounced in 2017. In 2018 up until August he had shown some fluctuation like previous years but is was between 210 and 140 and never close to approaching league average. For context in the last 25 years only Barry Bonds has batted at a higher wRC+ than 210 across as a season and 140 is equivalent to Manny Machado this season. So those levels are very good but can clearly see the decline at the end of 2018, the final peak in performance was on 22th August. So what happened to Jose, was he over performing at the start of the year, has he been getting unlucky recently? These are the questions I have looked into, in an attempt to elevate mine, and most Indians fans, fears for next season.
First thing I looked at was his strikeout and walk rates, if his strikeout to walk ratio had increased it would have an impact is offensive output. For the two season previous he had always struck out more than he walked but for almost all of 2018 he has had walked more. This stayed true during the ‘slump’ period. Also there had been multiple times during the season when they were very close. This is responsible for the wRC+ coming down but it isn’t main factor in the offensive decline, if everything else stayed the same it would have brought him back down to the 150 region. So lets look at what happens when he hits the ball.
To do that I have taken the Statcast data which has been made available by MLB via www.baseballsavant.mlb.com, within the dataset that is publicly available you can find the wOBA(weighted On Base Average) and the xwOBA (expected weighted On Base Average) for all pitches. Their xwOBA uses the launch angle and the exit velocity to determine the value of a hit like that based on others hit like that historically. This is the data that is used to give you the hit probability stats, if you have seen any of them during a broadcast or on a highlight reel. With these two stats we can see if a player has been getting lucky by comparing them and if there expected performance is increasing or decreasing.
The graph (36 game rolling again) above is for batted balls only (so excludes walks and strikeouts) and it shows that pretty much consistently over the last 3 season Jose Ramirez has been overachieving on the balls which he has put in play, which you could classify as being lucky. But he has also been steadily improving his xwOBA of the 3 year period as until August this year. This clearly shows that the expected offensive value of the balls that Jose has been putting in play at the end of the season were getting steadily worse and that translated to his actually wOBA and wRC+ significantly declining.
As I stated above Statcast uses the launch angle and exit velocity to determine expected value of the hit, so that means that in the latter part of this season his launch angle and/or velocity had taken a turn for the worse. So, we know that the ball is coming off the bat worse but is it for all types of pitches or is he suddenly struggling on a specific type, i.e fastballs or curveballs? I took the three most common pitch types that he hits and tracked their expected values in the same way as before.
The xwOBA for both 4-Seam and 2-Seam fastballs dropped of significantly over the later part of 2018, with the drop in the 2-Seamer dropping slightly earlier in July before the 4-Seamer sharply declined in August and September. The Changeup and all other types of pitches haven’t dropped anywhere near what the fastballs have, some of them have increased. In September 2018 Ramirez’s average exit velocity was 82.5mph for 4-Seamers and 85.5mph for 2-Seamers, these are 5mph and 3mph slower than the rest of 2018 respectively. These exit velocities are also slower than his averages for 2016 and 2017 as well and have lead to the decline in his offensive output,
So, by the end of the season Jose Ramirez wasn’t getting the same exit velocity on the fastballs he is hitting compared to earlier on, why is that? To try and answer that I looked at what pitches he was receiving to see if there been a change in what people were throwing at him.
As you can see, over the last 3 seasons, as he got better, he has received fewer fastballs and the percent of off-speed pitches he receives has increased. The big change at the end of this season, during the ‘slump’, is the volume of changeups he was receiving. That had hovered between 13-16% for the last 3 season before increasing over to 20% in late August / early September. The timings also do match up to when is xwOBA numbers against changeups started to improve, he improved against them after he starter to see more.
After all this analysis we sadly have to delve into the realms of conjecture. I don’t think Ramirez can read a changeup as well as he can read a fastball, simplistically this makes sense as he has seen far less of them over his MLB career (5,132 fastballs v 1,469 changeups). The evidence to back that theory up is his swing rate on pitches in the zone, that has decreased in the latter part of the season.
But he still isn’t swing on pitches out of the zone significantly, chasing after these changeup, so it seems to me he is expecting more of these changeups to be balls and not strikes. Or he just doesn’t want to hit a changeup and is waiting for the fastball but isn’t getting to it as fast as he was due to him being more weary of the changeup now. I think this is the main reason behind the decline in his season.
If I was the Indians I would give Jose a lot of exposure to changeups and off-speed pitches during the postseason and spring training. He is only 26 so still has time to develop as a player but for the next few seasons I expect his overall level to be more around the 150 wRC+ area than the 200+ area. He has shown a chink in his armour which some of the more analytical teams are definitely going to look to exploit, so this is definitely one to keep a look out for during spring training and the early part of next season.
To deliberate misquote the 2000’s British pop rock band The Hoosiers
‘The truth be told, the truth be told
I’m worried about the future holds, the future holds
I’m starting to worry, Jose’