Jays From the Couch explores the use of WAR to evaluate baseball players and whether it is giving way to something else.
It is unusual to read a column these days by a mainstream baseball journalist without coming across terms like “WAR” or “FIP.” Rather than a nerdy outsider, Bill James has been an executive for the Red Sox for years now, and front offices are dominated by Ivy League educated stat heads. The sabermetric revolution has been televised, but I would argue it is not over. At this stage in the evolution of analytics, I believe teams now hold a massive information advantage over the public, and likely players’ agents.
Terms like WAR have provided a common language for the industry and common folk to use when conversing in “sabermetricese.” From 30ish year old GM’s to legendary writers such as Peter Gammons and Bob Elliot, this new language has enabled fans to gain a better understanding of the game they love, and I think it has been a huge positive. However, I also think we may have reached a tipping point where the public is now using a language which is tantamount to Latin.
Let me first admit that this column is purely speculative, and I will be using anecdotes without much in the way of concrete evidence to support my views. As a result, this entire exercise is vulnerable to confirmation bias etc. With that warning out of the way, I think the publicly used language of WAR and related metrics are relics and will become even more so in the years ahead.
It has been widely known for some time that WAR is a flawed metric. The lack of public disclosure of the “formulas” for the different versions of WAR (Fangraphs’ fWAR and Baseball Reference’s bWAR perhaps being the most prominent) is just one issue. Defensive and base running metrics have been long debated, and how pitchers’ effectiveness is determined relative to the supporting defense has as well. In short, this is quite complicated!
Allow me to use the Blue Jays’ starting rotation as just one example of how assumptions and methodologies can drive significantly different outcomes. Fangraphs’ fWAR for pitchers appears to be very “FIPy” – meaning that it favors pitchers with high strike out rates and low walk rates. This makes intuitive sense, as K’s and BB’s are directly within the realm of a pitcher’s control. However, I would argue this ignores a certain set of skills which some pitchers have, which may enable them to induce weak contact on a regular basis. The Blue Jays just happen to have three such pitchers in Marcus Stroman and Aaron Sanchez (extreme ground ballers), while Marco Estrada‘s combination of his changeup and high-spin rate fastball induces a lot of infield popups and weak fly balls.
Let’s take the 2016 and 2017 vintages of Marcus Stroman and compare his fWAR and bWAR (which is reportedly based more on runs prevention than FIP). If you look at his peripherals, the only big thing I see which stands out between the two season was his Left on Base%. His BABIP, K% and BB% were in the same neighborhood, yet the results were dramatically different in fWAR vs bWAR, as well as in his ERA. His fWAR for the two seasons were close, with 3.6/3.4, while his bWAR was much more variable at 1.5/5.8. Was Stroman’s ability to strand more base runners in 2017 vs 2016 due to luck or skill? Which metric is more “accurate”? I don’t think anyone really knows, and I would argue, it doesn’t really matter.
Ross Atkins made an interesting comment immediately following the Randal Grichuk acquisition when he mentioned that the Blue Jays’ internal defensive metrics were more positive on Grichuk than the “industry.” Do people really think that the armies of young and smart baseball operations departments are just sitting around debating whether to use fWAR or bWAR, or UZR vs DRS?
The Statcast era is now upon us, and there has been an explosion in new data to be analyzed. Teams have been investing in data analysts and data management systems to handle this explosion, which just logically suggests they are doing something with the data! Statcast produces things like exit velocities, launch angles, route efficiencies, running speeds, and the quality of “jumps” outfielders get upon contact. I find it hard to believe that with so much brain power and new data that teams have not been building more advanced and accurate metrics than those the public uses. I may be incorrect, but I think what Mr. Atkins was referencing by “industry” was really metrics like those available at Fangraphs.
The potential for significant disparity in assigned value could be immense, and I also believe it may help explain what has been transpiring in the free agent market this offseason. We know that legacy defensive metrics are flawed and we also know that Statcast has introduced a ton of new information to better assess defensive effectiveness. We know that players peak physically, when it comes to speed and quickness, on average, in their early to mid 20’s. I put forth a theory that one of the reasons older free agents are being valued at lower levels is because teams have better information on defensive value and have figured out that slow/older players hurt a team’s performance more severely than previously understood. Things like exit velocity and “barrels” may also make teams better able to determine which pitchers have a repeatable skill at inducing weak contact.
The Statcast era may end up being to sabermetrics what Alexander Fleming’s discovery in 1928 was to medicine. Analytically speaking, baseball writers and the public are still bleeding patients intentionally using leeches, while MLB front offices may be on their way to bioengineered cures. Statcast was only introduced in 2015 and the dramatic shift in how the free agent market is paying older players began in earnest that following offseason.
When Mr. Atkins said recently at a Pitch Talks event that the industry isn’t going back, it’s likely because the Statcast era has catapulted forward teams’ understanding of the game and assigning value to players. Perhaps Scott Boras needs to recognize that the industry now knows he is a used car salesman and not a Sotheby’s auctioneer selling antique classic cars. Would you pay top dollar for even a Rolls Royce with 150,000 miles on it? Or worse yet, how about an 8 year old Civic with 200,000?
*Featured Image Courtesy Of DaveMe Images. Prints Available For Purchase.
HEAD ON OVER TO THE JAYS FROM THE COUCH VS ALS STORE AND GET SOME GREAT SWAG THAT YOU WILL LOOK GREAT IN AND YOU CAN FEEL GREAT ABOUT.
YOU CAN ALSO HEAD TO OUR JAYS FROM THE COUCH VS ALS FUNDRAISING PAGE TO MAKE A TAX DEDUCTIBLE DONATION DIRECTLY TO ALS CANADA.
THANK YOU FOR VISITING JAYS FROM THE COUCH! CHECK US OUT ON TWITTER @JAYSFROMCOUCH AND INSTAGRAM. LIKE US FACEBOOK. BE SURE TO CATCH THE LATEST FROM JAYS FROM THE COUCH RADIO AND SUBSCRIBE TO OUR YOUTUBE CHANNEL!