Sloan Sports Analytics Chat (#SSAChat) Follow-Up
This afternoon from noon to 1 p.m. EST, the Sloan Sports Analytics Conference (SSAC) hosted a Twitter chat featuring several speakers from the upcoming event. The conference, now in its 6th year, takes place from March 2-3 at the Hynes Convention Center in Boston (tickets for the event are still available but going fast!). Questions for the chat could be submitted using the #SSAChat hashtag and directed to:
@dmorey: Houston Rockets General Manager and SSAC co-chair Daryl Morey.
@ESPNStatsInfo: The ESPN Stats & Information Twitter feed, with Director of Production Analytics Dean Oliver fielding the questions.
@jluhnow: New Houston Astros General Manager Jeff Luhnow, who was formerly VP of Amateur Scouting and Player Development for the St. Louis Cardinals.
@RicBucher: Senior ESPN NBA reporter Ric Bucher.
The panelists all did a great job of fielding the many questions that were thrown at them. They even answered a few of mine:
Me: @ESPNStatsInfo Do you plan to create a Total QBR metric for FBS QBs? Possible to rate top 5 (Luck, RGIII, etc.) pre-April’s draft? #SSAChat
@ESPNStatsInfo: @whimsicalcookie College QBR will be done. Rating QBs for draft is similar but not the same problem. Harder problem.
Commentary: For those not familiar with Total QBR, it’s a comprehensive metric for rating quarterback play that ESPN developed this summer and introduced to fans via an hour-long TV special. Unlike passer rating, it accounts for all of a quarterback’s plays, including runs, sacks taken, and fumbles (regardless of which team recovers). The proprietary metric is derived from film study and attempts to assign credit and blame for good and bad plays more accurately than traditional stats. For instance, a 50-yard touchdown pass that is caught in the end zone would add to a quarterback’s QBR much more than a 50-yard touchdown on a screen pass, where most of the yards come after the catch. Finally, QBR is weighted based on the situation: down, distance, score, and time of the game. So if a QB throws a TD pass on 4th-and-goal from the 15, putting his team ahead as time expires, that would be given much more weight than the “same” 15-yard TD pass on 1st-and-10, early in the game, with his team already comfortably ahead. For more on Total QBR, see this Q&A with its main developer, Dean Oliver.
ESPN has calculated QBR for all NFL quarterbacks going back to 2008, and during the 2011 season they rated each QB’s play within about 12 hours of their game’s end. They have not yet tackled QBR for college quarterbacks, and with 120 teams in the FBS it is admittedly a much bigger undertaking. While Total QBR has been slow to take off, I think it’s an extremely useful stat, and I’m even more interested to see what it has to say about college quarterbacks. With a greater diversity of offenses in the college game, including many option attacks built around mobile quarterbacks, it would be great to see how Total QBR rates dissimilar QBs. Since Total QBR is a rate stat reflecting quality of play, and not a volume measure, it would also be a useful tool for comparing the performance of QBs in balanced, pro-style NCAA offenses, like Stanford’s Andrew Luck (404 pass attempts in 2011), with those from run-and-gun attacks, like Houston’s Case Keenum (who threw it 603 times).
Me: @jluhnow are #Astros using systems like FieldF/X and HitF/X? Did #Cardinals?
@jluhnow: Ever hear of Mike Fast? Great work, will help us.
Me: @jluhnow sure, @fastballs, I follow him!
Commentary: This was pretty neat — while I’m sure that some of Mike Fast’s friends and colleagues at Baseball Prospectus had already heard the big news, this tweet from Luhnow served as the official announcement that the well-regarded B.P. analyst had been hired by the Astros. If you’re not familiar with Fast’s work, such as his ground-breaking study on catcher framing, do yourself a favor and read some of it here on B.P. or here in the forums of The Book. You can also follow him on Twitter: @fastballs. With Luhnow at the helm and Fast on board, the Astros are quickly developing one of the more analytically-savvy front offices in baseball.
@RicBucher: Team exec said Dal had best FG% w/6 secs or less on shot clock. By far…
Me: @RicBucher do you attribute this to Roland Beech’s work with the team? Any insider info on players’ reaction to having a stats coach?
@RicBucher: @whimsicalcookie Don’t know who to credit, re: Beech. Head coach is the decider on how stats impact game plan. Players have less sway.
Commentary: For those who don’t know him, Roland Beech is a pioneer in NBA analytics. He founded the website 82games.com and was hired by Mark Cuban prior to the 2010-2011 season as the Dallas Mavericks Director of Basketball Analytics. Beech serves as one of the Mavericks’ assistant coaches, and is to my knowledge the only “coach” in the four major sports whose expertise is in analytics but who is a full-fledged member of the coaching staff, traveling with the team and sitting on the bench during games. Beech spoke on the NBA analytics panel at this fall’s New England Symposium on Statistics in Sports at Harvard, along with Sam Hinkie (from Daryl Morey’s Houston front office), and New York Knicks Assistant Coach Kenny Atkinson. For more on Beech and his contributions to the statistically-enlightened approach that helped the Mavs bring a championship to Dallas, see this column from ESPN.com’s John Hollinger.
If you liked this blog post, go to Twitter and read through the discussion on the #SSAChat hashtag for more. Also be sure to follow the Sloan Sports Analytics Conference Twitter feed at: @SloanSportsConf. Hope to see many of you in Boston in March!
Mike Cook © 2012