This is a continuation of yesterday's post discussing the baseball portion of the MIT Sloan Sports Analytics Conference. Some more notes on the discussion are below:
Differences Between Teams in Use of Statistical Analysis
The most interesting part of the discussion for me was the revelation that some teams are lagging far behind others in their use of statistics. Dewan, in particular, emphasized the variation between teams in their understanding of sabermetrics. He also claims there is a correlation between the level of understanding of statistics and success on the field. Naturally, the team representatives would not mention any other teams by name but the entire panel grinned when one of the attendees questioned them about the Royals. Someone else asked about the Pirates but Neyer responded that they actually do have a good handle on analytics.
Abbamondi (Cardinals) and Rehman (Diamondbacks) felt that statistical analysis gave them an advantage over other teams. Abbamondi went so far to say that he was concerned that new technology such as field f/x would level the playing field because it would make it easier for less sabermetric teams to use the data. He feels that the fact that fielding data is currently so difficult to process makes it easier for his team stay ahead of others who are not as good at teasing out the data.
The panelists felt that teams, in general, have gotten better at merging scouting with analytics. However, further education is needed. All teams have access to data now but some need to be learn how they use it better. For example, sometimes people new to sabermetrics are too eager to make conclusions on small sample sizes. Rehman suggested that sometimes people will try to slice the data too thin when using pitch f/x in particular. It's very important that statical staffs of teams not only know how to use the data but can present results clearly to those who are uncomfortable with math.
When Neyer asked panelists to give them one unknown that they wanted to know more about both Abbamondi and Rehman mentioned intangibles - work ethic, desire, etc. They admitted that teams are still not very good at determining which players have the best make-up. There is no good way to determine which players will respond to advertising, temptation to take drugs, etc. They both suggested that a lot of scouts mistaken politeness for make-up. Determination of intangibles is a very difficult problem to solve. They did not dismiss the notion that analytics could somehow help them in this area.
They talked about the many advances in the analysis of pitching over the years. Videos have allowed teams to study pitcher delivery and mechanics extensively. Biomechanical analysis has allowed doctors determine with pitchers are bigger injury risks and how to avoid some injuries. The new technology has led to new training and strengthening exercises and better instruction on pitcher deliveries.
Even with all the new advances though, pitchers are still unpredictable and injuries still plentiful. There are lots of different opinions on the best delivery and it's also important to get pitchers to be comfortable with new styles. One of the big problems is that the delivery which is best for a pitcher's health is not necessarily the one which allows him to be the most effective.
Duquette mentioned how teams need to understand differences between pitchers in order to determine the ideal pitch counts for individuals. Statistics and technology and been helpful in this area. Tippett reminded us that expanded playoffs have extended the season for many teams and that has an effect on pitcher management during the regular season. While it seems that workloads have decreased dramatically during the regular season, one can not forget the extra innings that pitchers pitch in post-season - very stressful innings.
Conference in General
My primary purpose for attending the conference was to watch the baseball analytics portion. However, I also attended some other sessions. The "What Geeks Don't Get: The Limits of Moneyball" was not at all smug as the title suggests. It was a very entertaining discussion highlighted by Dallas Mavericks owner Mark Cuban, Indianapolis Colts President Bill Polian and ESPN columnist Bill Simmons.
One of The themes was that analytics don't work as well in football and basketball as they do in baseball becuase those sports are so team oriented. Polian suggested that analytics don't help on the fotball field. Quantifiable elements are secondary to personal and secondary data. However, he said that analytics are important in management decisions. The salary cap in football makes the identification of undervalued assets essential to success.
It was a good day with little to complain about but it was a bit crowded. It was so well attended that it was difficult to find seating. I hope that next year they find a bigger conference room for the baseball discussion.