Wednesday, December 28, 2016

The Vault: Fancy Sports Numbers and Charts


This way back Wednesday is from 7 OCT 2016 via LinkedIn.

Fancy Sports Numbers and Charts: Why It's Turning Heads


Jerry Rice's consecutive reception record. Ted William's .400 lifetime batting average. Wilt Chamberlain's 100-point game. These are basic records that everyone knows at least relative to their favorite sport. But every so often there are new numbers introduced. Homer runs per season since baseball's "dead-ball" era. Total points per team per game since the introduction of the shot clock. The length of a hockey game with removal of the two line pass rule. In addition to charts like these, there are comparison charts to look after. So what's the deal?
Would you be able to examine why in the modern day that sport franchises are using these methods regarding their on-field product? While every team is at least a little guilty, those in baseball are the most notorious. If the introduction of Sabermetrics wasn't enough with adding more atypical player performance statistics, Amazon Web Services launched Statcast for the 2015 Major League Baseball season to provide information regarding the physics of the game as it happens instead of hours or decades later - particularly towards the abilities of the team members. No doubt, that advancement of the former and innovation of the latter was brought on (at least in part) by Oakland Athletics general manager Billy Beane. American gridiron football is the next closest, distant second from baseball but still well ahead of hockey and basketball. That is because of fantasy football, which has since paved way for other fantasy sports as baseball in particular took notice.
We haven't even brought up contracts yet. We pay players based on their past performance (net worth) rather than on their future successes or failures (projected value). This is before signing and performance bonuses. In addition to all of that, there is some form of payroll salary cap in most cases in which you can use to sign your entire roster. Many have argued - baseball especially (again) - that all salaries should be based on fixed value, which is to say having guidelines based on certain requirements or expectations at a player's given position. Then there is non-soccer free agency, which drives up a player's contract price thanks to competitive bidding for the player; whereas in soccer, one goes to another team for a transfer fee paid to a team in addition to paying the player. Thus, if a free agent, the team owes no money to the player's previous team but pays the player their wages. MLB's luxury tax is simply a tax where if a team can afford to pay its entire roster over the allotted league maximum, a percentage of the overflowing balance is then commandeered by the league and dispersed to league programs such as youth development. The percentage increases up to fifty percent beyond four consecutive years of exceeding the maximum.
How is it even remotely possible to pay for player services based on future performance and still meet the necessary payroll requirements? That is hard for the richer clubs, seeing as the New York Yankees have paid it every season since it's introduction (1997-'99, 2003-) that even with the tax are still fairly wealthy. However, it's effectiveness is mixed considering that spending on payroll doesn't always translate to success. For that matter, neither do all of the stats. The more stats, the more the stat categories clash with each other. Stats in Statcast can be used by MLB general managers when scouting perspective upcoming free agents or looking for trades and announcers have this information available during broadcast to inform the interested members of the tuned-in audience. The quarterback rating in football, was only initiated for fantasy football, but both the National Football League and National Collegiate Athletic Association football has adopted their own forms of a similar passer rating. This however is a stat created from the basic pre-existing economics principle of gross domestic product. We still haven't discovered what's the wow factor in all of this.
Following the end of the National Hockey League's Arizona Coyotes 2015-16 regular season, they hired John Chayka just a month before turning twenty-seven to replace their recently fired general manager Don Maloney, making him the youngest GM in the history of the league. Chayka had come over from running his own hockey statistics organization since 2009. Only one of the known other six teams with a stat junkie at the helm has seen success (Washington's Capitals) in recent years.¹ The big number in hockey isn't points (goals or assists), but rather the plus/minus rating. This rating is the player's team's goal difference when said player is on the ice. If his team has scored twice but was scored on once and the final score was to the opposing team three to four, the player's plus minus is plus one (the total of two minus one, regardless of the final score). But unlike that traditional statistic, new ones have been developed over the past decade such as the Fenwick (shooting with misses) and Corsi (Fenwick with blocking) numbers as possession-based stats or PDO (called shooting plus save percentage in the NHL) which estimates luck. The wow factor in hockey is the numbers in these three categories in addition to traditional hockey stats.
In his first big move, Chayka and his staff managed to obtain an additional first round pick which was conditionally owed by the New York Rangers (via Detroit) during the 2016 NHL Entry Draft after the Detroit Red Wings sought to move outgoing club legend Pavel Datsyuk's rights to free up cap space in a failed attempt to lure Tampa Bay's (then free agent) Steven Stamkos. If New York made the 2015-16 playoffs, Arizona would receive said conditional pick in addition to the multiplayer trade the previous off-season. Though Datsyuk's rights now belonged to the Coyotes, it was known long before this trade that he would be playing in the Kontinental Hockey League.
While he isn't Billy Beane in the baseball world, the Chicago Cubs' Theo Epstein (the next, closest person) did manage to take a supposedly cursed Boston Red Sox outfit to become two-time World Series winners over the course of four years prior to joining his current club in 2011. The Cubs now have become serious contenders for the first time in over a decade to win the World Series. With both of these gentlemen, there are stats and then there are sabermetrics. Look at some of the sabermetric stats that have likely been considered in acquiring talent. Similarity scores evaluate comparisons between players how are veterans versus players who have spent less time at a certain level. Adjusted and weighted statistics, such as with (adjusted) earned run average and batting average as well as (weighted) runs created, are revised versions of an original statistic in the category based on the home ballpark and league (American or National) the player typically plays in. There are also stats for value and wins by starting players over potential replacements, how many runs a player creates (an estimated number, not the same as runs batted in), how many runs a defense has saved, the combined ERA of a pitching staff attributed to each catcher, and a game score for individual player just to name a few.
We've established two sports in particular as to why advanced analytics are becoming important within organizations (perceived or actual). In football, the degree of analytics available have improved but are far from both baseball and hockey. Basketball is the worst among the sports that comprise the North American "Big Four" leagues. Most of the additional statistics are slightly more specific averages of pre-existing categories with the exception of similarity scores as in the other sports. The second half of the why is that technology never stops improving. Ten years ago when hockey was first getting into advanced analytics, technology was beginning to speed up as things including dial-up internet was phasing out of the commercial mainstream and broadband connection was replacing it as the new standard. With new capabilities also came new statistical categories.
Granted some statistical measures are overboard. But there still happens to be pre-existing stats that have not dated themselves, and most never will. That's the beauty of sports - we value our history, and we tell that story through the stats. One of the best ever games in NFL regular season history, Buffalo's "K-Gun" offense at San Francisco's (the original) west coast offense in Week 2 of 1992 can be remembered (the "No Punt" game). Or perhaps one of the many moments in baseball, which outnumber those of other sports as baseball is about the moments and not necessarily individual games - think of no-hitters and perfect games when considering entire contests. Or simply 9.69 seconds - Usain Bolt's then-world record in the 100m sprint at the XXIX Summer Olympiad in Beijing in 2008 (which he broke again in 2009). Some are straight to the point, others have a far deeper context. However you look at it, stats are (if not already) the wave of the future.
¹http://www.tsn.ca/statistically-speaking-chayka-makes-history-1.484390

No comments: