Bowling Philosophy-September 2012

Lane of Ignorance
By Thomas Scherrer

I. An Introduction
I have never fully prefaced my reason for writing as a whole.  It is, from where I stand, the most logical form of communication I know of.  Oftentimes, writers write from a form of private trauma or fear and some just have the innate ability to stay in their rooms and write for hours upon end.  For me?  It is a private struggle.  For as I have a great strength to write, and oddly enough about bowling, it compensates for a far greater weakness.  If you know me well enough, you know what my greatest flaw is: I do not speak to people eye-to-eye, nor do I speak to them face-to-face.  I usually look down at the floor or talk while looking around people or sometimes, at an angle.  Psychologically, I cannot truly determine why I came to this personal impasse.  If there are 20 people, it is not a problem.  One person?  I am abjectly terrified.  

But writing?  I fear nothing.  I fear no one.  Some might call that hiding behind a keyboard.  I call it a strange form of courage and when it comes to bowling, it is not being afraid to challenge our collective thinking of how we analyze the game.  Over the last 6 weeks, I have stumbled upon a challenge to once again blow apart the status quo of bowling that will hopefully, over time, solve some questions I have been pondering.

II. The Nature of Bowling
It all started with a simple premise: why is bowling the only sport that cannot truly identify the best performer each week, namely on the PBA Tour?  Pardon me while I burst into a tangent (and not flames)...do we really find out who the best player in each PBA event is?  Are they (and we as fans, to a lesser extent) victims of an antiquated format that no longer gives the advantage to the top seeded player on television?  Consider this a challenge to how we evaluate bowling and perhaps, force some corporate geniuses in Arlington to consider hiring new and younger minds into establishing a search for new bowling information.  Taking our initial premise to an even more extreme extent: if bowling cannot truly identify the best performer each week on the PBA Tour, can we truly identify the greatest performer in PBA history?  You likely scoff at this premise already...we (and when I mean "we" I mean "you") can safely assume it is either Walter Ray Williams Jr. or Earl Anthony because they have the most titles in tour history, with 47 and 43 respectively.  While that may be true, why do we solely define greatness based off wins when a) the player with the highest average after qualifying isn't always the top seed and b) the top seed having to win the tournament twice?  Reflect on the latter for a second.  Yes, after initial qualifying and match play, the top seed after establishing that position throughout the whole week now must bowl one game and win that game in order to be officially declared the champion.  The reason the PBA did/does this?  Television.  TV has become the PBA's greatest marketing weapon while also becoming its most deadly form of manipulating the proceedings.  Bowling a PBA event goes from a game to game marathon into a televised version of bowling's 40 yard dash.  A sprint that might not totally capture what happened all week.  If a player led qualifying by 500 pins, then going into the ABC Saturday telecast shot 180 in the championship match, they were likely taking home second place money.  Forty-two games of consistent, if not dominant work flushed town the toilet.  Instead of championing the best bowler, we are left to criticize said player's lack of clutch bowling in a one game format (cut to Chris Barnes nodding wistfully...).  Here's the funny thing: we do know who the best player was each week, along with the best average.  We just need to do some research into having it all make sense.

III. PBA's Archived Standings
Ah, the beauty of the Archived Standings, where we can go onto pba.com and look at any tournament and find the relevant information into determining the best of the best.

Sorta.

From 1959-1963, the Archives of each season give you only the final placing of each player and does not have an average or their won-loss-tie record in match play.  Yes, I know qualifying formats were different back then, but still.  In fact, the only event they kept some stats on was the PBA National Championship.  It was not until 1964 until they kept track of how many pins a player accumulated during the week plus their match play record...except for the stepladder finalists!  What?!?  Look, I know the PBA was only a few years into its life-its toddler/kindergarten days, if you would but not keeping track of the final qualifying standings of the the final four or five kinda ruins the concept.  In 1968, we had our official moment to peek inside the Archives and see information.  Look at the North Phoenix Open statistics for a second and observe what we see: we know of a 40 game format, with the top 16 bowlers making match play and their W-L-T record.  Along with how many pins every player obtained, plus their match play bonus pins to determine the top bowlers (Note: W-L-T record does become important in this process and I'll explain shortly).  From 1972-2000, the PBA had reached the formats we most closely associate our childhoods with: 42 games for standard tournament, more games in majors, all complete with pinfall and match play records, along with the winner each week getting its due in the Championship Round section.  However, after 1998, they only start posting the Full Standings, meaning where everyone finished after it was all said and done without even pins or a match play record.  Again, I am fully aware the PBA was in financial danger at this time, but it seems as if this is an important bit of information missing into allowing us a look at how deep we can go.  Like coming up 200 feet short of the Titanic-you know it's somewhere in the bottom of the Atlantic, but you still need to see it to research it.  All Jack Dawson jokes aside, we now see that from 1968-1998, we can make a firm evaluation of each tournament and therefore each season, and perhaps, history.  The question: how?

IV. Z-score, Weighted Z-score, and Adjusted Z-score for Match Play
For the statistically inclined, Z-score (or actual score) is basically an application to determine particular score within a given data set or bunch.  It is important because each score can now tell us if it was about the mean, at the mean, or below the mean.  What it also does in bowling's case is that is separates each tournament and allows to evaluate each tournament as equally as possible, regardless of average.  In a sport where technology has allowed averages to skyrocket and, in some cases, dilute our consciousness as to who was better, this formula strips average away and allows to focus on the field average to determine how better or worse a bowler was above the mean for that week.  Noooowwww, we're onto something, right?  If we could find out how great a player was week in and week out, we can ascertain how great certain player were or were perceived to be for a season, or for a career.  With that being said, greatness is consistency to some extent, but winning does matter.  You do need to win.  Just ask Ryan Shafer.  Shafer by all sabermetrical evaluations is a Hall of Fame player.  One of the more exceedingly consistent, and at points dominant bowlers for a 4 or 5 year span, but he won only 4 titles.  That lack of winning does count for something, so how to we account for that.

Weighted Z-score can allow us to weigh each tournament depending on its significance.  For example, the first season I am currently working on (1998), each standard PBA event is 42 games of bowling, then the stepladder finals.  We can weigh the finals by giving the winner a +1 by means of winning the event, but we should not penalize the defeated by giving them a -1.  Take the '98 Peoria Open as an example.  Dave Traber went on to win the event with a Z-score of +2.90 which is a fantastic performance, considering anything that is +3 or -3 for a Z-score is in the 99th percentile of performances.  Traber went on to win the event, adding a point to his Z-score of +3.90.  Sounds outright dominant, right?  Yeah, but too dominant.  As of this time, Traber through 21 events made just 4 match play appearances.  If we were to simply add a point to his win and then add or subtract the Z-score for the other events, it would not compute fairly to those who made far more match play appearances during the season.  So Weighted Z-score will account for adding points per victory after we take the player's Z-score average.  Case in point, Traber's average Z-score for his season was .41, which means in his other 3 match play appearances, Traber was below the mean for it to shrink that much.  So, if we add a point to the back end of it for his win, Traber now has a +1.41 wZ-score.  
What about other events, such as majors, in particular, the ABC/USBC Masters?  I personally hold majors in a higher regard, namely they should be weighed greater.  With that being said, I elected a +2 for the winner of each major.  Even the Masters-again, winning matters and the Masters is a major.  We have to give it some credit, no?  With the Triple Crown majors (US Open, PBA Nat'l, and TOC) and the Touring Player Championship, we have the standard Z-scores to go along with it.  So, winning a major matters a lot more, which it always should be.  Then there is the case of the Showboat.  The Showboat event in Las Vegas had a 56 game qualifying format, which meant more games bowled and a higher prize fund, but not a major.  But it clearly mattered more than a standard PBA event, so how do we weigh it?  How does +1.50 sound?  Now we have given ourselves a weighing system to further separate the men from the boys to determine greatness.  
Then there is the peculiar case of another concept I'm testing known as Adjusted Z-score for Match Play.  What this does it it gives those that made enough match play appearances the opportunity to gain ground despite having lower average Z-scores.  Suppose Steve Hoskins defeated Steve Jaros in match play 220-200.  Since it is only 2 people in a given data set, the Z-score for Hoskins in that particular game was +1, while Jaros was -1.  A simpler way of looking at it would be the player's W-L-T record I alluded to earlier.  Suppose Hoskins went 14-10-0 in a week; we simply add 14 points for wins subtracted by 10 point for losses and we get a +4 for Hoskins.  Suppose Jaros went 11-12-1 in a week: add 11 points for wins subtracted by 12 points for a loss, plus 0 for the tie (Z-score being the actual mean for that game which is zero...and always zero) and Jaros comes in at -1.  Again, we'll add that after averaging Jaros' Z-score throughout the season.  Just like that, we have 3 ways of evaluating a player.  One by consistency in making match play, the other two begin to separate dominance by means of winning tournaments and games in match play. 

V. Endgame
With this information available, we can avoid biases based solely off averages and allow us to get a more accurate reflection of how good a player really was.  Yes, winning does matter, but given the format the PBA has constituted, we have devalued other players where they deserve their proper appreciation.  With that being said, I now have over 3 decades of information to process.  Also, we can now determine who had the most dominant performances in PBA history, including who had the greatest single season in tour history.  Is all of this a hypothesis?  Yes.  Might it fail?  Possibly.  Am I afraid of that?  Never.


Thomas Scherrer is an Associate Member of the Bowling Writers' Association of America.

For those with a love and knowledge for the Sport of Bowling, this IS Bowling Philosophy.  Namaste.