Friday, February 8, 2013

Sam Donnellon Is Not A Fan Of Using Stats In Sports

There are some things in life that never change.  One of these things is the old, white newspaper columnist taking joy in his own ignorance proudly writing a column so full of misinformation and mischaracterizations that the fact it got published in the Philadelphia Daily News and not some random former WIP Caller's new Blogspot is in itself mind-blowing.

Meet Sam Donnellon.  He has been writing about sports since 1981.  Most of the time his stuff is easy to ignore as it is normally buried in the Philadelphia Daily News and none of what he says is worth discussion.  Out of sight, out of mind, and all that jazz.  But occasionally, he comes up with something so ridiculous, so absurd, and so out of touch with reality that ignorance is an impossibility.

Meet Sam Donnellon's latest column, "Better to judge baseball talent with the naked eye."  It's worse than it sounds.  Let's not waste any more time and delve right into, FJM style.  Donnellon's words are in italics.

ONCE, WHEN the Internet was not as friendly and we sought to observe athletic greatness rather than quantify it, sports was a simpler world. You watched under well-lit skies and inside of electrified arenas and amid pivotal moments and you did not need to look anything up afterward to understand what you saw.

A long, long time ago, people watched sporting games and no one bothered to research what they saw.  Bill James did not start in 1977, and the internet is the first time anyone ever sought to research deeper sporting events.  We hold these facts to be self-evident in the eyes of Sam Donnellon.  Never mind that in the days before Bill James, ERA was used to as a measure to evaluate pitchers, and people would look ERA up in newspapers to help "understand what they saw."  Sure it is a crude and imprecise measurement by today's standards, but when advanced stats do not exist, you do the best you can.

There were numbers guys back then of course, keeping their own stats in notebooks, reciting them at the lunch table or later, at the end of the bar. They were interesting and, as long as they didn't go on too long or get too wrapped up in minutia, welcome too.

"As long as they didn't......get too wrapped up in minutia."  What. The. Fuck.  So stats in notebook were welcome, but the second they sought to dive deeper and provide further analysis, that's a big no no?  Then what the fuck is the point of stats?  If you hate stats, go out and say "I hate stats!"  That's bad, but saying you did not mind stats as long as they did not get too deep into it is worse.  If you don't want further analysis, then what is the point of even wanting to listen to stats in the first place?

And someone tell Sam Donnellon that these people he cites would not recite them at the lunch table or the end of the bar, but rather in their mother's fallout shelter.

Those guys are our high priests now, no doubt about it. Whether they are exuberantly noting a baseball player's WAR or WHIP, a hockey guy's plus-minus, a quarterback's "rating," they seek to quantify exactly what you saw.

What.  What?  WHAT?!?!  Okay, let's take this one at a time.

Donnellon seems to indicate that the people of the old days who would watch stats like W-L record and ERA are also the same people who have moved onto advanced stats.  No, Sam, that's not true, those are the same people who have moved onto sexually molesting children.  Bill James did not start out his career praising the virtues of pitcher W-L record and ERA.  And old newspaper columnists and writers who started out championing ERA have remained champions of ERA.

But worse yet, take a look at the stats Donnellon cites.

Wins Above Replacement (WAR):  Okay.  WAR is a stat worth knowing and taking a look at, but with all stats, it needs context.  There are some valid criticisms of WAR, and some not-so-valid criticisms of WAR that normally go along the lines of "war whats it good for nothing lol follow me on the twitter dot com".

Walks and Hits per Innings Pitched (WHIP):  Does anyone look at WHIP?  Anyone?  Anyone?  Advanced stats people?  Old fogey old stats people?  Proudly ignorant young old stats people? The "use your eyes crowd?"  The next person I find that will religiously take peaks at WHIP will be the first.

Hockey plus-minus:  Show me a savvy statistical person who uses +/-, and I will show you a statistical person who is not very savvy.  +/- is awful and tells you nothing.  In short, it rewards everyone the scoring team equally and punishes everyone on the conceding team equally, even though blame usually is not that proportionate in reality. To better understand why plus-minus is useless, check out this post from The Shutdown Line for a breakdown of the overall uselessness of +/- and this post from Broad Street Hockey which does a video review of one game when Eric Gustafsson was a +6.

QB rating:  There may be some people that swear by QB rating, I honestly stay away from it just because I have no idea what it is telling me.  Instead, I look for data that goes through play-by-play analysis like that from Football Outsiders.  They have their own innovative stats that paints a much clearer picture of what is happening than QB rating ever will.

And if it's not exactly what you saw?

They tell you that you are mistaken.

They turn a blind eye to your naked one.


Let's just assume that we are talking about practical stats here like OBP, wOBA, and xFIP in baseball, Fenwick and Corsi and zone entries in hockey, and DVOA and DYAR in football, and useless stats like WHIP, +/-, and QB rating.  On one hand, you have a set of objective data that has no preconceived notions or biases to it.  Stats are stats.  Stats represent what happened exactly as it happened.  On the other hand, the human brain can play tricks on you and you may be fooled by your memory.  If you believe Matt Carle sucks, then you are going to see things that Matt Carle sucks (turnovers leading to goals) and ignore everything good he does (playing good D against great players) while conveniently ignoring context (great d men will have a high number of turnovers because they handle the puck more and they handle the puck more because they are great at defense; in other words turnovers will correlate more with number of times handling the puck than it will Corsi or other metrics of skill).

What I am getting at here is on memory, the naked eye is not reliable and should not be the sole judge of what happened in a sporting event.

Thing is, we believe them now when they recreate reality. The Internet has allowed them to multiply unchecked, like Canada Geese. The door into mainstream has been kicked open for every cellar dweller with a laptop.

Thanks to the internet, people, myself included, have been able to expand their horizons and knowledge of sports through reading online publications not available at a newstore.  It's either that, or THE ZOMBIE APPOCALYPSE IS NIGH!

You know these guys. They get all frothy-mouthed each time a new acronym is invented to try and quantify value, yet show little or no reaction when a difficult doubleplay is executed, or a well-thrown pitch is fought off by a batter down in the count. They feed off each other, creating a culture more hell-bent in arguing a debatable point than reflecting what you actually saw.

Did you know writing this has given me a boner?  But seriously, what is wrong with an intelligent exchange of knowledge of ideas in an attempt to greater understand things?  Some people may enjoy sports for imagery, but others may seek to know what is actually happening and why it happens, and it forever boggles my mind that there are close-minded simpletons like Sam Donnellon who continually reject this simply on the grounds of the fact they have no desire to understand a new concept.  In spite of what people may believe, not all stats people are assholes, and there are some very well-written tutorials designed to teach stats to those that may have not heard about them but are interested in learning.

I was watching a smart debate recently on the Major League Baseball Network about the steroid era and the Hall of Fame when Al Leiter brought Jack Morris' name up. Morris, said Leiter, taught him the value of "pitching to the scoreboard" and not the stat sheet. Leiter said it was the best advice he ever got.

Maybe to win big World Series games and championships, but not necessarily the approval of the stat mavens that will rank your status all time. I brought Morris' name up in a dinnertime press room Hall of Fame discussion the other day and a numbers guy immediately reeled his lifetime ERA of 3.91 off the top of his head, spitting it out almost in disgust to discredit Morris' candidacy.

Push aside for a moment that his ERA ballooned 20 points over his last two seasons as he tried to milk one last big paycheck out of his abilities. Thing is, I saw Jack Morris pitch, and he was great. Great enough for the Hall of Fame? All I know is there are less great players in there.

Kind of ironic how Donnellon criticizes advanced stats people for getting frothy-mouthed over new stats, but he froths at the mouth here over Jack Morris and one thing Morris may have taught Al Leiter.

Now I am not going to go into great detail about Jack Morris's (bad) Hall-of-Fame case as others have done it better and more thorough than I ever could.  And the concept of "pitching to the score" is bullshit.  There is no evidence of it, there never has been any evidence of it, and there likely won't be any evidence of it.  Remember this for later.

It's one reason I don't use my Hall of Fame ballot. Statistics, even the more contrived ones, have value. Greatness, though, is a naked-eye assessment. If they're going to argue that Morris or even Curt Schilling are less significant than guys already in there, guys like Bert Blyleven, then they ought to call the place the Baseball Bureau of Statistics. Because to the naked eye, it's absurd.

Greatness can also be evaluated through statistics and naked eye assessment.  Take for example Roy Halladay in the 2010 and 2011 seasons.  You were told going in he was one of baseball's best pitchers, if not the best, and your eyes were treated to two great seasons, one of which featured baseball's 20th ever perfect game and baseball's 2nd career post-season no-hitter.  Most people with the naked eye were able to believe Roy Halladay was a great pitcher in those seasons.  But what if they wanted to find out just how great he was?  Check his Fangraphs page and Halladay's numbers in 2010 and 2011 for yourself.  Boom. Greatness.

But there are times when the eye and the stats will disagree, and during those times, I will trust the objective data over my subjective memory every day of the week.

Here's another thing that bothers me: The valuation of regular-season statistics over postseason ones. Some of the game's more selfish players have recorded some gaudy regular-season statistics. Others have built their impressive résumé playing for poor teams in pressureless environments. Statistics built in the AL Central over the last two decades are not equal to statistics built in the AL East.

The problem with valuing regular season stats and postseason stats equally is the unfair nature of the comparison.  Especially in baseball, there are significantly more regular season games than postseason games, and the sample size for postseason games is a small one that does not tell you enough information.  There is too much room for variance in the small-sample size postseason to make accurate judgments about a player's ability.  It's why "clutch" arguments are nonsensical bullshit.  Delmon Young is a shitty baseball player.  Delmon Young was named 2012 ALCS MVP.  Does that mean everything we know about Delmon Young is wrong because he got on a hot streak at the exact right time?  Of course not.  Delmon Young is a bad baseball player who is unlikely to improve.

Over the course of a career, quality of competition a player faces will generally normalize and is not a big deal as Donnellon is making it out to be.

That's a naked-eye assessment. I'd probably put Morris into the Hall too, probably for the same reason stat mavens would throw him out. He won more games than anyone in the 1980s, but many, including our own David Murphy, have compellingly argued that a pitcher's won-lost record is among baseball's greatest irrelevancies.

Murphy has mentioned Cliff Lee's 2012 season as recent evidence of this. There is no doubt that Lee deserved better. But the naked eye, the one that watched the season in its entirety, recalls at least a handful of times when he received substantial leads and could not hold them. Morris would say, I suppose, that in those cases, he failed to pitch to the scoreboard.


One could do the same thing with just about every pitcher in baseball.  Cliff Lee only sticks out more because his W-L record was talked about significantly and he appeared to be living down to expectations (even though his Fangraphs page may suggest otherwise).  He got written about, therefore, we recall more times when he gave up a lead last season, even if other Phillies pitchers had similar outings.  And again, "pitching to the score" is nonsense explained in a link above.

[Edit:  I just now discovered this post on Hardball Talk which talked about this part of Donnellon's column.  Cliff Lee blew a lead of 3-runs or more only once during the 2012 season, and Cliff Lee left that game with the score tied. But why trust what objective stats say when Donnellon can tell you he saw with his own two naked eyes Cliff Lee blew many more substantial leads.]

Clearly, statistics are not irrelevant. But they should be used to support the naked eye, not create an alternate reality. Discussing Schilling's Hall of Fame candidacy, a stat disciple mentioned that his win total averaged out to 12 games a season.

Without knowing it, Donnellon has stumbled upon a reasonable point in his first sentence.  Stats are not irrelevant, and the naked eye is not irrelevant.  Context is needed when studying tape, and film breakdowns can help provide you that context.  If you completely ignore context, you get posts like this.  But if you use your eye to judge everything and dismiss statistical analyses, you get ignorant columns like this one from Donnellon, where he cannot even correctly identify advanced stats people use.

I'm with Murphy on this one. It might be one of the most irrelevant statistics one can offer about the former Phillies ace.

Ok then!  Pitcher wins is one of the most irrelevant stats one can offer!

Schilling, by the way, was one of the first athletes I ever saw use a laptop inside of a clubhouse. He kept tabs on every umpire's strike zone, what pitches got batters out, what sequences he had used the last time he faced that night's team.

It's one of the tools that made him great. But if he ever does get a plaque in Cooperstown, I doubt it will mention that.

"Curt Schilling tracked umpire's strike zones and studied his opponents tendency and that is just one thing that made him great and Cooperstown won't mention it but hey did you know Jack Morris pitched to the score and was totally awesome for it and should be in the Hall of  Fame and Cliff Lee should learn from him?"

In closing, Sam Donnellon is a complete ass who knows nothing of what he is talking about, despite his hilariously desperate attempts to convince you otherwise.  And really, despite the fact that I went through this whole sad and sordid affair, Donnellon lost any credibility he might of otherwise had the second he claimed PLUS-MINUS WAS AN ADVANCED STAT IN HOCKEY! PLUS-MINUS!!!!!!!!  AN ADVANCED STAT!!!!!!! THIS HAPPENED!!!!!!

Perhaps if Sam Donnellon used his naked eyes to do 5-seconds worth of research he would have learned otherwise.