The Recency Error

January 25, 2019

In his Twitter polls, which are fun, Bill is often getting at a general feel for what his followers’ preferences and tastes and opinions are, without a shred of scientific validity for the polls’ findings. His sizable following (45,964 as of this morning) is a blip in terms of reliable sample size, considering how few of them respond to a given poll, and the bigger invalidator is the self-selection of his followers. So the polls are all about the fun.

Bill’s recent polling, on a series of face-offs between pairs of basketball coaches, raised the question of "recency bias."  I argued (on Twitter) that the overwhelming outcome of virtually any pairing will favor the more recent candidate, and suggested that Bill do a dummy poll to test for recency-bias. I suggested Bill do it rather than me doing it, because he has 750 times the followers I have (most of mine are lost souls who wandered into my Twitter feed by mistake) but since he’d need to have 750 times the number of his followers to have any sort of respectable sample size, it doesn’t matter. In fact, results in general don’t matter here. It was just that I was interested in running a dummy poll (in which you’d be asking one question—I suggested "greatest 2B men?"—while actually looking for an answer to an unrelated question, which in this case would be "If I ask about which 2B man you think is greater, will you consistently give me the more recent one as your choice?")  Anyway, Bill shot the idea down, partially on the principle of "I don’t work for you," I suspect (i.e., "if you want to run a poll on recency-bias, Goldleaf, then run your own damn poll"), but also on the problem inherent in my idea that any such poll would give backwards results in the very recent past. Bill explained that younger people would give responses that contradict older people’s answers, so the results would be all over the map in terms of recency-bias. I shot back the suggestion that he could further refine the poll by asking each respondent to identify himself by age, which Bill properly ignored as my assigning him more work to do that he didn’t care a damn about.

I know about as much about basketball coaches, mind you, as I do about best-tasting fatal poisons. I even lack clue #1 about what makes a good basketball coach, especially on the college level, which I consider to be a sport in the same sense that I consider getting drunk or toenail-cutting a sport. I don’t know that I’ve ever watched 30 consecutive seconds of a college basketball game—no, wait, I’ve never watched 30 consecutive seconds of a televised college basketball game, but I’ve actually attended numerous complete college basketball games at the Pace U. gym, where I cheered our team, including some players who had been in my class that morning. It was good fun, they were pleasant students, and I enjoyed myself in the same way that I’d enjoy a student debate or a student talent show: nice kids deserve support, but would I pay a nickel for the spectacle itself? No, of course not: I’ve got a world of professional talent available on Youtube for free, and a world of professional debaters (I can waste hours watching the late Christopher Hitchens demolish some lazy full-of-himself rabbi’s foolish debating points) and I’ve got the NBA, with the best of the best of the best of the best of the best of college athletes, all performing at a level well beyond the hopes and aspirations of most undergrads.

So I’m not sure what a basketball coach actually does, and I’m not sure that the difference between coaching in college and coaching the pros doesn’t actually describe two entirely different activities, like managing hedge funds and weaving potholders. There’s some commonalities: devising plays that will work best against specific opponents, knowing when to make substitutions, judging talent in general, but it seems very different coaching kids who have a lot of competing pressures in their lives (grades, parents, money problems, acne, career goals, roommates etc.) and coaching grown pros, though what do I know about coaching? Maybe college basketball is played on a higher level than I’m imagining—it would almost have to be.

The most exciting night I can remember at a Pace B-Ball game was the night our coach, a young, vigorous, trim, healthy-seeming guy had a heart attack during the first half, was whisked to the ER at the major hospital across the street from the Pace gym, but sadly dropped dead before the second half got underway. He was a nice guy, and a good coach, Darrell Halloran, but this was above-and-beyond in terms of providing excitement. Our basketball program never really recovered, and was moved entirely to our suburban campus a few years later—the night Darrell Halloran died may have been the last college basketball game I ever sat through, now that I think about it.

I could propose Halloran as the greatest basketball coach ever, though. Seriously, if you Google him (Halloran Pace basketball coach  obituary heart attack—any of that will get a few hits, I think) you’ll see how well he was thought of, as a coach and as a man, and he really did turn a so-so B-ball program into a major regional force in Division II hoops. Looking over his obits, I notice that his fatal heart attack, and his surprising coaching success, came at the same age as that of one of my other heroes, Gil Hodges, twenty-five years earlier also at age 47.  So if I could propose someone you never heard of before reading this piece as the greatest B-Ball coach ever, equally deserving as John Wooden (who won Bill’s poll), does this mean I don’t know nothing about Coaching Basketball (which is true as Could Be, or Clair Bee, or Cardi B) or that anyone’s opinion is equally valid when expressed in a poll?  This is a major problem with democracy, that a fine idea—everyone gets one vote—is negated by the fact that not everyone knows as much as his neighbor, and some voters are colossal dunderheads.

Where was I? Oh, yeah, comparing college basketball coaching to coaching in the pros—are they even comparable? More to my point, what’s the point of running Twitter polls on things that may not be quantifiable, or measurable, especially with a self-selected polling group, on questions that are often ill-defined? I’m not busting Bill’s chops here, as much as I’m noticing how these polls seem like fun, and are, but the only real use I can make of them, aside from the fun, is simply finding out in a binary way what I, myself, actually think on a subject. Mostly what I think is "How the hell would I know?"

Bill is currently polling to see who people think the greatest defensive catcher is: his choices are Bench, Pudge Rodriguez, Campy and "other." What I found out is that I think I have no methodology for judging who the greatest defensive catcher is, and would feel like a blowhard checking off any of the boxes, based on my world-class ignorance. So I didn’t. I’m reminded of players who batted against Lefty Grove in 1928 and Bob Feller in 1938 testifying who had the better fastball—obviously their memories are profoundly affected by a ton of factors aside from the actual fastballs: how aware they were of their own deteriorating reflexes over that decade, how intimidated or confident each batter felt in each period, whether they’re more inclined to report nostalgia for the old days that got fossilized in their memory or to exaggerate the more recent experience as vivid in their memory, etc. The one thing you can be absolutely sure of is that they were all incapable of rendering an objective judgment on who was faster, Grove or Feller, just capable of rendering opinions based on their feelings. Maybe some of them felt that Grove was throwing fast in the 1920s and maintained his good fastball until Feller had debuted, so they could compare them more or less simultaneously, while others felt that, no, Grove’s fastball had lost several MPH by the mid-1930s so they would need to compare events from a decade apart.  It’s aw a muddle, as Dickens said, and opinions are like lower colons—everybody’s got one. Respondents to the poll are sometimes going by their own personal version of "dWAR" and other times going by "I saw this guy do such-and-thus, it was in 1957 or 1975 or something like that, and I was nine beers to the wind, but it was the greatest performance behind the dish I ever saw…"

The fun here is just in learning what Bill’s self-selected polling group-- which is scewed wildly from a cross-section of the population, or from baseball fans, or from a wide age-range of baseball fans—believes to be true, based on nothing but their gut feelings. The whole recency thing, though, is what gets me: I’d be interested in learning how recency biases this poll, and most of Bill’s polls, and most of anyone’s polls. Bill argued (I hate paraphrasing his arguments—no one approves of anyone’s paraphrasing, but Bill has never said anything about any of mine other than "Totally Wrong!!!") that younger fans mess up the recency bias-testing by all voting for Jose Altuve and Robinson Cano as the greatest 2B-man of all time while us older ferts will vote for Biggio and Littleio and Mediumio, while close-to-dead ferts will go for Hornsby and Eddie Collins and Nap Lajoie, so how do you measure recency bias when everyone’s Golden Age is different?

So speaking of voting for one’s idols from one’s youth, I must take issue with the whole idea behind the choices Bill’s poll laid out, Bench, Campy, Pudge or other. Of course it’s "other"! And it’s probably someone really, really obscure, someone who barely meets the minimum of "games caught," "years played," or any other standard you want to introduce. For me, my IFOY ("idol from one’s youth"—I disagree with Bill about making up phony abbreviations, too: "IFOY" is funny to me) is Jerry Grote. Not just because I grew up watching him, and loving Jerry’s RA ("Red Ass") qualities, his hustle, his belligerence, his intensity, and the respect that all his pitchers showered all over him. Though I must add here that, for all the love his contemporary Johnny Bench gets in the poll as the GOAT (not gunna explain that one, not gunna do it), no less brainy a pitching authority than Tom Seaver, who pitched to both of them in his prime, argued that Grote was better behind the plate, and Bench himself maintained that if he played on the same team as Grote, he’d (Bench would) be playing third base.

But Grote is a good prototype of my argument, which is: How the Hell is it Likely that a Dude Who Hits Well Also is the GOAT Defensively?

Oooh, oooh, I know, teacher! Pick me! "Not likely at all." It’s mathematically possible, of course, that someone who hits like Johnny Bench or Roy Campanella is also supremely gifted at the entirely-different skills it takes to squat behind the plate, call a brilliant game, throw 120 feet like a laser gun, and frame pitches to convince umps that the plate is actually 10 inches outside, but is it likely? No, the astonishing part of Bench’s or Campy’s mega-talent is that they were able to hit 45 HRs and drive in 16,000 runs while catching better than almost anyone else around. But clearly Grote or Jim Hegan (or somebody) whose offensive game couldn’t get themselves arrested  for marching naked in the Easter Day parade were catching better than ANYone else around, including Bench and Campy. The Mets or the Indians would have been crazy to think that dink hitters like Grote or Hegan were valuable offensive assets—they owe their careers (long and deep careers) purely to their defensive skills. I’m not talking about their individual skills here (though I’ll tell anyone who asks how good Grote looked at the time), I’m talking about all the guys who eke out careers batting in the #8 hole, flailing away hopelessly at decent pitching, solely due to their skills while squatting. It only makes common sense that one of them—maybe Grote, maybe Hegan, maybe Al Lopez or Johnny Bassler—is the GOAT. Bill’s three main options are really for the GOAT Defensive Catcher Capable of Batting 3, 4, or 5. (Just curious—I looked up Grote’s stats: 257 total Plate Appearances batting 3, 4, or 5 in 4844 lifetime PA, which he rewarded with exactly 1 HR.  This includes double-switching appearances into the heart of the order. Bassler literally never started a single game batting 3, 4, or 5.)

Which is fine: Bill can poll anything he chooses, and respondents can give him any misguided thoughts and beliefs and opinions they happen to hold at that moment. It’s all fun, all good. This complaint is more about a different kind of bias than recency bias, but it’s still a kind of bias that few acknowledge even exists: the kind that includes offensive performance in a question that’s ostensibly measuring pure defense.  If discussions of great defensive catchers get skewed in favor of those who are great hitters, as this poll is, that’s kind of nutty, isn’t it? I mean, you’d rightly go nuts if I ran a poll about great hitting catchers, and listed as your choices players like Yadier Molina or Duffy Dyer, right? (Oddly enough, by certain measures, BTW, Dyer shows up as Grote’s defensive superior, though neither could hit his way out of a barf bag.) That’s just crazy talk, but the other way around, it’s seen as reasonable.

Think about it, when the Benches, the Mauers, the Munsons get old, do they retire? Of course not—they play 1B or the outfield. But when the light-hitting catchers can no longer catch, we stick a BBQ fork in them. Ain’t nobody putting Johnny Edwards in the lineup at first base—the only reason these guys are in MLB at all is their defensive prowess, and as soon as that reverts to mediocrity, See Ya! As groups, I think the all-time greats, Berra, Bench, Campy, Cochrane, Hartnett have to be ranked below Hegan, Grote, Molina, Johnny Edwards etc. defensively just because the only thing that keeps the latter catchers in the major leagues is their ability with the glove. I don’t even see where it’s arguable.

The only valid-seeming argument you could make, I think, is the one that goes "Bench et al. are great hitters because they’re great athletes—the hand/eye coordination that makes them into world-class hitters bleeds over into every other part of their game, so of course they’re also great defensive wizards as well," which is clearly nonsense because they’re obviously inferior at certain other parts of the game that don’t include holding a bat in their hands. Baserunning, for example.  Most of these guys are clods on the bases, and no one makes the case that Campy was a fabulous baserunner because he was such a great batter. One has NUH-thing to do with the other, and no one thinks it should. Baserunning and hitting are two discrete skills, as are catching and hitting.

Probably there has been a thoughtful study of how to poll people while testing for biases of various kinds—recency, decency, Liam-Neeson-see—that I could look at to see how smart people have accounted for bias in polling, and how successful they have been. But something occurred to me as I was finding fault with Bill’s polls giving skewed results (or meaningless ones) because his respondents are alive today (most of them, anyway) and so will tell you that Joe Mauer is a greater hitter than Mickey Cochrane because—wait, who’s Mickey Cochrane again? Is he the one that invented the face mask? Or the one who caught a grapefruit thrown from an airplane? What occurred to me as I began preparing my mocking remarks about these dopey kids today is that I agree with them. I’ve been arguing here for months now that the Joe Mauers of this worlds are much better hitters than the Mickey Cochranes who are in turn greater hitters than the Buck Ewings. I’ve been making the case that baseball players of all stripes-- pin-, Zebra-, and prison-,-- have been improving so steadily that it’s entirely possible that every full generation is markedly superior to its predecessor, and that 19th century stars couldn’t, therefore, start on a ladies’ weekend softball team today.

In other words, as skewed-up and screwed up as recency bias is, it just might be perfectly correct. It might (and probably does) reflect reality, for all the wrong reasons.  The greatest players, overall, are the ones playing right now, and the second greatest ones are the ones who’ve just retired, logically speaking. This principle cannot apply across the board, at every position, for every skill, in every statistical category, but as a rule of thumb, it’s not a bad one. So all these smug little pishers proclaiming Mike Trout as the GOAT just because they’ve never seen Mantle jack one out, or Willie run first-to-third on a simple 6-3, or Dimag glide elegantly to nab a ball in Death Valley—they might just be right. They probably are right.

COMMENTS (9 Comments, most recent shown first)

Steven Goldleaf
Well, I wrote a whole column (I think two whole columns) on the subject of what factors go into players' improvements, and I don't think anyone wants me to repeat myself more than I already do, you least of all, Frank, but I'm not sure what the individual bases of improvement have to do with this discussion. Either way, improvements support the validity of recency bias, right?
3:18 PM Jan 29th
Steve, not wanting to rehash old arguments lets just assume that you are correct and that each generation of players is better than the previous one. How much of this improvement is due to increased player ability (i.e., larger people, faster runners, stronger, more knowledge about game including analytics, detailed film analysis, etc.) and how much is due to technology (i.e., bigger gloves, smoother fields, better medical knowledge, drugs, PEDS, etc.)???

Here's an interesting discussion:

on technology driving much of measured athletic improvements. Is there anyway to separate the effect of these disparate factors for baseball? Are the generational improvements 60% better player ability and 40% technology? 90/10? I don't know.

As an aside, IMO, a 1905 Einstein if brought in Mr Peabody's Way Back machine to today would quickly become one of the if not the top physicist of today if he wanted to do the work. For Newton it might be too far of a leap ......
9:21 PM Jan 28th
I've commented on this site before about watching telecasts of games from the 50's on YouTube, and (especially when a centerfield camera view is used) seeing that the pitchers didn't throw anywhere near as hard as the pitchers of today, nor were their breaking pitches as good. I hesitate to draw conclusions from this because A) I could be wrong in my judgment of the efficacy of the pitches and B) that doesn't mean the hitting stars from that era couldn't hit today's pitchers either.

Honus Wagner was known for working out and staying in shape in the off-season; would he be a great player today? We'll never know the answer to this and other such questions, but it's fun to think about.​
9:58 AM Jan 28th
I shall not attempt to define what a great defensive catcher is, and perhaps I could never succeed in intelligibly doing so. But I know it when I see it.
4:57 PM Jan 26th
Steven Goldleaf
There's a crucial difference, Frank, between having the ability to think in totally outside-the-box ways that innovate (invent?) an entire field, and having the ability to master that field many generations later. Do you doubt that today's top physicists, having learned everything Newton and Einstein figured out, including the things they figured out wrong, are better at modern physics than Ike and Al would be if they were plunked down (in violation of the laws of physics) in 2019? Maybe A. and I. would be able to catch up after a few seminars (and maybe not, depending on how stubborn they are) but they would not be clearly better physicists, IMO. In any event, I'm not going to argue here the thesis that I've gone on about endlessly in previous articles, but simply wanted to assert that I'm using that thesis to validate the bias in recency that I'm decrying here. Trout is standing on the shoulders of Giants like Willie Mays, but that doesn't mean his head isn't five feet higher than Mays' head is.
8:27 AM Jan 26th
FrankD, I don't disagree, but there is more to it: today's players are better because they are in better physical condition, which is a result of their making enough money to spend off-seasons working out rather than working for a living.
5:17 AM Jan 26th
well... a long winded piece which I sampled at points .... lets go another way in long-windedness about recency bias. Are the top physicists today better than say Einstein or Newton. Now all the top physicists today learned when young and understand Einstein and Newton whereas Newton never knew/grasped Einstein. And of course Einstein spent more time failing to come up with a "theory-of-everything" then he spent on either the Theories of Special or General Relativity. Einstein spent a lot of time attacking Quantum Physics, which he never accepted ("God does not throw dice"). So are today's physicists better than Newton or Einstein? I think a poll of modern physicists would be a resounding "no". I'll ask Steven to do this poll, I could be wrong. Getting back to baseball: are today's MLB players better because they've mastered a game that has evolved from unkempt grounds, spitters/emery/darkened pitches, no off-season training, no videos, ad infintum ... to what there is today? If the players today look greater its because, as paraphrasing Newton: Todays players look great because they stand on the shoulders of Giants ............
9:50 PM Jan 25th
Steven Goldleaf
No, no, no, Elliot. I have no criteria for Greatest Defensive Catcher. Absence of offensive ability does not = presence of defensive ability, not at all. I'm sure it's possible to suck at both catching and hitting, me being the most convincing example (to me, as I'm sure you are to yourself). But I'm clueless how you'd find a method for choosing the best defensive catcher.
4:23 PM Jan 25th
If I correctly understand your criteria for Greatest Defensive Catcher, I would have to select Bill Bergen.​
3:15 PM Jan 25th
©2019 Be Jolly, Inc. All Rights Reserved.|Web site design and development by|Terms & Conditions|Privacy Policy