It was quite an exciting weekend for Steeler Nation. I suspect some of us haven’t yet recovered from the overwhelming joy of the David DeCastro pick, not to mention Alameda Ta’amu, Mike Adams, and so on. Only time will tell how good a draft this actually was, but at the moment it seems to be one for the record books.
But now we’ve got all these lovely draft picks, as well as a bevy of UDFAs. What are we going to do with them? It’s a good time to explore some of the issues surrounding the coaching of our shiny new proto-Steelers. The Steelers are obviously one of the best teams at identifying talent, historically at least. But picking great players is only half the equation. The other half is how well you develop them.
In general Mike Tomlin and his coaching staff appear to have done quite a good job of player development. But there have been some less-than-ideal outcomes with some fairly high draft picks such as Kraig Urbik and Limas Sweed. And the Steelers have a higher-than-average failure rate in the lower rounds. Therefore I thought it would be interesting to look at the issue of how the coaching staff perceives the abilities of each player.
The reason I began thinking along these lines was a paper I recently encountered, thanks to my annoying elder son, titled "Can People Distinguish Pâté from Dog Food?" Authors John Bohannon, Robin Goldstein and Alexis Herschkowitsch gave the following abstract:
Considering the similarity of its ingredients, canned dog food could be a suitable and inexpensive substitute for pâté or processed blended meat products such as Spam or liverwurst. However, the social stigma associated with the human consumption of pet food makes an unbiased comparison challenging. To prevent bias, Newman's Own dog food was prepared with a food processor to have the texture and appearance of a liver mousse. In a double-blind test, subjects were presented with five unlabeled blended meat products, one of which was the prepared dog food. After ranking the samples on the basis of taste, subjects were challenged to identify which of the five was dog food. Although 72% of subjects ranked the dog food as the worst of the five samples in terms of taste...subjects were not better than random at correctly identifying the dog food.
Here was their methodology:
The dog food tested was Canned Turkey & Chicken Formula for Puppies/Active Dogs (Newman's Own® Organics, Aptos, CA). The four meat products used for comparison were duck liver mousse ("Mousse de Canard," Trois Petits Cochons, New York, NY), pork liver pâté ("Pâté de Campagne," Trois Petits Cochons, New York, NY), supermarket liverwurst (D’Agostino), and Spam (Hormel Foods Corporation, Austin, MN.) Each product was pulsed in a food processor to have the consistency of mousse. Samples were allocated to serving bowls, labeled A - E, garnished with parsley to enhance presentation, and chilled in a refrigerator to 4°C.
The issue is perception. (Note the "garnished with parsley to enhance presentation.") Potential participants were told the purpose of the study, and were also told one of the samples would be dog food. The researchers then asked for volunteers. They were only able to find 18 people willing to participate in the study. I’m a bit surprised they could find any at all. But the interesting thing is this: although almost 75% of the subjects disliked the taste of the dog food more than any of the others, they were unable to tell which was the dog food. I suspect the numbers would be closer to 100 percent of the subjects disliking the dog food, had they known which sample it was. And that might have been the case even if the dog food had tasted better than most of the other samples. Perhaps I’ll run my own experiment with better tasting dog food, or worse pâté.
The factor the researchers were trying to eliminate was bias. Bias often effects how we relate to other people as well, whether we want it to or not. The way you view another person often depends to a great extent on how you initially perceive him or her.
Take this example. Let's say you were introduced to someone with an intriguing foreign accent, and you had been told beforehand you would be meeting the president of Portugal. We'll assume that you, like most U. S. citizens, have no idea who the president of Portugal is or what s/he looks like or if they even have a president. (They do, and he looks a bit like Mitt Romney, actually. I had to look it up.)
You would almost certainly react and relate to your new acquaintance quite differently than if you were told beforehand s/he was an escaped drug lord from a foreign prison. You would possibly even find the person you believed to be the President of Portugal to be better-looking and more well-spoken than you would had you believed you were speaking to a criminal. You would perhaps be equally nervous in either case, but in an entirely different way.
And then there is the Pygmalion Effect. It is a phenomenon in psychology which was given this name after a study done in 1966 purported to demonstrate "one's expectations about a person can eventually lead that person to behave and achieve in ways that confirm those expectations." In the last decade it was tested in a large-scale, five-year study done at Duke University. The study seems to have verified the effect: if you teach ordinary primary school students as if they were gifted, independent testing later identifies a much larger percentage than expected as quantifiably gifted.
What does any of this have to do with coaching football in the NFL? A lot, really. Take late-round draft picks and UDFAs. The success rate for these players, league-wide, is a tiny fraction of the success rate for players taken in the upper rounds, particularly the first round.
The charts below demonstrate how drastic the drop-off is. The numbers are from the calculations for my draft assessment post last week, which compared the drafting success of each team in the NFL. They were derived by adding up the "Career Approximate Values" as per Pro-Football-Reference of all the players each team had drafted in each round between 2001 and 2010. There was a great deal of variation between the best and worst teams in each round. There was also a great deal of variation between rounds for each team.
But whether a team drafted well or poorly there was a dramatic drop in eventual player success between the first round and the second, and a steady if more gradual drop thereafter.
Here are the numbers for the two overall best teams at drafting, the Colts and Patriots, and the two overall worst teams at drafting, Tampa Bay and Detroit. And of course I’ve included Pittsburgh for interest.
As you see, there’s an occasional bump up or deeper-than expected drop, but the numbers follow a classic curve. If you average the five teams it is even more obvious:
On the whole, the best players are taken at the top of the draft, although individual players may drop into a lower round because of non-football issues. Overall the scouts appear to be doing a good job of player evaluation. But is it possible there is another explanation for at least part of the typical lack of success achieved by draft picks in the lower rounds? Could there be an inequity in the opportunities the players receive?
In an Advanced NFL Stats article titled "Career Success by Draft Order," author Brian Burke attempts to determine this very thing after a comment to an earlier article raised the question. Burke said:
Player success has a lot to do with opportunity, and that needs to be factored into the discussion. He [the commentor] suggested that top picks will get the opportunities to start, (ostensibly because teams have the most invested in them). So regardless of differences in ability, top picks would naturally be expected to become Pro Bowlers more often simply due to opportunity.
Burke plots the data and concludes the answer, basically, is no:
The top picks are more likely to become successful even accounting for opportunity (at least in terms of Pro Bowls, an admittedly imperfect measure). In fact, my hunch is that this would over-account for opportunity because Pro Bowl selection and being a starter are both directly proportional to player talent. So we’re really dividing talent by talent + opportunity.
But is inherent superiority the only reason players taken at the top end of the draft are so much more successful than those taken in the later round? It isn’t a 100% correlation, after all. Some of the best players in the league were drafted in the lower rounds or were even undrafted. Perhaps this is where perception and the Pygmalion Effect enters the picture.
The actual reasons any given player did not draft higher than he did are as numerous as the players themselves. A given player’s low draft position or UDFA status may have been completely justified. But once he reports to work for his new team, what does the coaching staff see? Clearly, there were reasons they took a chance on the player in the first place, but there are also reasons they didn’t value him any more highly, or they would have taken him earlier rather than risk losing him. Consciously or unconsciously the staff is not just looking at Player A or Player B, they are looking at Player A, Second Round pick, and Player B, Seventh Round Pick.
Although I have no idea what coaches generally do, I would assume a thoughtful coaching staff like the one headed by Mike Tomlin would be very honest with their new players. I would think they tell each player what they see, why they chose him, and why they, along with the rest of the league, didn’t value him more highly. They help him understand what his deficiencies are, and which ones he can fix. They then help him to find ways to deal with the things he can’t fix, such as being too short for his position or what have you. They then pat him on the back and send him out to work his tail off.
When the players come to camp, in theory it’s a strict meritocracy. Just because you were picked in the first or second round doesn’t guarantee you a place. If you are a UDFA the coaching staff gives you the same chance to compete for a "hat" as anyone else.
My question is, are they able to actually do that? Particularly for the head coach and the coordinators, surely they have a certain investment in their high-round picks. They are going to be judged and second-guessed for their draft choices by the talking heads, the fans, and most critically by their ownership. How well have they judged the talent available? How well they have utilized draft picks, a limited and precious resource? Few people care if a seventh-round pick doesn't pan out, but everybody cares if a first-round pick is a flop.
Because high round picks are more expensive in every sense of the word, and because failure has greater consequences, it is difficult to see how the coaching staff would not provide a greater amount of assistance and opportunities to succeed for their high draft picks. Whether it is conscious or unconscious, this bias surely exists.
In their paper, "Great Expectations? An Investigation of Teacher Expectation Research," co-authors Bruns, McFall, McFall, Persinger, and Vostal said the following:
According to Jussim, Smith, Madon, and Palumbo (1998), "By far, the strongest influences on teaching are usually students' past performance and motivation" (p. 27). The effects of these expectations are cyclical. A student who performs well in the past is expected to perform well in the future, just as a student who performs poorly in the past is expected to perform poorly in the future. If a student who usually performs well happens to perform poorly on a specific assignment, the teacher may conclude that the student is capable of doing the work but did not put enough time and effort into the assignment. Likewise, if a poor performing student performs unusually high, the teacher may conclude that the student had a burst of luck. Despite this new assignment, the instructor will continue to treat both students based on prior performance. The first student will most likely continue to be praised and continually do good work. The second student is likely to be criticized, encouraging a belief that he/she cannot do the work, and causing his/her continued poor perform[ance.]
I have taught at the college level for a number of years, off and on. My first impulse upon reading that statement was "I would never do that." But as I considered it I realized it had a great deal of validity. My assessments of the students in my courses were typically not based upon their performance prior to appearing in my classroom, because generally speaking I did not know the students before they signed up for my class, even by reputation. But I realize I quickly made judgments as to how interested a student was in the subject material, how good a student s/he was, and even how intelligent s/he was. Some of this was almost certainly a result of how well s/he responded to my teaching style, which makes it a sort of double-whammy.
Although I worked very hard to develop an objective system of grading, a music course can't be graded like a math course. There isn’t necessarily one right answer or solution to a question or problem. In fact, there may not necessarily be any "right" answer, just a number of possible ways of approaching a question. This necessarily introduces a certain amount of subjectivity into the grading process. I worked hard to grade fairly and objectively, but I’m quite sure I didn’t entirely manage it. I couldn’t avoid a certain amount of bias. Much of the bias I unavoidably held was based upon how good I had decided a student was. S/he might manage to change my mind later (either for better or worse) but as I reflect on this I realize how remarkably persistant those early impressions were.
I’m sure the coaching staff is affected by these inherent and possibly unnoticed types of bias. After all, the coaches were studying the players and looking at film on them before they drafted them, and are quite familiar with their deficiencies as well as their desirable qualities. They most likely have a certain level of expectation for the player based upon what they have seen before, and possibly a certain unconscious "ceiling" they believe the player is capable of achieving. While they may be persuaded to change their mind if the performance of the player warrants it, I suspect it takes a longer run of sustained high performance from a player they had rated rather low in terms of their expectations. And yet if the Pygmalion Effect is real, they might help a player to achieve more than either the player or the staff actually believe is possible by acting as if a higher ceiling is reasonable.
I can’t help but think of Mike Tomlin’s very public assessment of Mike Wallace as a "One-Trick Pony." Tomlin might well have studied Wallace’s character and psychological makeup sufficiently to realize this was the best way to motivate him to achieve more. But it could possibly have also had the effect on Wallace of causing him to believe he is incapable of excelling in other types of situations besides those requiring a great deal of speed and not much else. And neither he nor Coach Tomlin may be aware of it.
I have no idea whether this is true. I only give it as an example. It could well be the Steelers' coaching staff already knows how to get the very best outcome with each and every player entrusted to them. But improvement may also be possible.
I have an immense amount of admiration for Mike Tomlin, and a great belief in his abilities as a judge of people and a motivator of men. This doesn't mean he can't and won't continue to improve as a coach and a mentor. As he himself has said, "If you’re not getting better, I don’t care what business you’re in, you’re a dead man. I try to look critically at the mistakes that I make and try to learn from them..." Amen to that. I look forward to seeing how well he, and the staff, develop this crop of young men who seem so full of promise. And I also look forward to cheering for the "surprise" player—the one who achieves far more than anyone expected, except, perhaps, for his coaches and himself.