Al Davis, the late owner and general manager of the Oakland Raiders, had a famous motto for his team, “Just win, baby.” Nothing else mattered, or so he claimed. The Raiders did that on the opening week of the season Sunday, beating the New Orleans Saints, 35-34, when Derek Carr hit Michael Crabtree on a fade route for a two-point conversion in the last minute of the game.1
“We came here to win,” Raiders coach Jack Del Rio insisted. “I thought, ‘Let’s win it right now. Every part of our strategy was focused on getting the win.”
It was surely a gutsy call2 by Raiders coach Jack Del Rio and one that seemed to buck conventional football wisdom, which is generally highly conservative. The NFL introduced the two-point conversion – long a staple of college football – in 1994, but it has been attempted only relatively rarely. Before Sunday, a team trailing by seven in the final two minutes that scores a touchdown went for two just 11 times in total and not once since 2013. Only three of those attempts were successful.
Ron Rivera of the Carolina Panthers has undergone an amazing transformation from a typically risk-averse NFL head coach into “Riverboat Ron,” a coach willing to “go for it” and, not coincidentally, focusing on doing so when the odds are in his favor. Also, and also not coincidentally (I suspect), when he morphed into Riverboat Ron he started to win a lot more. But he’s a focus because he is so rare. Most NFL types, despite being desperate to win, haven’t paid a lot of attention to this research or to analytics generally. They continue to wallow in their highly conservative conventional wisdom. Baltimore Ravens lineman and mathematician John Urschel explains why.
“Just because the expected points of one endeavor is greater than the other, doesn’t mean it is what coaches are going to do.
“Why? Because, as you might have surmised at some point, NFL coaches are risk averse. Coaches like low variation, and a difference of .03 expected points per extra point is not nearly enough to deter them from the safer choice of going with a slightly longer kick (which has variance of .07) as opposed to the much riskier two-point conversion (which has variance .25).”
Now compare it to the much more typical and much less analytical approach of former Steelers coach Bill Cowher. “If we all listened to the professor, we may be all looking for professor jobs,” said Cowher. But that rejection of analytics is changing at least somewhat, led (not surprisingly) by Bill Belichick (despite his careful denials).
But in this particular instance, Del Rio might have been bucking more than just conventional wisdom. His decision also seemed to buck the math, even though football analytics generally suggests that teams should go for two much more often than they do, per noted University of California Berkeley economist David Romer (full paper here), especially now that traditional PATs are kicked from a line of scrimmage at the 15-yard line rather than the two.
This occasion may have been different.
As the tweet notes (with fuller explanation here), based on ESPN’s win probability model, going for two dropped the Raiders’ win probability from 51 percent to 44 percent prior to the conversion. But it’s hardly a no-brainer decision, as the full ESPN piece also makes clear.
“That assumes that teams make extra points 95 percent of the time and convert the two points at a rate of 48 percent, which is based on NFL averages over recent years. The Raiders would have had to convert their conversions at a rate of 56 percent for it be a break-even proposition.”
The ESPN model may simply be constructed differently from Professor Romer’s model (and it is). But there is more involved here as well. What the ESPN model can’t account for is the particular situation to which the general data would be applied – in other words, the specific human factors applied to that single game. Thus the question, from a probability standpoint and as specified by the math, even assuming the quality of the ESPN model, was whether Del Rio reasonably thought his team could convert at that particular time at a 56 percent rate.
Let’s take a look at the evidence.
For starters, consider that the Saints offense is great and was playing great. Drew Brees threw for 423 yards and four touchdowns against the Raiders while mostly playing from ahead, which should have depressed his numbers. Assuming the Raiders could keep the Saints from scoring in the final 47 seconds (which they barely did but which they had to do whether they kicked the PAT or went for two), did it make sense for Del Rio to avoid giving Brees the ball in overtime? There’s a very good argument that it did.
Also consider that Crabtree’s catch percentage in tandem with Derek Carr, while an only okay by league standards 63 percent in 2014 and 58 percent in 2015, was still well above the 56 percent threshold. Additionally, it was an exceedingly high 77.8 percent in that game, so far above the threshold (even given the small sample size) as to suggest a very positive match-up advantage. Moreover, on the fateful two-point conversion attempt, Crabtree was lined up against undrafted free agent rookie cornerback Ken Crawley (who was in the game only because Delvin Breaux had been sidelined with an ankle injury earlier) in single coverage, all against the backdrop of the Saints only having had the 29th-ranked defense against number one receivers last season. Del Rio may even have been particularly confident about his two-point conversion schemes – the Raiders had already converted one earlier in the fourth quarter, giving them as many successes on Sunday as they had all of last season.
There are also ancillary benefits to Del Rio’s decision that he may well have considered. Being aggressive would likely instill confidence and aid team chemistry even if the move failed, especially in that it was the season’s opening game. “It is just that kind of belief and that kind of trust for a coach to believe in us like that,” said Raiders quarterback Derek Carr. Belief by and in teams is often crucial to success. “It just gives us so much confidence to have him do that for us.”
There is no definitive slam dunk conclusion here. The move was surely gutsy, since few would have criticized Del Rio had he kicked the extra point to tie that game while many would have criticized him had he failed. Although I run the risk of outcome bias in saying so, the decision Del Rio made was clearly defensible based upon the analytics and based upon what we know about Del Rio’s team (which, very significantly, is far less than Del Rio knows).
Ultimately, Coach Del Rio had the last word, even if only because the move worked out. He deserves it, after all, given that his neck – like that of all coaches – is on the line every week.
Del Rio’s decision worked out. Good for him (since I have no strong rooting interest in either team). But what, if anything, does it tell us about investment success?
We all know that the outcomes in many activities of life combine both skill and luck. Football is obviously one of them. Jack Del Rio could have made a solid decision, well supported by tangible, intangible and statistical factors (as I think he did), and still lost the game. The world is wildly random. With so many variables, even the best process can be undermined at many, many points. Thus we’d all be wise to pay close attention to the “fundamental attribution error,” one aspect of which is the error we make when we overweight the role of the individual and underweight the roles of chance and context when trying to explain successes and failures.
Remember back to the 2012 NFL season when a controversial (to say the least – it even forced the NFL to settle its lock-out of its regular officials) touchdown call at the end of the game gave the Seattle Seahawks a 14-12 victory over the Green Bay Packers. That blown call had an enormous impact on the regular NFL officials, the teams, the standings, and on gamblers. If Seattle’s desperation pass had correctly been ruled an interception, Green Bay – as 3½ point favorites — would have won by five, covering the spread. Instead, the replacement officials’ call shifted the win from those who bet on the Packers to those who took the underdog Seahawks. Remarkably, that result meant that as much as $1 billion (that’s with a “b”) moved in one direction as opposed to the other. It’s perhaps the ultimate “bad beat.”
That a clear win was eliminated due solely to the almost unbelievable incompetence of the replacement officials isn’t remotely foreseeable, as least in the particular. It was just dumb luck for Green Bay and for bettors, and why it can be so frustratingly difficult to wager successfully on both our favorite teams and, not coincidentally, our favorite stocks. It may seem like the system is rigged, but investing successfully is just really hard (like gambling), as Tadas Viskanta so eloquently and regularly points out and I routinely reiterate.
Investing involves an almost bizarre consortium of luck and skill. Understanding the relative contributions of each can help us assess past results and, more importantly, to anticipate future results. Luck (i.e., randomness) is a huge factor in investment returns, irrespective of style or manager. “Most of the annual variation in [one’s investment] performance is due to luck, not skill,” according to CalTech professor Bradford Cornell in a view shared by all the relevant experts (Nobel laureate Daniel Kahneman talks about it in this video, for example). Even more troublesome is our perfectly human tendency to attribute poor results to bad luck and good results to skill – classic self-serving bias.
As a consequence, in all probabilistic fields, the best performers dwell on process. This is true for great investors, great gamblers, and great coaches. A crucial element of Nick Saban’s football coaching success at the University of Alabama is what he unoriginally calls the “Process,” a simple but profound way of breaking a difficult situation down into manageable parts. If the players focus only on winning each play, with no focus on the end result, success would take care of itself. It’s a matter of preparation over payoff. In Tuscaloosa, it borders on obsession (or perhaps goes way past it).
Maintaining good process is really hard to do psychologically, emotionally, and organizationally. But it is absolutely imperative for consistent and ongoing investment success.
Due to what Kahneman calls the “planning fallacy,” our ability even to forecast the future, much less control the future, is extremely limited and is far more limited than we want to believe. In Kahneman’s now iconic book, Thinking, Fast and Slow, he describes the planning fallacy as a corollary to optimism bias (think Lake Wobegon – where all the children are above average) and self-serving bias (where the good stuff is my doing and the bad stuff is always someone else’s fault). Most of us overrate our own capacities and exaggerate our abilities to shape the future.
The planning fallacy thus manifests itself in our tendency to underestimate the time, costs, and risks of future actions and at the same time to overestimate the benefits thereof. It’s at least partly why we underestimate bad results. It’s why we think it won’t take us as long to accomplish something as it does. It’s why projects tend to cost more than we expect. It’s why the results we achieve aren’t as good as we expect. It’s why I take three trips to The Home Depot on Saturdays. We are all susceptible – clients and financial professionals alike.
As Nate Silver emphasizes, forecasting – which is imperative to outperformance – is really hard. There are simply too many variables and too much uncertainty (Donald Rumsfeld’s infamous – but accurate – “unknown unknowns”) for forecasting to be anything like easy. As I keep repeating, information is cheap; meaning is expensive. For example (per Leonard Mlodinow), we are tricked into thinking that random patterns are meaningful, we build models that are far more sensitive to our initial assumptions than we realize, we make approximations that are cruder than we realize, we focus on what is easiest to measure rather than on what is really important, we build models that rely too heavily on statistics without enough theoretical understanding, and we unconsciously let biases based on expectation or self-interest affect our analysis. Yet we’re still surprised when things don’t turn out as we’d like.
The great investor Seth Klarman, founder of the Baupost Group, makes a terrific insight at this point. “Value investing is at its core the marriage of a contrarian streak and a calculator.” The contrarian streak means that a good investor must be willing and able to do something different from what the consensus is doing. Still, investment success draws a crowd and dilutes future success, meaning that one can never “rest on her laurels.”
The issue is complicated further in that sometimes the consensus is right. If the movie theater is on fire, you should run out the door with everyone else, not in. So adding the calculator part is crucial. That’s why the probabilities are so important in analyzing Del Rio’s decision. In the investment world, being a contrarian makes sense only when it leads to a mispricing between fundamentals and expectations. That is how it becomes a market opportunity. Finding active managers with that outlook as well as the ability and the psychological strength to execute it well is astonishingly hard.
But there is yet another problem: a portfolio’s results are largely dictated by overall market performance during the applicable time period. Thus the more risk-averse strategies will tend to generate better returns in a difficult market by protecting the downside and the reverse will also tend to be true, that managers with higher risk tolerances will be more likely to succeed during periods of strong market returns. The conventional method of mitigating this dilemma is to “risk adjust” the results, comparing nominal returns with volatility, but this approach is uncertain at best in that volatility and risk are hardly the same thing. Even worse, the conventional investor response is to chase performance and thus buy what is risky and what’s less so at precisely the wrong times.
Everybody – advisors and clients alike – wants the same basic thing in the end: high relative returns with a minimum number of sleepless nights. Human nature being what it is, however, investors are often their own worst enemies. Risk-averse investors, for instance, should want to underperform the chosen benchmark in a bull market. It implies a strategy of risk management that will protect them when, inevitably, the benchmark heads lower. Of course the benchmark is going to outperform that strategy in a sustained bull market. But what tends to happen, again, is performance chasing. It’s hard for us to live with our choices when it’s easy to see people who are apparently doing better, even when their goals, situations and choices are very different.
Again, in all probabilistic fields, like investing and football, the best performers dwell on process. A great football player focuses upon a good approach, good effort, good technique and good execution. If he does that – maintains a good process – he will suffer negative or ineffective plays and losses sometimes (even when his process is excellent), but the wins will take care of themselves overall. Again, maintaining good process is really hard to do psychologically, emotionally, and organizationally. But it is absolutely imperative for investment success.
The NFL, like investing is so often, is utterly results oriented. It’s a “Just win, baby” league in the same way that investors and investment managers are generally judged by their performance. But results are far from entirely controllable. That’s why process is so important. A good process can’t guaranty success, but it will offer the best possible odds. And that’s the most we can hope for.
1 Ironically, it was just over three years ago that Crabtree was unable catch three straight fades in the opposite end zone of the Superdome, any of which would have given the San Francisco 49ers a Super Bowl victory over the Baltimore Ravens.
2 Saints quarterback Drew Brees called the move “ballsy,” while Saints head coach Sean Payton said the call “wasn’t unexpected” and “on the road, in a game like this, I probably would have done the same thing.”