Dynasty, in Theory: Creeping Determinism

A reminder that we did not, in fact, know it all along.

With championship games on the books and the small matter of winning leagues finally dealt with, it’s time to turn our eyes towards more important things. With postmortems to be conducted and offseasons to be won, I’ll be devoting a column to each this week.

First, I wanted to go over a few examples of extremely flawed thinking people might have engaged in during the year. I’m sure a lot of you are going to be nodding along in agreement as I go.

Example #1— How on earth was Rob Gronkowski able to fall to the third round in so many early drafts? The guy has handily outproduced every other tight end in the league— including Jimmy Graham— whenever he’s been on the field, and he was healthy. It’s not like there were so many great alternatives- Graham had a history of dealing with nagging injuries, and Julius Thomas was the third option at best in his own offense.

Example #2— Why on earth did no one question the fact that Emmanuel Sanders was going at the 6/7 turn in most drafts? The guy he was replacing was coming off of back-to-back top-10 seasons, and the offense he was joining was the most prolific in history, (including becoming the second offense in history where three players topped 10 receiving touchdowns and the first offense where four players hit the mark).

Example #3— So Marshawn Lynch was the #5 RB in 2011, the #4 RB in 2012, and the #4 RB in 2013. And yet he fell to the 2nd round in many drafts and was considered outside of the top 5 pretty much anywhere. And, of course, he finishes the 2014 season as the #3 fantasy RB. Quelle surprise!

When we look back at each of these situations, it’s so unbelievably obvious how badly the consensus screwed everything up. Fantasy owners couldn’t see their own nose as it was staring at them from the middle of their face.

But here’s the twist: the faulty thinking isn’t making those mistakes in the first place, it’s assuming those mistakes could have been obvious ahead of time if we’d only paid more attention.

Rob Gronkowski was regularly going in the 3rd round because opening the season on the PUP list was a very real possibility, as was returning and playing most of the season at less than 100% of his pre-injury form.

Emmanuel Sanders was going at the 6/7 turn because he was a 4-year veteran who had never topped 750 receiving yards and who had just signed a contract for a mere $5 million a year.

Marshawn Lynch was falling to the end of the 1st round because he was disgruntled with his contract, holding out, threatening to retire, and because his offensive coordinator was openly discussing how they were going to use more of a committee approach at the position.

If any of those three players had busted spectacularly, I could have just as easily led this column with anecdotes saying “how could anyone have taken those players as high as they did when it was abundantly obvious that they were going to disappoint?” The ingredients were in place to explain either success or failure, but our minds are designed to forget the factors predicting outcomes that never materialized.

This phenomenon is called “creeping determinism”, also known as “hindsight bias” or the “I knew it all along” effect.

When Determinism Creeps

Regular readers of this column will not be surprised by the fact that our memories are bad. Like, really really bad. Or that our faulty mental software is riddled with errors and prone to fallacious reasoning. In fact, it could be argued that these biases are not a bug, they’re a feature.

With that in mind, the idea that we are prone to remember the past as being more predictable than it actually was makes perfect sense. If we believe that the past was predictable if only we’d done a better job, then naturally we’ll assume that the future will also be predictable. We will be more motivated to do a better job, assuming that our efforts will easily be rewarded.

Creeping determinism is often the first ingredient of a particularly dangerous bias cocktail. If we believe the past could have been anticipated if we’d only done a better job, we’re going to believe that the future is likewise predictable if we only put together the right clues.

Believing that the future is predictable leads to us taking more aggressive risks going forward. Aggressive risks, by their very nature, are more likely to fail, but will also return bigger rewards when they succeed. After a few of those risks pay off, selective memory starts to kick in, causing us to remember our successes more than our failures and reinforcing to us the idea that we are quite good at what we are doing.

This confidence that we are quite good at what we are doing quickly creates a positive feedback loop, leading to more aggressive risks, which leads to more memorable payoffs, which leads back to more confidence. This is how economies are crashed; when conducting the postmortem of past crashes, we see the “obvious” signs and feel like we now know what to look for, so we quickly feel safe taking aggressive risks again because now we’re better able to guard against failure.

These risks rely on the idea that the original crash was predictable, and therefore future crashes will likewise be predictable. But if the original crash was predictable, it would have been widely predicted; the fact that it wasn’t is all the evidence we need that any predictability is an illusion, an artifact of 20/20 hindsight, a post hoc rationalization.

Hindsight is 50/50

Fantasy football is a game of making calls. We draft one player over another because we think he will have a better season. We start one player over another because we think he will have a better game.

If we drafted our entire team by taking the top two remaining players by ADP and simply flipping a coin to decide between them, we’d be right roughly half the time. Coin flips are 50/50 events, and we can easily recognize that to be true.

The problem is that, because we are biased to see the past as more predictable than it was, we are also biased to see our correct predictions as the result of skill rather than luck. The past was predictable, we made a prediction, our prediction was correct, therefore we are good at making predictions.

I’m not above this in the slightest. In fact, if there’s one thing that I hope I have impressed upon you this season, it’s that no one is above any of this. We are all prone to cognitive biases, and even being aware of them doesn’t obviate their impact. I knew all about confirmation bias when I fell prey to it with Stephen Hill.

I was banging the drum for Rob Gronkowski as hard as anyone this last offseason, so it’s tempting for me to say that I predicted this season from him. Fortunately for me, I have proof that I did not. To quote myself: “The real problem with Rob Gronkowski in 2014 is not the question of how many games he’ll miss, it’s the question of how effective he’ll be in the games in which he plays. It’s true that in 2012, Adrian Peterson showed us that it is possible for a player to return from an ACL tear in a short timeframe without losing any dominance. It’s also true that players like Wes Welker, Robert Griffin III, Rashard Mendenhall, and Heath Miller have shown us that sometimes players who return from ACL tears in a short timeframe do, in fact, lose a little bit of their dominance. I would rather have Gronkowski playing in 75% of his games at 100% of his peak effectiveness than Gronkowski playing in 100% of his games at 75% of his peak effectiveness. The former is a the difference-making talent you’re hoping for when you draft him in the second round. The latter is a pale shadow who would find that loft ADP awfully difficult to justify. As a result, news that Rob Gronkowski will likely play in week 1 is great, but what I’m really waiting to hear is news that Rob Gronkowski looks like his normal self. Until we see that happening, there will always be some extra risk surrounding Gronkowski in 2014.

Rob Gronkowski started looking like his normal self around week 5 of the season. From that point on, Gronkowski’s dominance really was an inevitability. But would it be correct for me to say that I predicted it in advance? No. I highlighted it as a possibility, even a probability. I noted that Gronkowski was a fast healer and a potential difference maker. But for all of my praise of Gronkowski, I never advocated taking him as the first tight end off the board, or drafting him in the first round. And, thankfully, I have archives of my thoughts on the subject so I can remind myself just how guarded my optimism really was.

In fact, I would even argue that “guarded optimism” was the correct state regarding Rob Gronkowski. Up until week 5, it looked like he very well might have been a big disappointment who needed an extra year to regain his pre-ACL form. Predictions should not be a black-and-white affair, because the reality of the NFL is drawn in shades of gray.

If we view our correct predictions as obvious and inevitable, and our incorrect predictions as easily avoidable errors, we will quickly begin to minimize the role that random chance plays in fantasy football. By believing we are better owners, we will become worse owners.

How do we combat creeping determinism? Well, now we’re coming full circle. My first practical advice of the season was a reminder that our memories are faulty and a suggestion that the best way to combat that was to keep a journal to accurately record our thoughts at the time we thought them. It seems fitting that my last practical advice of the season be a reiteration of that point.

Being predictable with the benefit of hindsight is not the same thing as being predictable. The best way to combat a belief that you “knew it all along” is an objective reminder that, in fact, you did not.


More articles from Adam Harstad

See all

More articles on: Dynasty

See all

More articles on: Strategy

See all