Prediction Charlatans

Faust
5 min readDec 14, 2020

“A good forecaster is not smarter than everyone else, he merely has his ignorance better organised. “ — Anonymous

Talk is cheap, but it shouldn’t be when it comes to predicting the future. In fact, it should be expensive and should always have a cost. That cost is what one is willing to put on the line. Win for being right, or be punished for being wrong. Without this mechanism, it is difficult to weigh the opinions that everyone has when it comes to the future. Some of these opinions may be absurd, some may be plausible but weighted incorrectly, and some may be both plausible and accurately weighted. The problem is that nothing differentiates these opinions when there is no skin in the game. Ultimately what propels the process of improving accuracy is the feedback one gets when they win or lose. In order to survive — that is to not go broke — one must meticulously assess their bias, their ignorance, and everything in between. It’s no longer about what one wishes to be true, but rather about what is the best assessment of outcomes. One-off outcomes by themselves may not tell us much about a person’s thought process, but repeated success or failure can help to diminish the role of luck and signal a higher value of their decision-making.

The other problem with predictions is that there is no incentive to change your mind when nothing is on the line. In the absence of incentives, one could argue that the risk of emotional attachment and bias is actually more likely. This drives a tendency to double down on one’s position even when new information may make such an action seem foolish. This is not the case with money.

Enter the bettor. Bettors are a mix of everyone from degenerate gamblers to sharp quants. It’s largely irrelevant what the ratio of degenerates to sharps is, because as long as a handful of sharp bettors can take advantage of the odds they believe to be incorrect, markets stand a better chance of finding a more accurate equilibrium in comparison to pundits that don’t wager anything. The smart bettor is willing to update his understanding of the world as time goes by. The same cannot be said for the pundit. “I changed my mind” doesn’t sit well with the talking heads, but it would if they had something to lose. This speaks to a larger problem in our world regarding the shaming of views one may have held from years or decades past. A politician may indeed be a flip-flop or may have genuinely changed their mind on an issue or policy. How to tell the difference? Well, it’d certainly be easier if we adopted the gambler’s mindset. Strong opinions, weakly held.

Improving The Signal Problem

“He who lives by the crystal ball soon learns to eat ground glass.” — Edgar Fiedler

Probability is hard. When we look at history, all we see are events that actually happened. We don’t see — and often don’t know — the other possibilities that could’ve happened, nor the likelihood of any of them. Whatever the probability is a priori, it can be hard for people who think in terms of certainty to grasp that unlikely events are also possible. Emotions can make it difficult to judge the unlikely, especially in non-repeated events. Knowing the unlikely probability of being in an airplane crash is of no consolation when you’re in the plane that’s crashing. Even so, something that happens 1% of the time is something that can still happen. Underdogs do win games after all.

But the prediction problems we face right now aren’t just ones of statistical illiteracy. Even when some professions grasp the nature of uncertainty they still let their bias distort the outcome they wish to see, whether that’s for political reasons or otherwise. The pressure to conform can also supersede the desire for accuracy when one has nothing on the line. It’s why stock analysts would rather be right with the herd than be wrong alone. Being a contrarian only pays off when you’re right.

So what can be done about these problems? Similar to the tragedy of the commons, we live in an age in which producing information, especially broadcasting opinions, is just too damn easy. The worst part about risk-free prognosticators is that they are incentivized to say the extreme to increase their engagement. If they’re right, they make a name for themselves. If they are wrong, people forget. It doesn’t matter if you incorrectly call for recession repeatedly if you are able to correctly call one. But it would matter if every time you called it incorrectly you lost money. To paraphrase Taleb, “don’t tell me what you like, show me your portfolio.” It’s too easy to have conflicting incentives when we talk about uncertain events and don’t attach any value to them. Imagine if we had a way to easily access every pundit’s track record, or better yet, if it was displayed next to them whenever they talked. Would we listen as much to someone who had made hundreds of predictions and managed to get 5% of them right?

These frictional costs would do wonders for public discourse. Think of the internal conversation for those spewing thoughtless information too freely. “Gee I stand to lose material wealth today if I make stupid predictions again.” How many more talking heads would say “I don’t know” rather than contribute more verbal diarrhea to the pile of noise?

Humans crave certainty, hence the popularity of those who are charismatic and extremely wrong versus those who voice doubt and are partially right. Though it may feel weird, we would all do better to push back on these valueless opinions, even in our daily lives. Asking people to assign probabilities to what they are saying may feel like an affront to them at first, but if they really start to think about it, ultimately it benefits them just as much as you. This allows you to call their bluff too, which is a much needed feedback mechanism. If someone assigns an absurd probability, make them bet it. In a world of fake news, distrusted institutions, and tribalism, this is desperately needed. Money isn’t everything, but it certainly serves as a better scorecard for who is actually thinking critically about the future rather than just saying what’s on their mind. Betting may be one of the only ways to stem the tide of noise brought on by cheap and easily disseminated opinions. Let’s make these opinions expensive.

--

--