Wealthall: Election
Thursday, 3 January 2013
Nick Wealthall spells "poker"
I ruddy love American politics. There, I’ve said it. I’d probably love UK politics, too, if I could get over the fact that MPs are like a bunch of corrupt supply teachers.
I’ve always loved it in a geeky over-analytical way. I’m more interested in how voters are influenced, and the prediction of the outcome, than I am in the policies or who actually wins.
The internet is a wonderful thing. It allows us to play poker in our pants, and it’s also brought the murky world of psephology (the study of political elections) and election prediction into the mainstream. This election was interesting because we saw a battle between mainstream media talking heads and geeky stats analysts. The talking heads – perhaps biased by the need for a compelling story – constantly stated that the Obama/Romney campaign was “too close to call”. While they were doing this, probabilistic analysts like Nate Silver and Drew Linzer were consistently predicting an Obama win, mostly claiming it had an 80% to 90% chance of happening.
This gap produced some bizarre Nate Silver bashing – mainly from Republicans who didn’t like the fact he was predicting that their guy would lose. It’s a bit like in poker when the internet generation began cleaning up all the cash and the previous generation sat in denial claiming the “new way” of playing couldn’t possibly be right (spoiler alert: it was).
Seeing what you firmly believe, or want to happen, being contradicted can lead to so some odd behaviour. A great example of this is the mostly mental website www.unskewedpolls.com, a Republican site which claimed all the polls analysts were using were incorrect and biased. This isn’t a party political point, by the way. Democrats claimed exactly the same thing before the Bush/Kerry election, which they lost.
All of this produced amazing market inefficiency and a great opportunity to make a profitable bet – sorry, investment – which I’m unashamed to admit I took advantage of. If you trusted the analysts and the “maths”, you knew that Obama’s chance of re-election was between 80% and 90% for the entire campaign. However, depending on when you bet, you could invest as if he had as little as a 60% chance. That is a big dollop of positive expectation that doesn’t come along very often.
The sad thing is that because the maths geeks were right (big fucking surprise – those dudes are always right. Damned facts!), the cat may be out of the bag. The knowledge that once belonged to a few is now in the mainstream. Nate Silver and Drew Linzer have done countless TV shows and other media as “the men who called the election” and it will be very difficult to find such a market misconception in the future.
Again, the parallel with poker is there – for a while, just a few young poker players had broken the game down and knew how it should be played, and they cleaned up. Then, as time passed, thanks to shared knowledge, poker forums, training sites and so on, the knowledge turned from expertise into general awareness, making the market more perfect and making it harder to find big positive expectation spots. In simple terms, the games got tougher. Maybe this is the price we pay for a more efficient, prosperous world – the demystification of everything!
One thing that continues to fascinate me through all this is people’s total inability to grasp the idea of probabilistic thinking. Or if you want to put it another way, people struggle to get beyond a black and white result. After the election, the analysts that had predicted it were carried shoulder high in the media as New Age Gods. The fascinating thing is, what if Romney had won? Well, apart from a war in Iran, a lot of millionaires celebrating, and a lot of sad liberals, the stats geeks would have been utterly dismissed. After all, they’d got it wrong – they’d have failed!
This would have totally misunderstood their predictions and the basis of their work. When someone using probabilistic thinking says there’s an 85% chance of something happening, it means it won’t happen 15% of the time. A Romney victory could have been a completely plausible if relatively unlikely event.
All of us humans struggle to think like this. Our brains constantly try to simplify everything to one answer. We’ve been programmed over thousands of years to merely assess whether something we meet should be eaten, run from, befriended or sexed. This limits our thinking in almost everything.
It’s completely essential to think probabilistically to do well in poker. If you’re only thinking about whether you have the best hand or not in a black or white sense, you’re going to make some really big mistakes. You will almost never know your opponent’s specific hand, so instead of guessing their hand, assigning a probabilistic range is correct. You shouldn’t be making a bluff because “you think they will fold” but because you’ve assessed the probability of them folding.
Most crucial of all is to break the link between viewing your results and the outcomes of hands as the determining factor in whether you made a good decision. Just because your opponent called you doesn’t mean your bluff was bad, and just because you ran into the top part of their range doesn’t mean your call was bad, and so on. It’s not easy to do, as you’re conditioned to think in terms of black and white results, but with a little practice anyone can start thinking probabilistically in poker and in life.