Thursday 10 November 2011

The death of momentum

As Ian Dury once said, there ain't half been some clever bastards, and one of them is Daniel Kahneman, who won the Nobel Prize for Economics in 2002, even though he's not an economist, he's a psychologist.

Kahneman is the star of Michael Lewis's piece in the new issue of Vanity Fair, a story that fills in a little hole drilled by Lewis's book Moneyball. Kahneman, as you might expect of a man who knocked off a Nobel in his spare time, had the answer to a question that Moneyball left hanging, namely, why, if baseball coaches had spent their entire lives watching baseball, had they got player selection wrong so often, and by so much?

The solution lay in cognitive psychology and something Kahneman called 'the availability heuristic', which was the notion that human judgement is often based on the most easily recalled information. He explained this by means of one of his experiments: a roulette wheel was rigged to stop on one of two numbers, 10 or 65. Kahneman asked the groups he assembled in front of the wheel to write down the number they saw. He then asked them an unrelated question: 'What is your best guess of the percentage of African nations in the UN?'

The average answer of the groups whose wheel landed on 10 was 25 per cent, and of the groups who landed on 65 was 45 per cent. In other words, the unrelated number affected their guess.

Kahneman called this 'the anchoring effect'. He conducted lots of other strange experiments too, like creating a character called Linda, who 'was bright, majored in philosophy and who was deeply concerned with discrimination and social justice'. He asked his subjects which statement was more true: i] Linda is a bank teller ii] Linda is a bank teller and is active in the feminist movement. Eighty-five per cent of people opted for number ii even though it is logically impossible [if number ii is true, then number i must be equally true].

Daniel Kahneman developed all of this stuff into 'prospect theory' which was about economics and ultimately, many years later, won him the big one. A Harvard undergraduate called Paul DePodesta, who had been hired by Billy Beane at the Oakland As, became interested in it. Along with Bill James, their maverick statistician, they exploited the 'willful ignorance' of the baseball player market, and revolutionised the way the game was measured.

Michael Lewis thought of all of this when he stumbled on a letter written to him in 1985 by Bill James. 'Baseball men have an entire vocabulary of completely imaginary concepts used to tie together chance groupings,' James wrote. 'It includes momentum, confidence, seeing the ball well, slumps, guts, clutch ability, being hot and my all-time favourite, intangibles'.

Kahneman's work seemed to answer Bill James's question. Baseball coaches often based their judgement on nebulous concepts and 'instincts' rather than empirical evidence of the kind rooted out by Bill James.

Cricket does it too. Australia dropped Simon Katich for being too old in the face of all available evidence: in the previous three years, he was the only Australian batsman to average over 50, had scored more runs, home and away, than anyone else, and two payers who kept their places, Ponting and Hussey, were older than Katich. There are plenty of other examples: how long did Steve Harmison's 7-12 affect opinion of his game?

During an insane day at Newlands yesterday, when Australia were bowled out for 284 and then South Africa were bowled out for 96 and Australia's second innings score stood at 21-9, Robin Jackman asserted on commentary that 'South Africa have the momentum here'.

How did Jackman make that judgement? Probably because, in his mind, South Africa taking 9-21 was further forward than the knowledge that Australia were 209 runs ahead on a day when 20 wickets had fallen for 128 runs.

Momentum is king of those nebulous concepts affected by the availability heuristic. In truth, not even Daniel Kahneman could tell you what's going to happen at Newlands today, other than that someone's going to win, because there's almost nothing to compare it with. Try one for yourself: Next time Australia bat, which will be in Johannesburg, how many do you reckon they'll score? Not that easy is it, when your availability heuristic is all over the place.


Ian & Nina Graham said...

Dear O.B.
The extent of your hinterland beyond the boundary is always a source of wonder and delight to your readers.

But in this case, I think you're being a bit hard on Mr. Jackman.

In general terms, I think 'momentum' is a more tangible reality in cricket than in most sports. How often do we watch a decent attack toil in the field for a day and a half, and then their batsmen falling in a heap on the very same wicket ?

Specifically as to Newlands: I watched the Australian collapse in real time, though I'd blinked and missed the South African one. Philander and Morkel bowled with skill and intensity. And although the wicket encouraged them, I saw nothing to suggest it was a beast. Conversely, Australia undoubtedly lost it - the wickets of Hussey and Haddin being the most notable instances. So although the calculation of ultimate advantage was an intriguing one, I think Jackman's assessment was fair enough. There are all sorts of recoil involved in a situation like this, and to start an innings with a good lead having dismissed the opposition cheaply, and then be shot out yourself, is an enormous psychological reverse. Conversely, to recover from all out 98 to produce that performance in the field shows pretty clearly who was on the up-beat.
Jackman was at not that stage predicting the result, notice. It was a snapshot judgement.

But to return to the top - keep them coming ! Perhaps you could a host a sort of 'rule of six' competition - "challenge the OB to make THIS the crux of a blog!"

diogenes said...

you might want to games last a few hours. Test matches last 5 days...(or so). Over 5 days, different psychological dynamics can come out. As an exsample, headingley 1081 or the other Ashes test in 1896(?) - too lazu to look it up. Yes teeams can be crushed by a huge first innings...other nteams get inspired. The ruoles need more analysis. And what about Dev Mnalcolm's match against the Saffers?

Chrisps said...

That's a really stimulating piece. I've had a ponder on Declaration Game about how selectors and others weigh up the value of a batsman who benefits from the most visible source of good fortune: dropped catches. I suspect they go with the momentum argument - scored some runs so has the confidence to do so again.
Baseball does insulate itself from this particular Issue by counting hits arising from missed chances as 'errors' and not crediting them to the batter.

Tom Redfern said...

you're all far too clever- the blogger and posters in the comments section. I've read this blog three or four times and I still don't get the link between heuristic availability and 21-9.
What is the connection?

Pay per head bookmaking said...

sorry dear blogger, but I have to disagree with you on this, I agree with Nina Graham's comment, you were a little bit hard on Mr. Jackman, he is a good man

pay per head bookmaking said...

grrr I accidently hit the post button, I was saying that Mr. Jackman is a good person, I met him years ago and I realized this