Thursday, May 1, 2014

What the Last Ten Months of Market Action Say to A Quant

The goal in quant investing is to use data to shade the odds in your favor. There is nothing magic and nothing guaranteed but if you can get enough bets down when the odds are in you favor, even as low as 51-49, and if you have a big enough bankroll to handle the semi-variance when betting at and above the Kelly optimum* you will end up with a lot of chips.
Here's an example from PastStat:

we are reluctant to make it open source , but anyway , higher highs and more dizzy highs on $SPX 

ten higher highs on $SPX monthly chart
ten higher highs on monthly $SPX chart

below the trading odds , for the next 1/2/3/4/5/6/12 months, when $SPX makes ten or more higher highs in row, data since 1950

Exit # Wins % Wins Avg% Med% Avg Win % Avg Loss % Pay Off Max Loss %
t+1 22 16 72.7 1.87 1.78 3.49 -2.43 1.44 -8.54
t+2 22 18 81.8 4.08 4.31 5.31 -1.43 3.71 -3.54
t+3 22 21 95.5 6.57 5.19 6.95 -1.47 4.73 -1.47
t+4 22 21 95.5 8.20 6.42 8.79 -4.25 2.07 -4.25
t+5 22 20 90.9 9.90 9.19 11.09 -2.05 5.42 -4.02
t+6 22 22 100.0 12.33 12.28 12.33 NA INF brokerage
t+12 22 22 100.0 24.88 24.33 24.88 NA INF brokerage

below the prior instances of $SPX making ten or more higher highs in row, since 1950.....MORE

We are in a bull market until further notice. But you knew that.

*Previously:
Repost: Dreamtime Finance (and the Kelly Criterion)
What Proportion of Your Bankroll Should You Bet? "A New Interpretation of Information Rate"

One of the quantfathers:
"How did Ed Thorp Win in Blackjack and the Stock Market?"
Markets, Risk and Gambler's Ruin
Journal of Investment Consulting: Interview With Edward O. Thorp

And many more. Here's the synopsis: