The famous MIT Blackjack Team based their systematic approach to blackjack on the idea that they wanted to risk significantly more capital in situations where they had a quantifiable edge. We can use a similar strategy to put more of our capital to work in high probability situations, and we don’t have to worry about getting thrown out of the casino.
The MIT Blackjack Team had players that would place table minimum bets for hours waiting for a deck to become advantageous for the players. When that occurred, they would signal a teammate to join the table and bet the maximum. This allowed them to risk as much money as possible under conditions where they knew they had an edge, without alerting the casinos to their strategy.
In a recent post on Trader Edge, Brian Johnson explained a quantitative trading strategy that is very similar to what the MIT Blackjack Team was known for. Brian suggests that traders should use a filter to identify the trades with the biggest edge for the trader. This filter will produce very few trades, so it won’t be a tradeable strategy itself, but it can be used to identify the most potentially profitable trades.
Brian’s Filter Example
The example that Brian uses to illustrate his concept is a basic strategy that buys securities that are experiencing short-term pullbacks during long-term uptrends. While he doesn’t give the specifics of this strategy, he does tell us that it produced 790 trades with a profit factor of 3.78 during his backtesting period.
Brian explains that adding a Stochastic filter to his strategy has a varying effect on its performance depending on what value is used for the filter. He produces a chart that shows that using any value between 60 と 100 for the filter has zero effect on the strategy. しかし, as the value of the Stochastic filter is lowered from 60, it begins to impact the strategy by increasing the profit factor and lowering the number of trade signals.
Brian finds that the sweet spot for his Stochastic filter appears to be a value of 20, which produces a profit factor of 7.77 上 324 取引. He explains that this is where most traders force themselves to make a difficult decision.
Filter or No Filter? Both!
Brian now has two strategies. The strategy with no filter makes less return on a per trade basis, but signals many more trades. 反対に, his filtered strategy makes far more profit per trade, but generates fewer signals.
What he suggests is the same idea that the MIT Blackjack Team would utilize: Trade both strategies! His simple, but brilliant concept is to take all of the trades that the unfiltered system produces and increase his position size on the trades that meet the criteria of his Stochastic filter. This will allow him to risk more capital in higher probability situations, and less capital in lower-probability situations.
This concept could be applied to almost any trading strategy and filter combination with proper backtesting. Brian makes a point to remind us that we should be cautious not to increase position size past an acceptable overall risk.