Probability of Predictability

 In a recent discussion with a very experienced trader looking for new ideas to deploy, I seemed to pique his interest with a concept I hold dear (but in the spirit of this blog, that I am willing to share with the universe!) - that of gauging the probability of predictability. At first, he was convinced I meant using volatility as a gauge of how predictable/deterministic the market is behaving, but I had to clarify: a highly volatile market can still exhibit predictable signatures, and a "boring" market could be highly unpredictable in what move it makes next. The example I gave was my experience trading the FOMC announcements - there's the primary big move when the interest rate decision is announced, the secondary move after Powell starts speaking, then the establishment of the prolonged market bias. Sure you can play the first move, which is often more of a gamble for large profits, but is it really repeatable?

More often than not, windfall profits will be evened out by Father Time as it exerts its benevolent force with the power of Probability. It's sort of like getting 5 heads in a row in a game of coin flips and expecting to have a huge "heads winning percentage" by flip 100. That edge will surely disappear by then. In addition, gambles are a double-edged sword. Instead of a windfall, one could experience an outsized loss, whose recovery is much, much harder than would have been the case by simply not taking the trade (more of a matter of just mentally getting past the lost opportunity) or taking a more reasonable trade for a smaller loss (which is mathematically much easier to recover from).

By gauging predictability, I am not saying that "this security has x % chance of reaching price target Y". I am simply saying that given its behavior and price level, its subsequent behavior is much more predictable than it was 5 minutes ago, when the setup was just forming. Whether you're gauging a price target, a breakdown, or a simple mean reversion is more of a function of your specific strategy and trading style. The main concept here is to first assess whether the damn thing is moving predictably, i.e. in a non-random fashion, and then to exploit that non-random behavior. This then becomes an umbrella heuristic to filter out undesirable market conditions that are often hard to generically define across the board (like earnings announcements, choppy price action, etc).