Test everything?

There is a wonderful scene in Ted Lasso where Ted hustles the club’s snooty owner at a game of darts. The clip is all over TikTok so I recommend you find it, but the gist of it is that the arrogant owner thinks a dumb American like Lasso couldn’t possibly be any good at darts and proceeds to lose a major bet to Ted who it turns out is somewhat of a darts shark due to his early childhood days of spending Saturdays playing the game in a bar with his Dad.

During the soliloquy Ted launches into a monologue about a Walt Whitman quote about the need to be curious rather than judgemental.  That’s a great sentiment by Whitman and I thought about it the other day when I passed a man on Broadway with the following text emblazoned on his T-shirt, “Test everything, results matter.” The man was clearly an engineer and on the face of it the statement was highly laudable. After all if you are building a skyscraper you definitely do want to to test everything to make sure that the structure is sound.

But when it comes to trading that type of judgemental thinking will ruin you. How many back tested systems succeed in real life? Let me save you the trouble of guessing. Zero. Yes, exactly zero backtests will make money in real life going forward. There is no doubt that some of them may make money for a little while, some may do so for a matter of years but eventually all backtested systems fail.

In Jack Schwager’s book Unknown Market Wizards he interviews Marsten Parker a systems trader who managed to compound an astounding 58% for 30+ years running yet in matter of months Parker nearly wiped out all his life’s earnings when the market regime turned horribly against him and all his back tested strategies turned sour at once.

I have always believed that there is no backtested system that I could not destroy in five minutes flat by simply changing the market regime going forwards. Have a breakout strategy? I’ll just put you into an endless loop of false breaks until you are either out of money or forced to bet so small that it becomes meaningless. Silver had 17 false breakouts at one time and it's still basically a shell of its former self from the heydays of 1970s and 1980s. Want to trade trends? Good luck to you if you find yourself in the years 1966 through 1982 when the Dow hit 1000 so many times and fell that most investors were convinced stocks were dead forever - thus the famous Business Week cover. Ah but perhaps you like mean reversion well then have a ball shorting Internet stocks during the dot com bubble. You would be margined out in one day.

Backtests in trading are truly the biggest waste of time. They prejudice you in two ways. They make you believe that what you’ve just tested on past prices will work forward or - and this is even worse in my opinion - they make you dismiss some strategy ideas because they performed poorly in the past when in fact they may be the perfect fit for the upcoming future.

In either case the human need for verified tested, sanitized results of market performance is truly one of the colossal mistakes of human arrogance in which we  constantly conflate the repeatable, replicable and highly stable results of the physical universe with the utterly unpredictable, fanciful and illogical aspects of human behavior.  

The engineer’s slogan of Test Everything is 100% applicable to the world of nuclear reactors. Nuclear reactors of the 2050s will be exactly the same as the nuclear reactors of the 1950s if  not in degree but certainly in kind as the principles of nuclear fusion will not change. But market principles change all the time. Stocks can trade at 25 times earnings if investors think rates will decline or they can trade at 5 times earnings if they think rates will rise. Earnings themselves may not matter at all because  trading is a game of expectations rather than reality and if expectations change deeply enough they can change reality. Trading does not comply with the inviolable laws of nature  but rather with the artificial social constructs that we create every day.

That’s why when it comes to trading the slogan needs to change from Test Everything to Try Everything, because you never know. You never know what you may see in real time market action that could help you tweak your strategy to the current market environment. You never know how a particular idea you had could morph into something new and wonderful that you would have never found in a back test. You may even discover that some of your best trading setups like Viagra are a completely accidental find.It has certainly happened to  me. A trend trading strategy I worked on was a perpetual loser, but when I applied it to news trading it became  a near infallible winner.

Don’t test everything, try everything. Marsten Parker eventually turned it around, by scrapping all his old ideas and finding new market systems to trade. He is back to making money now because he learned that profits are made forward not backward. So next time someone shows you a beautiful 45 degree equity curve, project it forward to its most likely destination - zero. 

---