I’m undertaking a 1000-day reinvention project, blogging here daily to track my progress. In Thursday Thinker, I share a smart idea or theory.
I’m thinking about buying Brent Donnelly’s new book Trade Outside the Box: Advanced Thinking for Professional Traders, but the Kindle version is $37 which is a lot for an ebook. I heard Donnelly interviewed on Forward Guidance and liked what I heard so I’ll probably go for it.
Here’s one of the ideas he shared, related to the mathematical concept of ergodicity:
Reminiscences of a Stock Operator [a fictionalized memoir detailing trader Jesse Livermore’s experiences] reads like an awesome “how to trade manual,” but he went bankrupt three times and committed suicide so it’s not a good model for how to trade…. [Livermore] was an unbelievable trader but also had one fatal flaw, which was he didn’t manage his risk and he faced ergodicity problems….
If you’re trading a strategy that’s extremely convex in the wrong direction, at some point you’re going to blow up. And so really internalizing that and just never having strategies that are that way is another thing I talk about in the book.
In an ergodic system, the average of a process over time is the same as the average of all its possible states at a moment in time. A single person performing an activity many times will yield the same average outcomes as many people performing it once.
Many traders rely on ensemble probability — what happens to a large group of people at the same time) — rather than considering time probability — what happens to one person over a long period of time. Traders may assume that because a strategy such as holding the S&P 500 over time has a positive expected value across all market participants over time that they will have a positive result.
The issue is that in a non-ergodic system, individual outcomes over time can differ drastically from the group average. You can be wiped out at one point in time and that means your average return cannot be the same as the average of all traders. You can no longer trade once wiped out (without raising more money somehow).
Donnelly would say you want positively convex trades, not negatively convex. In a positively convex trade, your potential upside is explosive while your downside is capped. In a negatively convex trade, you make small and predictable gains most of the time but rare, catastrophic losses can completely wipe you out.
Nassim Taleb promotes positive convexity trades like tail-risk hedging, which pay off explosively when they do pay off, but lose small amounts of money consistently over time. Other positively convex trading approaches include long volatility trades such as buying options straddles and strangles or VIX call options before a major structural shift (such as a company earnings report or the CPI release), and venture-style micro-cap investing, where you spread your money across many highly speculative assets, and if only one pays off massively you make money.
What are some negatively convex trading approaches? One is selling naked options such as deep OTM puts or calls to collect small, steady premium income is one — and similar to my options wheeling. The win rate can be very high, but in a crash, you can be bankrupted if you get assigned on your naked puts all at once, or face assigned calls for stocks you don’t own all at once.
Another is the “martingale” style system in which you double down on a losing position as it drops, assuming it must eventually mean-revert. A single prolonged trending move against the trader can lead to margin calls and liquidation.
What I’m doing is negatively convex (but not extremely)
The options wheel is negatively convex, however, not extremely negatively convex the way I do it.
Most of my short puts generate small profits, but occasionally I will face a situation where many or most positions end up with large negative losses. I have multiple ways I manage this risk:
- I secure my short puts with money held in a money market fund, so I will never be wiped out by margin calls and bankruptcy, though I may end up holding long positions in stocks with large unbooked losses
- I tend to sell at low deltas of around -.25 to -.20 lowering the probability that I will get assigned
- I allocate only 50% of my total available funds to short puts, so that if there is a marketwide crash, I cannot be totally wiped out, and I have funds to deploy at lower stock market prices
Still, I’m pretty uncomfortable with the asymmetry of the trading I’m doing, and I’m feeling it right now as my semiconductor short put positions are under pressure (I have $MRVL, $SMH, $INTC, and $TSM — a bit overloaded for sure!)
I have experimented with bull and bear spreads as an alternative to selling puts outright, but it hasn’t captured me for ongoing trading. I’ll probably stick with what I’m doing, and if a crash happens, work my way through it by selling calls on assigned positions, and watching charts for new opportunities to sell puts and make money.