Optimal Short Put Strategy Backtest by Specialkayme

I've run around 1,000 backtests so far in an attempt to figure out an optimal short put strategy, and I'd appreciate any feedback. I've seen conflicting conclusions with other sites (I've seen some backtests that show 6 delta as the optimum while others are 15 and 30, some that show super short dte, most in the 30-45 dte range, while others show 90-120). So I just decided to run the data myself using eDeltaPro. All of the data is from 2005-2022 (some data that far back isn't super reliable, fyi) and uses EOD data. All of the trades don't include commissions, and are done sequentially, so you only have one position on at a time. Not that I'll trade that way, but to get an apples to apples comparison I figured it was the best.

To start, and in an attempt to avoid "form fitting" the data, I took seven indexes (SPX, SPY, NDX, QQQ, RUT, IWM and DIA) and ran backtests on dte (7, 15, 30, 45, 60, 90 and 120) and delta (5, 10, 15, 20 and 30). Initially, no early exit (profit or days) and no stop loss. Here are the results, broken down by index, DTE and Delta.

So far the data is appearing to show that the higher the delta, the greater the return on capital (not surprising), but also the larger potential drawdown (also not surprising). But what his surprising is the DTE is appearing to show a smile in profitability, where the profitability is the highest at 7 dte and 120 dte, and the lowest around 30 dte (with some variability). I expected to see more of a linear chart. Plus, it appears to almost counter the work of Tasty, showing max profitability and max theta decay to be around the 30-45 dte mark. But it should be mentioned that this is ROC, and not return on margin. The margin requirement of a 120dte position is a fraction of the margin requirement for a 7dte position. So when you consider return on capital in a margin account, the longer dated options will take a significant bump up.

Next I took what I considered the best performing groups (dte of 7, 60, 90 and 120, as well as delta of 15, 20 and 30) and compared what it would look like if I added a stop loss at 100%, 200% and 400%.

Breaking the information down based on delta was the most notable, as you're clearly seeing a smile like result, where a 100% stop loss reduces overall returns (not surprising) but a 400% stop loss appears to increase overall returns.

Overall, the stop losses decreased the win rate, increased the average loss, decreased the average hold period, and typically slightly reduced the average win. But there were mixed results on the "max loss" number, as some stop losses appeared to reduce the max loss, while others appeared to have little change. I suspect the reduction in hold period improved the results, as the test immediately reestablished another position, and with a shorter time period allowed you to compound gains faster. But that's a guess. In any event, I wasn't expecting to find that data. I was expecting to find that all stop losses would under perform over hold to expiration, as it's the data that Spintwig and Option Alpha typically show. Maybe this is an outlier or not. Not sure. But interesting.

Then I compared an early profit (25%, 50% and 75%) when compared to HTE (100%). The data shows (unsurprisingly) that taking an early profit, overall, increases your profit. What was surprising was that it was not the same across all indexes. Some (QQQ) showed significantly greater returns (over double) by taking early profits. Others (SPX) showed a greater profit by taking gains at 75% over 25% (oddly, the data wasn't mirrored by SPY). But, overall, it shows greater profit for all by taking some sort of early profit over holding to expiration.

This is again shown when you compare just the DTE comparison. 90dte appeared to react differently than 60dte, for example. Some were a little more flat than others, but all early profit taking was better than holding to expiration.

I still want to check what will happen if you add some trend component (index above 200SMA, for example), and what the impact would be for spreads vs. naked.

Beyond that, I'm interested to see if there are any "stacking" results that come out, such as a 200% stop loss under performs a 400% stop loss when both don't have an early profit component, but flips when you take profits at 50%.

Anyway, that's all I have completed so far. Just was hoping to get some feedback on direction. Thanks for taking the time to read my ramblings.