Hello All,
I'm not new to options trading (have a few years experience), but more wanted to ask for advice on what the best strategy would be given the results of my model. I felt this warranted its own post given the length of the post - but I am happy to repost into the safe haven thread if mods feel that's best.
I've been playing around with equities and machine learning models for a couple years now and have a decent model that I would like to start testing with paper trading options but am not sure which parameters to set up.
My model essentially uses a handful of predictors to predict whether SPY will go up at least X% from Monday's open during the week. I say X% because it uses the median weekly high from the Monday open (calculated in the training period to ensure no lookahead bias) - which typically is between 0.9% and 1%.
The model performs quite well across equities but especially so with SPY, QQQ, and IWM. Using a 10 year/1 month rolling training and testing period, I have achieved relatively high accuracy relative to baseline in predicting whether the ETF will hit 1% during the week. You can find my results below.
|**Ticker**|**Strategy**|**Weeks Traded**|**Win Rate**|**Avg Return/Trade**|**Avg Max Profit/Week**|**Avg Hurdle Imposed**|
|:-|:-|:-|:-|:-|:-|:-|
||||||||
|**SPY**|Strat 1 (Scalper)|171|72.51%|0.3339%|2.1458%|0.92%|
|**SPY**|Baseline 1|518|49.61%|0.1328%|1.3506%|0.95%|
|**QQQ**|Strat 1 (Scalper)|155|68.39%|0.1618%|2.5775%|1.25%|
|**QQQ**|Baseline 1|518|54.25%|0.2068%|1.7988%|1.25%|
|**IWM**|Strat 1 (Scalper)|159|67.92%|0.3213%|2.5997%|1.33%|
|**IWM**|Baseline 1|518|51.93%|0.0855%|1.8699%|1.35%|
Here you can see for all 3 tickers the model is able to predict with 13% (QQQ) - 22% (SPY) better than baseline. Average return/trade means what happens if you have a strategy of simply selling when that hurdle is hit and we see that the average return is higher for both SPY and IWM, but not QQQ. We also see that Avg Max Profit/week (that is the average max profit possible) tends to be higher than baseline as well.
If you have a strategy where you buy at Monday open and hold until the end of the week, results look like this
|**Ticker**|**Strategy**|**Weeks Traded**|**Win Rate**|**Avg Return/Trade**|**Avg Max Profit/Week**|**Avg Hurdle Imposed**|
|:-|:-|:-|:-|:-|:-|:-|
||||||||
|**SPY**|Strat 2 (Holder)|171|58.48%|0.5827%|2.1458%|0.92%|
|**SPY**|Baseline 2|518|57.34%|0.2361%|1.3506%|0.95%|
|**QQQ**|Strat 2 (Holder)|155|54.19%|0.4886%|2.5775%|1.25%|
|**QQQ**|Baseline 2|518|58.88%|0.3543%|1.7988%|1.25%|
|**IWM**|Strat 2 (Holder)|159|54.72%|0.3758%|2.5997%|1.33%|
|**IWM**|Baseline 2|518|53.09%|0.1424%|1.8699%|1.35%|
Win rates - that is weeks where you are profitable are roughly comparable between the model and baseline, but the average return is higher in for all 3 ETFs.
My question is based on these results, what's the best strategy to trade with options? My initial thought is to buy ATM 30DTE calls at open on Monday when there's a signal and sell when the underlying hits the minimum hurdle, but I understand that becomes sensitive to tail risk and a high win rate would need to compensate for that.
Would a bull call spread be better here, and then closing the spread when the hurdle is hit? Would love to hear how people would trade given they had this information. Perhaps options is not even a suitable strategy here.
Also feel free to ask any questions or criticize my results as you see fit.