What to do when you find the Holy Grail

As I have mentioned in several interviews, I am always looking for new strategies. One area that fascinates me is stock options. Because it is difficult to get good data and to do backtests, I believe that there are good edges here to be found. A few weeks ago, I found myself with lots of time and having read presentations on options, I went into a testing frenzy.

After lots of work, I found a strategy that greatly exceeded my expectations. It seemed to be the Holy Grail. I ran the strategy by my trading buddy, Steven, and he thought it was great too. We were so excited!

Option backtests are very complicated. Besides having to deal with the usual issues, one must deal with expirations, additional data like the Greeks, lack of the option trading, and many more issues.

The 7 steps to do when you find the Holy Grail.

Source: https://www.flickr.com/photos/spiritual_marketplace/2207966935

Step 1: Assume you did not find it

In the excitement of finding a strategy that makes lots of money, one forgets there is always a possibility that you did not find it. The better the results the more likely there is a problem.

Step 2: See Step 1

I really must emphasis this point. The more excited you get the more you make up stories to justify why your strategy is working. We started to fall down this path.

Step 3: Check the code for future leak

The first obvious step is to check the code for future leak. This means are you looking into the future to get in or out of a position. For example, you check to see if today is a down day, if so, you get out of the open. Yes, I have done this before.

After checking and checking, I could not find any future leaks. I even made the code more pessimistic, and entered longs on the ask and exited on the bid. For shorts, I entered on the bid and exited on the ask. The strategy still held up great. My excitement continued to grow.

Step 4: Check outsized winners

Are a few trades giving you all your winnings? There were some big winners but nothing that outsized that if they did not occur the strategy would be a losing one.

Outsized winners are also good for checking for future leak. None could be found here.

Step 5: Check random trades

This is similar to step 4, but now I just check some random trades. Again, everything looked fine.

Step 6: Paper trade it

Steve got so excited that he actually placed a real trade to see if he could understand it better. Of course, that trade turned out to be a big winner. I did not have that much trust yet. I kept reminding him and myself that there probably was a problem. Even though I really hoped not. Instead I placed a trade with InteractiveBrokers paper trading count. Often going through a trade, one will find the problem of future leak or something else. Opening the order, I found no issues

Step 7: Get external verification

This is the best way. Find someone else to code up the strategy and see if they get the same results. I have a research friend who will often verify results for me for free. But given the difficulty of testing options strategies I was ready to pay him.

Holy Grail Lost

Steve and I were going through the trades yet again. Then we found it. We found an exit that did not make sense. The data seemed wrong. Oh, oh. How often did this happen?

Upon investigation, it turned out there was a problem. It was two-fold. One, I had exported the data out of the options database incorrectly. Two, my code was dealing with data holes incorrectly. Because of these two issues, the strategy went from making great money to barely making money.

As much as I was expecting this, it always sucks when it happens. But better to find it now, then in real trading.

I drank from the wrong cup. I chose… poorly.

https://www.youtube.com/watch?v=qIitjokEJwg

Good quant trading,