Yesterday I received an “interesting” (abusive) email from someone we’ll call “Mr Angry”, in response to my previous blog post with a case study of how Market Internals student Christopher reduced drawdowns by 59%.
The email from Angry is ridiculous, so I won’t publish it here, but in his ramblings he did raise a good point which I want to share with you, because it’s a huge problem in our industry and one that personally bugs me too, here it is:
“Your case studies are cherry picked, fake, overfitted and over optimized.”
OK Angry, you got me there…
I have to admit to the “cherry picked” claim, because it’s the only example I was sent, so I “cherry-picked” it out of a group of 1… my bad!
However, would there be a lot of additional value in me sharing an example of it not working when I clearly explained in the blog post that it doesn’t work for all strategies?
I doubt it.
Here’s the exact statement from my previous blog post:
“Now of course Market Internals techniques don’t work for all strategies (nothing does)…”
I do think we can learn from analysing things that don’t work though, so if I come up with an example of it not working that has a good lesson behind it, I’ll be sure to share it.
As for the suggestion that the case study is “fake” – as far as I can tell Christopher is a real person and not a mythical creature, as are all the other traders sharing their results here, including Maxim Schulz who used Market Internals to take out 3rd place in the World Cup Trading Championships of 2017. Those results are audited by a 3rd party and are definitely not fake, but as I mentioned in my previous blog post:
“don’t take our word for it – prove it to yourself.”
On tradingmarketinternals.com we have a case study, with ACTUAL code and backtest reports, so you can download it and prove it to yourself. Don’t just believe me, run the code on your own trading platform, apply it to your own strategies and see if it works for you.
I’m not sure how much more transparent we can be than that!
As for the “overfitted and over optimized” claim, this is definitely a huge problem in the industry, and one that also pisses me off at times (but not enough to fire off abusive emails…).
In the Market Internals training we talk about the importance of Out-of-sample validation, because overfitting is a challenge and risk that can never be eliminated completely.
For this case study, it’s tough to say if it is or isn’t over optimized without seeing the code, or without being a creep lurking quietly in the corner of the room, watching traders as they apply these techniques to their own strategies, so again, my suggestion is to prove it to yourself.
If it works, then that’s great – you just added another technique to your trading arsenal.
If it’s not for you, that’s fine too, you’re free to move on.
You can find the case study, code and backtest reports at tradingmarketinternals.com
Have a great weekend (yes, even you Mr Angry!)
Are you looking for a proven solution to CUT DRAWDOWNS and significantly improve the performance of your trading strategies?
02 March 2018