In the spirit of full disclosure, I will tell you that I trade my IRA in this account, which is now available for others to mirror. Of course you can mirror it without using a retirement account, and you'll find it a bit more conservative than the standard approach. What I personally do is trade retirement funds in the new model, and all other investment funds in the old model.
(Note: When you look at the model on the Covestor site you will notice that the performance chart shows that it started in December, but there is a flat line until the first of February. That's because I didn't actually start trading until February. Sadly, the Covestor record started earlier, so the SP500 got a headstart on me. If the dates had started at the same time we would be much better off relative to the market.)
Good Trading...
Mr Smith
ReplyDeleteWhat performance can be expected with the IRA system versus the first model?
Regards
Ken
Hi Ken,
ReplyDeleteThanks for asking. Naturally I can't make any hard predictions about the future, but I will do my best to give you some things to think about. And, as a quantitative trader I do a tremendous amount of testing an analysis, so what I can share has more to do with the past than the future... Hope that makes sense.
The IRA system (as I call it) is basically the same with two primary differences. First, we can't use margin. Second, there are no short trades. By definition, that has two effects on returns. No margin means we will have either smaller position sizes, or less positions, or both. In my case I have chosen to have smaller positions sizes so I can have more positions. No short trades, simply means less trades.
When I test all of this historically the backtests (ten years worth) are pretty much identical on a "per trade" basis. Each trade has a return of approximately 3%, and the win ratio is about 78% (again, this is the backtest). Ultimately then, the IRA system is simply different because we get less trades.
Oops... I published that too soon.
ReplyDeleteSo, the big difference is in the quantity of trades we get. Backtesting would indicate that we still get about two-thirds of the trades, as compared to the legacy system. But there is one more twist.
On a given day I might enter twenty trades (just an example), and only three of them will be triggered by hitting the limit price. The problem is that I don't know which three. With a margin account I can enter them all; but with a non-margin account I can only enter the trades that my account can afford if they all trigger. As a result, I am required to "enter" less orders. I might only be able to enter ten, and I might miss the three that actually trigger.
The net result of all this, again, is less trades to put our money to work. I am still trying to measure this impact and be able to get accurate backtests. Once I have that I will be posting the more precise backtest numbers. Until then, my estimates are that I get about half the number of trades, which is still quite good. Given the "per trade" characteristics I described above, this makes me comfortable enough to trade my own IRA money with the system.
I hope that helps. Stay tuned for more numbers before long.
Good Trading...
Mr Smith
ReplyDeleteSo for the IRA system if there are 1/2 as many trades does that mean the returns of the IRA system will approximate 1/2 of the first model?
From the covestor website it appears your model (first model) returns 32.36% SI but the average sub-account (the investor) received a 17.72% return SI. Can you explain this disparity and can I invest with you direct to get better returns?
Regards
Ken
Hi Ken,
ReplyDeleteMy current observation and best guess on the IRA system is that it will be a little better than half the performance. I prioritize the trade candidates and choose the ones that historically have had higher probability. Of course it's impossible to know if the future will bear out the same as the past, but that's how it has historically turned out.
As far as the disparity from the average subscriber, that would be a good question to ask Covestor. I am not allowed to see any subscriber information so I can't really analyze it. Some it, for sure, is due to timing. For example, if someone subscribed when I started (June-2010), they might be up 35%. But if they started a week ago they might be up only 3% from just the last few trades.
Last I saw I had 13 subscribers, and at least half of them have joined in the past couple of months, so there is no way they would have the full return from the beginning.
Aside from this, there may be some other details that Covestor could provide. My guess is that this is the biggest factor.
I hope this helps.
Mr Smith
ReplyDeleteCan I invest directly with you managing the account?
Ken
Ken,
ReplyDeleteThanks for asking. Sorry to say, not at this point. To date I don't do any private hedge fund management. I've been asked many times, and that may eventually happen, but not at this time.
Bill