Today, QID, SDS, and SMN all gapped beneath their stops for this system. This resulted in QID and SDS losing 2-3% more than projected, which is not terrible. What is terrible was SMN, which resulted in a loss 2x the projected 1% risk amount.
Since the gap down would have triggered the stop market orders, each position was closed at today’s opening print.
The latest trades have really put a hurtin’ on the system performance. As of today, the system is in a 12.5% drawdown from its highs. I will have to check to be sure, but I’m fairly certain this exceeds the max drawdown from highs realized during backtesting.
Trading is such a psychological challenge. Remember that I changed the stop to 8% and increased the position size to 2% as of the November 25 SDS trade. Wouldn’t you know that the two winners since then were smaller than expected, and all three losers were much larger than expected. In fact, the combined losses from these three closed trades were more than the combined losses from the seven previous losing trades. Ouch.
The system is still showing gains of 9.41% since inception (ignore the cash amount, I forgot to update it). But to look at the spreadsheet, would anyone want to trade this system? Probably not. As I’ve said at least a half-dozen times, this is why I’m tracking the system on the blog. I hope that watching this in real-time has helped or is helping traders think about the difficulties in developing and then trading systems.
On the bright side, I’m hoping the next trades are going to be from the long side.
No New Entry Signals
Shed, I think you’ve learned as much as we have in going through this exercise. I appreciate your posts and showing us how you work on developing a system.
Now that the experiment has proceeded this far, may I ask how you feel now about modifying your stops and position size during the experiment? I’m just curious.
Boca, I was asking myself the same questions. I still think it was the right thing to do. I think that today’s gap down was just bad luck.
However, I will go back and refigure things using the original stop and position sizing and see what difference it would have made.
It really wasn’t the stop or the position sizing that hurt this last round of trades as much as the fact that they were all about 2% from being stopped out. So when it gapped down, they suffered hard.
As I think more about it, the stop allowed these last trades to stay in the game, without stopping out, for a decent amount of time. They did have room to breathe. Unfortunately, they all exhaled at the same time 😉
I understand exactly what you mean and understand why you changed it.
I’m not too familiar with testing methodology, but I had always thought that testing design assumed that parameters don’t change during the actual test, unless your test subjects die from the new drug…. hahaha oh man. 🙂
Well, with backtesting, one can change the parameters in midstream to see how it affects the results. Its called curve-fitting, or over-optimization, if it is applied with a heavy hand.
An example of extreme curve fitting may be that you have a system that buys and sells based on the 50 day average. Well, through testing, you realize that actually a 43 day moving average works best. In my opinion, that is curve fitting.
What we have done I don’t believe is nearly over-optimizing the system, as we have not altered the entry or the exit criteria, just the stop and position sizing.
OK, thanks for the explanation Shed, it makes sense now. Curve-fitting is a term I had heard in other contexts (non-stock related).
I really hope you do find a set of parameters that delivers the goods on a long term basis.
Just curious…is the criteria for entry/exit…
…and, the technical parameters that go to make up this “system”…posted anywhere ?
If so…where ? tia
.
Bummer. Maybe VIX needs to be under 40 for any trades or something. 12.5% isn’t that bad but it is good it happened on paper and before you made this your covestor system. Go kill covestor anyway!
so I didn’t initially buy SMN, but noticed it today and bought it.
Can buying a smaller amount, like if you intend on having a 2% position, buying 1%, and then buying instead of selling on the stop work as effective money management?
Then if it goes even lower, buy even more?
As my name implies I’m a stubborn fuck who hates losing, and tends to not want to follow stop losses, because in my nature, if I bought low, not buying lower would not only be admitting I’m wrong, but it would be flipping on the position, and giving up on it AFTER it has already lossed.
Basically I want some system that can work by adding lower… Any advice?
Alf, near the top of my page, under “About” you’ll find “Big Bamboo Trading System.” That link will take you to the first few introductory posts and should answer most basic questions. From there, go to categories, and click on *Big Bamboo, which houses all posts on the system.
Other than that, I’ll be happy to answer most questions.
Aumana, thanks. I am thinking about adding a short strategy for equities to my covestor account. I’ve taken a few of the trades in my covestor audited account because Scottrade couldn’t locate shares and Tradestation could.
Stubborn, I think I can develop a rough framework to help you, but it will be fairly complicated and require at least basic knowledge of expectancy and system trading.
Let me know if you are game, and I’ll put up a comment. I just want to make sure you are really interested because it will take me some time to flesh it all out here.
Thanks for the reply !
From what I’ve read in the Bamboo thread…you are unwilling to post the “entry criteria” for the system.
Is this true…or, did I miss it ?
————————
I see the RSI(2) exit rules…
I see the position sizing stuff…
I see the ATR stop loss stuff…
I did NOT see the “entry criteria” !
.
Alf, that is true. The entry criteria, as far as I can tell, is novel, and I haven’t disclosed it. Briefly though, it seeks to catch an overdone pullback within a trend. It is somewhat a mean reversion system, but it fires much more quickly than a lot of the MR systems I’ve seen.
…got it !
But…it’s NOT an RSI(2) “extreme” in a prevailing trend…correct ?
I think I read where you said it was NOT an RSI(2) entry.
——————–
FWIW…you can also use a short-term MACD or Price Oscillator to catch entrys in a pullback.
For example…you can use the Slow Line of a MACD as a “Trend Indicator” of short-term momentum…and watch for the Fast Line to pullback to the Slow Line and then turn UP/DOWN…to go LONG/SHORT ! I’m looking for a minumum of 2 days of “slope opposition.”
It’s a kewl little way to use a MACD…other than the typical, garden variety “signal line cross”…or, the “zero line cross” !
———————
Good Luck with the Bamboo system !
.
ooops !
minumum = minimum !
.