My Metastock strategy

Option_Attack · May 30, 2003

There are at least 20 systems of this same type here, with the same or better backtest results:

http://www.wealth-lab.com/cgi-bin/WealthLab.DLL/browsechartscripts

They are of the DipBuyer, Band Violation, Simplicity, Keltner, Bollinger, hold 2-3 days, etc. type.

They work but are difficult to trade, unless you have an order system that can enter 100s of limit orders of fairly small size.

jaiko · May 31, 2003

if you trade nyse stocks the open print on a chart isnt always the open price of the specialist------so you cannot backtest an opening strategy for nyse stocks (at least not with q-charts or tradestation)!!!

good luck

DaveN · May 31, 2003

jaiko,

Good point! I often try to get around this by having my code check the volume of those prints or the 1 minute bar, depending on my data resolution.

Clearly, this is easier to do on the thicker NYSE stocks, and it's certainly not foolproof, but it helps weed a lot of those errant ECN prints out. I often require that the first bar/print have at least 5000 or more shares in it.

rtstrading · May 31, 2003

Quote from baggerlord:

I didn't do any optimization, and my backtest has produced thousands of trades.(100+ for many of the stocks in 4 years.) I am putting my money on the line, starting earlier today. Not much money of course, just enough to see if it works. I'll keep you guys posted, even if I fail miserablly.

"I updated my system to sell all stocks that don't reach the target within 20 days. I lowered the target to 2%."

No Optimization?
How did you come up with your profit target? Yes, that is optimizing in order to get the best results. And how did you arrive at 20 days? Yes, that is optimizing...why not 18 days?
Why not 2.5% target?

Think about it!

Bob

baggerlord · Jun 2, 2003

Hello again. I hope everybody had a pleasant weekend! So far so good on my strategy. I raised my target back up to 3% and slightly tightened the entry requirements. I guess you could call this optimization, but when I am doing it for all of the 100 stocks I think it is OK. You asked why the 20 day stop, and this is because I am interested in options right now and am seeing how this would perform on short-term options. I don't intend to trade that way, I was just curious. My system does appear to work, and last Fri I made three trades that I closed today, all winners. Based on my testing it should net about 4% a month, so I may use it, but I am pretty good at traditional TA and will probably just stick with that. This one seems like way too much work and too many trades. I like getting to know just a few stocks and playing them for bigger profits. Thanks for all the feedback!

G_Morgan · Jun 2, 2003

It is absolutely essential that you use a 1 day delay if you are going with the open, -probably-, since some of your indicators (moving averages, macd, stochs), use the end of the current day data... which you don't have until the close.

baggerlord · Jun 3, 2003

I do use a 1 day delay.

kgharris · Jun 3, 2003

Is there a reason you are not sharing your system parameters?

ges · Jun 3, 2003

Quote from baggerlord:

My system does appear to work, and last Fri I made three trades that I closed today, all winners. Based on my testing it should net about 4% a month, so I may use it...

That's a very short real time test. Does your system trade both long and short?

g

baggerlord · Jun 3, 2003

It netted about 4% a month on average over the last 7 years. Basically all the system does is identify stocks that have dropped sharply while in an uptrend/risen sharply while in a downtrend. Position is taken and held for 3%, or sold if target isn't reached within 20 days. It works but it isn't nearly as fun or profitable as finding my good ol' triangles so I doubt I will ever use it.