5 replies to this topic

[Tor]

Hi there, I just ran some numbers onto my system and came up with some crazy numbers/targets. To be clear, I wont be trading off em, as I dont believe them, but wanted to post them, just to be on the record. Top in the SPX 1690, within 4 months!!! IF, and its a BIG IF, I trade off this, I will only ever do this via call options. I wont be doing it right now ie today. accordingly, please take this as just a public record comment of my system, and of no other value. Thanks.

Edited by Tor, 18 June 2007 - 08:22 AM.

[Sentient Being]

What is a "pilot system"? I'm using Metastoch at home, was working on some new, mechanical trading systems over the weekend. Programming is a pain for us non programmers.

[Tor]

What is a "pilot system"?

I'm using Metastoch at home, was working on some new, mechanical trading systems over the weekend. Programming is a pain for us non programmers.

Pilot, as in I am testing it.

I am simply using a system based on the MACD and COT combination, also using the ADX line to spot a trend move.

It is based on info dating back to the early 1990's so not very long hence my doubts.

[Sentient Being]

Thanks. I work in IT so I was thinking in term of the name of an application or a device. This weekend I built a system based on various stochastic requirements and a trailing stop, Now I'm testing it against a broad range of stocks to see if any are in a mood or mode to work well with the system after optimization.

[Tor]

Thanks. I work in IT so I was thinking in term of the name of an application or a device. This weekend I built a system based on various stochastic requirements and a trailing stop, Now I'm testing it against a broad range of stocks to see if any are in a mood or mode to work well with the system after optimization.

My conclusions so far i think I need an adaptive system according to envoronment. They will change depending on conditions.

[Sentient Being]

I guess there's all sorts of ways to skin a cat. I started out with the idea that the more data the better...train a system on up markets down markets and sideways markets. I think that's the common accepted practice. Also hold back enough date to test and see if it works well on data it wasn't trained on. But lately I've come to the conclusion that I'm not looking to trade up markets, down markets, and sideways markets. Rather than make the system fit the stock or abandon the system, I'd like to find a stock trades well with the system in the past several months. Find a stock that for whatever reasons is presently inefficient and trade it until the inefficiency and statistical advantage change. Looking at various items such as win-loss, total trades, return, draw-down, avoiding profit spikes as opposed to profit mounds, etc. I think a lot of times you can simply look at the price action and see if it's behaving as it has recently or if the market is changing.