You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Patterns that work in many instruments
- Thread starter tim888
- Start date
If the same system is profitable in many markets, then the probability that it has an edge is higher, although I have no idea how to prove this.
The problem with backtests is there is always a chance that the positive returns are just a fluke and not a result of identifiable reoccuring market condition. Having a larger number of trades on the backtest or the idea working on a longer historical period increase confidence that it's not a fluke. Testing the idea (both on a backtest and in real trading) on many markets increases confidence in it similarly to testing it on a longer period of time: one gets more observations.Quote from intradaybill:
If the same system is profitable in many markets, then the probability that it has an edge is higher, although I have no idea how to prove this.
One fair way to do it is to pay based on the number of hours or days the work takes. Go to a local it contractor website and check what the rates are like. If you want to "fix" the price, you can further ask the programmer how long teh work will likely take and base the agreed price on that.Quote from tim888:
Does anyone know how much will it cost to develop such software for offline use (not real-time)?
Quote from tim888:
I found someone at the local College to develop code for me but I have no idea how to price the work because I have never done this.
Does anyone know how much will it cost to develop such software for offline use (not real-time)?
What exactly kind of tool do you want? Maybe it already exists or an existing tool could be supplemented?
Quote from Mr_You:
What exactly kind of tool do you want? Maybe it already exists or an existing tool could be supplemented?
I'm familiar with one tool that could be supplemented for what I want. Company promised to do that in a future release but no result yet. I think what I am looking for is the key to finding an edge.
Quote from LeeD:
The problem with backtests is there is always a chance that the positive returns are just a fluke and not a result of identifiable reoccuring market condition. Having a larger number of trades on the backtest or the idea working on a longer historical period increase confidence that it's not a fluke. Testing the idea (both on a backtest and in real trading) on many markets increases confidence in it similarly to testing it on a longer period of time: one gets more observations.
What you wrote makes sense but there is a definite lack of a rigorous foundation.
Quote from tim888:
For example if I identify a pattern that works well in 4 or 5 currency pairs instead of just one, will it tend to be more robust? Has anyone done any research in this area? Any software that does this automatically?
Maybe this is something relevant: http://bit.ly/sBNuZn
Have you actually purchased and used this product? It's $3k (!) /year or $10k (!!) /unlimited use.