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.
Any pure mechnical traders or system traders?
- Thread starter feng456
- Start date
I'm too stupid to trade any other way.
I believe in dialogueing publicly for the benefit of the entire community, so I don't do PM. If you can post without revealing proprietary details, have at it.
ok well im a mechanical trader and i build systems. i'll give you an overview of my procedure and maybe you can see something wrong that i dont.
1. i come up with an idea first..i.e. buy everytime x crosses y.
2. I backtest a year's worth of data to find the optimal stops and targets (if it works at all).
3. I crossvalidate with 3 months of virgin data, from the past to confirm similar results with similar parameters.
4. I go live and trade exactly in everyway as I would when i backtested so that by keeping the variables the same it can be as relevant as possible.
When I backtest I only count target fills when price as exceeded the target. Since I trade the ES, slippage has not been an issue for me. Also I've live traded my latest system for 4 months now with 30% of expected performance so either I'm:
a) in a losing streak, which a similar thing did occur once in my backtesting
b) missing something crucial and I don't really have an edge
c) doing something wrong fundamentally in my entire approach.
EDIT: I also make it a point to keep things as simple as possible so that I can actually keep track of all the variables and approach everything scientifically and systematically when backtesting. Therefore my systems generally have few rules and is more like 'if a crosses b...' than 'if a crosses b and it's a monday and yesterday was a high range day of 20+ pts and 50 ma has been hit and...'
1. i come up with an idea first..i.e. buy everytime x crosses y.
2. I backtest a year's worth of data to find the optimal stops and targets (if it works at all).
3. I crossvalidate with 3 months of virgin data, from the past to confirm similar results with similar parameters.
4. I go live and trade exactly in everyway as I would when i backtested so that by keeping the variables the same it can be as relevant as possible.
When I backtest I only count target fills when price as exceeded the target. Since I trade the ES, slippage has not been an issue for me. Also I've live traded my latest system for 4 months now with 30% of expected performance so either I'm:
a) in a losing streak, which a similar thing did occur once in my backtesting
b) missing something crucial and I don't really have an edge
c) doing something wrong fundamentally in my entire approach.
EDIT: I also make it a point to keep things as simple as possible so that I can actually keep track of all the variables and approach everything scientifically and systematically when backtesting. Therefore my systems generally have few rules and is more like 'if a crosses b...' than 'if a crosses b and it's a monday and yesterday was a high range day of 20+ pts and 50 ma has been hit and...'
Quote from feng456:
...I also make it a point to keep things as simple as possible so that I can actually keep track of all the variables and approach everything scientifically and systematically when backtesting. Therefore my systems generally have few rules and is more like 'if a crosses b...' than 'if a crosses b and it's a monday and yesterday was a high range day of 20+ pts and 50 ma has been hit and...'
You think you are being simple? Seems to me as though you are more than willing to confuse yourself.
Feng, please identify the time frame you are charting and trading from. Thanks.
and how many trades are in back tests
btw-if this cross that never works for me..
also-please realize, that you ar stepping into probably most efficient and liquid market. it's like driving F1 in manhattan with zero driving experience. you have to compete with best of the best in the world...
i trade for about 10 years,about 5-6 simple mech system,but i'm not going to touch that thing- ES, if you ask me. too hard..
btw-if this cross that never works for me..
also-please realize, that you ar stepping into probably most efficient and liquid market. it's like driving F1 in manhattan with zero driving experience. you have to compete with best of the best in the world...
i trade for about 10 years,about 5-6 simple mech system,but i'm not going to touch that thing- ES, if you ask me. too hard..
Crosses are only to be borne.