There are seveal monte carlo methods,including scattering trades,or equity with discrete and continuous distribution,what's the pros*cons of each of them?
I found many guys discussed monte carlo methods,but none talked about it,it should be useful for the monte carlo beginner as me to tell the differences and pros and cons of them,any reference/guide/suggestion please?
Monte Carlo Simulation
Hi, I did the following on an excel spreadsheet I prepared and that I'm currently testing:
I copied the trades I got from tradestation's report
I scrambled these trades allowing recycle (in practice I filled new arrays of the same length picking randomly from the trades sequence)
I analysed the returns and drawdowns of the new series on which I applied a fixed fraction money management approach
I plotted the % return grapgh and %DD grapgh to have a glance at the scenario.
Done this I made following considerations:
I have 635 trades starting in year 2000 on EURUSD on forex and these are the trades I worked on.
On the original trade sequence I applied money management and got the final equity.
then I picked only the first 300 trades and constructed 5000 new trades sequences of 635 trades each picking random from the 300 trades I chose.
On these sequences I applied the same money management approach to see what outcoming scenario I could have.
In practice I considered to be in the situation I had only the first 300 trades I was trying to figure out what could happen after 635 trades.
The final scenario was worse than the real scenario i.e. the simulation gave poor probabilities to get the return i'd got from the real sequence.
Do you think this is due to a better market behaviour after the first 300 trades or do you think the model i constructed is poor?
Bye
Smodato
I copied the trades I got from tradestation's report
I scrambled these trades allowing recycle (in practice I filled new arrays of the same length picking randomly from the trades sequence)
I analysed the returns and drawdowns of the new series on which I applied a fixed fraction money management approach
I plotted the % return grapgh and %DD grapgh to have a glance at the scenario.
Done this I made following considerations:
I have 635 trades starting in year 2000 on EURUSD on forex and these are the trades I worked on.
On the original trade sequence I applied money management and got the final equity.
then I picked only the first 300 trades and constructed 5000 new trades sequences of 635 trades each picking random from the 300 trades I chose.
On these sequences I applied the same money management approach to see what outcoming scenario I could have.
In practice I considered to be in the situation I had only the first 300 trades I was trying to figure out what could happen after 635 trades.
The final scenario was worse than the real scenario i.e. the simulation gave poor probabilities to get the return i'd got from the real sequence.
Do you think this is due to a better market behaviour after the first 300 trades or do you think the model i constructed is poor?
Bye
Smodato
Maybe I got part of the answer myself, I simply tested the system using the last 300 trades instead of the first 300 and the simulation after 635 trades gave extremely good results, much better then the real ones.
Any suggestion on how to take into account a different market behaviour? The system was good in the first 300 trades and better in the last 300 trades, I'd have traded a system with the results of the first 300 trades and I'd get obviously very excited with a system showing the last 300 trades.
Smodato
Any suggestion on how to take into account a different market behaviour? The system was good in the first 300 trades and better in the last 300 trades, I'd have traded a system with the results of the first 300 trades and I'd get obviously very excited with a system showing the last 300 trades.
Smodato