Hi
I am testing a VERY simple system of CMF<-0.12, exit after 5 days, only on SPY. For some reason it is running VERY slowly, about 2 days per second.
I have never seen this before. At this rate it will take 20 minutes to test 1 symbol over 10 years.
All my software is up to date, and I am on a very powerful computer.
Any ideas please?
Thanks and regards
Rod
PS: Update Granular Intraday Data is not checked, nor is Use Granular Limit
I am testing a VERY simple system of CMF<-0.12, exit after 5 days, only on SPY. For some reason it is running VERY slowly, about 2 days per second.
I have never seen this before. At this rate it will take 20 minutes to test 1 symbol over 10 years.
All my software is up to date, and I am on a very powerful computer.
Any ideas please?
Thanks and regards
Rod
PS: Update Granular Intraday Data is not checked, nor is Use Granular Limit
Rename
Uncheck ALL event providers.
Thanks. I think I found another problem also. I was selling 5 days after last exit, rather than last entry. That seemed to impact it.
Even with Event Providers, testing 1 symbol should be very fast with maybe some seconds of delay at the start.
1. Is that a block strategy?
2. What interval?
3. Data Provider?
1. Is that a block strategy?
2. What interval?
3. Data Provider?
Hi
It was a block strategy, daily, with Alphavantage. Alphavantage is working fine with other systems. I really think it was due to the rule "5 days from last exit" - I accidentally had that as a sell rule, which likely confused it. Thanks Cone.
It was a block strategy, daily, with Alphavantage. Alphavantage is working fine with other systems. I really think it was due to the rule "5 days from last exit" - I accidentally had that as a sell rule, which likely confused it. Thanks Cone.
Okay, I see it. We'll have to work on an optimization for that rule!
That's what happens when you count all the trading days from now to eternity (the exit date when you haven't exited) about 2500 times. We didn't imagine the rule would be used for an exit signal (I'll make change for that too), but the TradingDaysBetweenDates() function needed that optimization anyway.
Fixed for Build 93.
Fixed for Build 93.
Cheers:). My mistake, but glad you will stop someone else making a silly mistake like I did:)
Not only that, I'm going for broke -
In build 93 you'll be able to apply this rule to:
1. both entry and exit rules for any configuration, and,
2. the current symbol or All symbols in the backtest using a new parameter.
For example, as shown above:
Exit only if 5 Days have passed since Last Exit for All Symbols. That means this condition will only be true if no symbol in the backtest has exited in the last 5 days.
Or, if you change "All Symbols" to "Current Symbol", the "Days since" only applies to the last exit for the current symbol.
In build 93 you'll be able to apply this rule to:
1. both entry and exit rules for any configuration, and,
2. the current symbol or All symbols in the backtest using a new parameter.
For example, as shown above:
Exit only if 5 Days have passed since Last Exit for All Symbols. That means this condition will only be true if no symbol in the backtest has exited in the last 5 days.
Or, if you change "All Symbols" to "Current Symbol", the "Days since" only applies to the last exit for the current symbol.
Wow:)
Your Response
Post
Edit Post
Login is required