[windsurfer]

Barry,

I'm on the road on my annual trip north for the summer. I'd like to look at your inputs, but won't be able for a week. I'd suggest you start by taking out all profit targets and stops. If you get valid results, add additional inputs 1 at a time. Start with a very large stop, then narrow it. You should also start with a longer period, such as 6 months, using @ES as your symbol.

Good luck!

[barryd]

Ok I mailed the developers and I got a response that cleared up the 2 issues:

1) I was being an idiot - a profit target of 1 in the EC is HUGE as it moves in 0.0001 increments. Something more realistic is 0.0015, apparently Clayburg uses something similar

2) The system requires an entry in the Stop_Loss field to work at all.

Windsurfer - thanks, yes I have found stop_loss values between 150 and 300 seem to work best.

I have been running optimizations for individual 1 week periods in the @ES to see how much the B1 S1 B2 S2 parameters change from week to week. The developers told me to use values between 8-20 as a guideline so I have been optimizing between those.

It seems they change quite a lot - it's hard to know how often the system should be optimized to keep it profitable, perhaps a few times a day? If so, that will need a really high-end CPU and tons of RAM to chomp through so many input variables...

I will continue to back test and may try running live in simulation mode tomorrow.

[daedalus]

I've been following clayburg for sometime and I think the self adaptive indicator could really be the next "big" thing in systems trading.

I have to give credit where its due, and its a fairly logical and elegant step to adress the underlying optimization issues all systems traders have.

He eludes in some of his videos that you could even re-optimizes every hour during the day... I wonder how that would fair.

It can't be the hard to code if we have the "beta" free version. Doesn't it just track the equity curve of 50 different settings in the indicator? If so, most of the parameters can be found in the free videos on his site and on TTM's site.

[10xray1]

I'm getting some weird anomalies as well...

For some reason, it looks as if the End Of Day exit is labeled different in some cases. I get "$ Trailing Stop", "C eod", and "S eod." Other times, "$ Trailing Stop" and "S Day Loss" is used for the Day Loss feature with a trailing $ amount.

Also, I find the trailing stop after Target 1 does not work at all. Perhaps I have some settings configured wrong. I haven't quite figured out what it's doing. It looks like its exiting the appropriate amount of contracts, however, they are exited right next to each other, disregarding the amount/% trailing.

Is anyone else having those types of issues? Perhaps a correct config for the trail?

Thanks guys!

[swansjr]

Originally Posted by 10xray1 »

I'm getting some weird anomalies as well...

For some reason, it looks as if the End Of Day exit is labeled different in some cases. I get "$ Trailing Stop", "C eod", and "S eod." Other times, "$ Trailing Stop" and "S Day Loss" is used for the Day Loss feature with a trailing $ amount.

Also, I find the trailing stop after Target 1 does not work at all. Perhaps I have some settings configured wrong. I haven't quite figured out what it's doing. It looks like its exiting the appropriate amount of contracts, however, they are exited right next to each other, disregarding the amount/% trailing.

Is anyone else having those types of issues? Perhaps a correct config for the trail?

Thanks guys!

I've learned there is a bug in the code in regards to the trailing stop. I guess that will be fixed with next next release.

As for the different labels they seem to be working on my charts. "C eod" is "Cover end-of-day" and "S eod" is "Sell end-of-day." Likewise for "S Day Loss" and "C Day Loss".

[rsi77]

I'm getting very impressive optimizaiton results on EC using 512 and 144 tick charts. Profit factors better than 2. I'm going to start forward testing and see how these hold up. I also know that TradeStation counts a touch on a limit price as filled, whereas in real life it usually aint so.

I'll probably try to code it up in Ninja and see how it fares there. In any case it's pretty interesting for a start of a system.

rsi77

[sevensa]

Originally Posted by rsi77 »

I'm getting very impressive optimizaiton results on EC using 512 and 144 tick charts. Profit factors better than 2. I'm going to start forward testing and see how these hold up. I also know that TradeStation counts a touch on a limit price as filled, whereas in real life it usually aint so.

I'll probably try to code it up in Ninja and see how it fares there. In any case it's pretty interesting for a start of a system.

rsi77

You can get very impressive optimization results with pretty much any system...

[swansjr]

Originally Posted by rsi77 »

I'm getting very impressive optimizaiton results on EC using 512 and 144 tick charts. Profit factors better than 2. I'm going to start forward testing and see how these hold up. I also know that TradeStation counts a touch on a limit price as filled, whereas in real life it usually aint so.

I'll probably try to code it up in Ninja and see how it fares there. In any case it's pretty interesting for a start of a system.

rsi77

Please post your live testing results in a few weeks. Would love to see them.