Machine Learning in ProOrder ProRealTime
- This topic has 454 replies, 32 voices, and was last updated 2 years ago by
.
-
Grahal: So if I set increment = 2 then in any one bar, ValueX or Value Y should not change by more than ‘2’ (up or down)?
My observations are that often ValueX and / or ValueY does change by multiple x incremental value in any one bar.
Change by multiple x incremental value occurs when increments have not reached their maximum value … anyone else get the same?
PS to @Bard
I take it you found Renko Type … it is / was in the code of whatever System we were talking about!?
I can’t keep up with all these new changes … I need more supplements for the brain … I’ll try a Vinpocetine today!! 🙂
1 user thanked author for this post.
When everyone was having a s&*% time trading
Wow yeah … took me a while, but I got same settings as you … we could ALL have been millionaires now and part of the ‘Greek Set’ drinking beer with Vonasi on his Yacht!!!!!
EDIT / PS
In case you are wondering wtf we are on about … check out attached!!!
This one has your name on it @Nonetheless so we’d all have had to chip in and buy you a Yacht to say thanks!! 🙂
1 user thanked author for this post.
Following up to the ML conversion I made of a great Renko Trading System (see previous p11) – here are new random date ranges, new time frames and this time the Spread added, (only accidentally omitted from other tests due to creating a new system instead of duplicating an existing that had my 3.8 spread). The previous examples on p11 were without spread, but it makes no difference when spread is added. Tick by Tick testing remains very profitable whether Coffee Arabica (KC) is trending or ranging.
Note, this newer version of my ML system (see below for itf file) swapped the ValueX boxsize for optimising the Stop Loss and fixing the boxsize at 100 and the Target Profit of 500.
Here are the 1 and 4 hourly results. Again a more realistic but still aggressive money management algo is applied of 10 x capital/close. Low drawdowns remain.Could anyone apply a Vonasi Robustness Test to these results? Thanks.
1 user thanked author for this post.
For me it’s too beautiful, there’s a bug.
1 user thanked author for this post.
This one has your name on it @nonetheless so we’d all have had to chip in and buy you a Yacht to say thanks!!
you guys have taken this and run with it, which is great. I think I’m responsible for about 1% of its DNA by now. Maybe less.
Been through some weird ups and downs this one, but looks like it might fly after all. Thanks to all contributors!
1 user thanked author for this post.
For me it’s too beautiful, there’s a bug.
That’s what I think, hence my “£55 million on coffee, that’s got to be a mistake surely?”
I ramped up the Money Mgt algo to a ridiculous 30x to push it and see if it would blow up, as I’ve got a lot of past performance experience on other systems with this money algo and know it’s limits. When I got past 10 I was surprised… and then 20… and finally went “full hog” to 30, I couldn’t believe it. As you say it looks too good, so then I left the same ML settings that I’d used for the original 5x capital system and then applied these exact same settings again but with different instruments like the Dow, Brent Crude, and then finally Coffee Arabica… then I thought, okay lets “break it” with a different TF, so I tried it on the 4 hr, 1 hr, even going down to the 10 mins, but I couldn’t. So then I changed the dates and there were solid results there too… A little baffled tbh, hence the request for a Vonasi Robustness Test. My task to figure out today, hopefully.
Find exclusive trading pro-tools on 
