Lost Sock Evolution - Automated

This is definitely worth a shot!

If you go through the TE's, one thing to appreciate is that they typically never absolutely Nut a chart for entries nor exits...meaning?

There is some flex on either side.

This may help out with a wider spread to help out with the extra gains.

Question is...

Are the spreads 25% higher than you can get with other brokers?

Because, if they are, you are pretty much at a wash with extra cash back to tighter spreads...kind of pound of feathers to pound of bricks, idea.

Yes on most pairs they are, so you are probably right. My thinking is that positions will close for $1 anyway, so does it really matter if they have to travel a few extra pips to get there? So if you could manage to close out the same positions almost just as quickly, you would be better off. Of course if tight spreads means that we can close out 4 positions for every 3 that the wide spread account closes out then we are back to being flat like you say.

That's why it'd be worth giving a shot...doing the calculations manually, with a flex in spread and crazy shit at roll over...really hard to tell which way things would go.

FINALLY!!

Some useable fucking DATA on these!!



Revolution, Evolution 1.1 and Evolution NN were all switched over to Summ Start - meaning...they add incrementally to the previous position.

The idea behind doing this was to see which of the logic would prevail over EXTREME conditions.

Now...it looks like we have our winner - Revolution!

Revolution has a secondary code in it to be a little more aggressive in counter triggering. This has added in additional positions, but has also allowed the flex in market to keep the DD well under control.

Where Revolution is running on 19 pairs the Evolutions on only 7...pretty easy to see and say that $155 in the shitter is a lot less than the $2064 and $1668 on the Evolutions.

Here's the thing...

We couldn't go flat out with all of the same pairs on the Regular Version of Revolution as we maxed out the 200 trade threshold of the most brokers in less than 2 weeks...we wouldn't want to go with only the same magic 7 pairs on the NN version, because the profit, though steady, might make picking bottles in ditches more appealing...

So...

To sum up...

LSE, killer EA. I've bumped lots on this up to .03 and $3 profit...figured we were close enough to that next 100% that it'd be worth a shot.
Revolution NN...can run very well with either Fixed Start or Summ Start and continue to churn the cash...risk is the only consideration under summ start and adding in additional lots for Fixed with a wider profit might be the way to go. With the Filter Added in the current version, and on a smaller risk account, this motherfucker is GOLD!

From starting with this triggering idea, some 10 months ago...finally at a place to start improving?

Still the daily filter to drop into Evolution and get released...got this started a couple weeks ago, hopefully polish it off over the weekend.

And then...testing through Christmas!

w00t!

Alright...so I might have cheated this a little...but seriously...


60 days at .01 with a $1 profit to double an account. Doubled the lot size and profit and it took another 30 days to get to...well, okay, just short of 200%

And really, I only cheated by like 2 days...



Lot size .03 and profit $3...

Will we see the next 100% add in 15 days?

Made out like a fucking BANDIT only the dragon last night.

It wasn't cheap...but was FUCKING AWESOME!

Mad Price Rise into London Open, leaving a HUGE retail gap. Shorted and with the help of a little logic specific for this event...closed close to 300 pips!

Sure...there was a little pant shitting going on, but will be sporting some silkies by end of the weekend with the profits from this motherfucker alone!!

All While watching this...I had some thoughts about Revolution and it's Trade Logic...

TBC!

Loon was another paydirt vehicle today, though, sadly, I missed exiting one position to allow this to rise...to hit profit of the GMO60 prior to me being able to do this and consequently exited for 80 pips...Net on Loon this week...bout a buck and a half...

Great fucking Score.

There is definitely something to the use of the GridMeOut logic that could really help tend...

Not an entire fantastic close on the week...however...


On my micro - Fuck Around - Account...I was playing with another GMO idea.

Fairly Aggressive and you have to maybe add this to a longer than M1 chart and maybe only look at enough pairs that your account size and margin can handle...

Meh...

The idea is solid and may be married to one of my others for trend to counter-trend trading...again...longer time frame...less pairs...more funds...will be testing.

Evolution and Revolution - No Nedge...

The idea on the trading to a Homeostatic Market Median is still ROCK fucking SOLID!

BUT!!

When writing in the No-Nedge Logic, we've seen the failures of this, as applied to Evolution...a little less on Revolution.

When originally coding the traders, they were set to directionally trade...if you look at the success of Evolution - 200% gains in 90 days...easy to see we are doing something right...right?

But...

When dropping in the No-Nedge, as applied to Revolution requires a separate out of trend logic to keep it in the good books. I'm not a fan of this, even though it has proven to be fairly successful..in wicked ranges on goofy pairs, this motherfucker could spell ultimate demise for any account.

Both Evolution and Revolution NN versions are still set to only initiate a trigger when they are in proper cycle trend and trade towards market median. The problem with this is...this creates a really narrow trading range and has actually caused the EA to start tripping over it's own nuts causing catastrophe on Evolution NN (with the incremental add)...

So...

The logic works in trend...falters out of trend and can get hung by the balls in extreme situations.

Ergo...the bug in this idea is that since we are trading towards a target, we know we can hit...why wait for true multiple sequencing on trend to work backward?

I think I can take ideas from both of the NN's...drop it into a hybrid...throttle back the sensitivity...and remove the Revolution out of trend trigger.

I know...I am supposed to try keep focus and have a couple of EAs that are way past promised dates...but I think this one will jump line along with at least one basic idea for the GMO concept...

Should have some time this weekend and already stocked up with the weekend provisional suds.

Time to get fucking HAMMERED and Tap into Creativity!

Enjoy your weekend!!

\m/

GMO - Trend60...

So...the idea behind this, which may be obvious from the pic...

Is that instead of using a price grid, it uses time and is specific to criteria...

Looking at it..rather than buying as price slides against it...it triggers into the primary and looks to shorten the range for profit.

This is one of the modified version trades it made...I was changing code with almost every trigger for a while, just to get it to do what I wanted it to do...

Overall...it does have potential, but again, viewing this on an M1 and especially if you fire this motherfucker up in shallow waters, could get you into some serious troubles.

Anyways...this is just a sneak a peak at the idea.

Not sure what I was looking at or working on when I wrote Evolution NN...some pretty fucking wild ideas written into that motherfucker...

Primary version of Hybrid NN written and testing in Strat Tester to make sure triggering is correct.

From all indicators...which isn't live testing...I think with the Hyrbrid, we can improve the win rate and the average trade value.

This should help make up for fewer trades and less overall Draw Down.

If all goes well, will have this fucker posted tonight!

You've got the entire weekend to Backtest your heart out and NOT POST BACKTEST RESULTS IN MY THREAD!

Hybrid NN V1 -> HERE

I think you will really like this one...



Alright...I'm half in the bag...but I fuck'n love you man...

The perfect shot of a Homeostatic Median Trading EA...

Yeah, it's from the strat tester...but it's not 70 fucking pages of different settings...

It's a sample of the actual logic as was designed.

Homeostatic Median Trading!!


Very Kewl!

One last thing to mention about Hybrid NN...I have coded in the market median into a longer look and because of the less restrictive trade ideology of the new trader, it can actually be applied to an M5 chart...due to logic restrictions, M5 is probably the longest chart it will trade on.

Might not be that beneficial, but it's a start...

Hi Mr. Magix,
is your EA restricted for Live trading? I tested your RevNN.1.2 at micro acc where is min. lot 0.10 and suffix (e.g. EURUSD.)
Unfortunately none trade opened for over a week - and no errors in expert.
Thanks for answer.

Updates

The pictures below are self explanatory. Summary of the Revolution and Evolution trading
- I changed all to sum start
- I do daily check of ADR movements to adjust the step
- All on M1 (The ADR adjustments will help limit the trades)

Evolution 1.1 / 35 days / 18%
- 6 pairs
- Step increased to 200 to give room for stacking


Revolution NN.1 / 22 days / 8.2% / Live and still in red
- Reduced to 16 pairs
- Step increased to 150


Revolution NN.2
- Same as NN.1


I just setup Hybrid NN on my VPS
- Same settings as NN.1
- 14 pairs (selected based on WTR:ATR ratio)


For all pairs the WTR:ATR is between 1.5 - 2.7
- With this information, I know the max range for each pair
- With the introduction of median filter, I know the range from which the EA will trade
- By applying the ratio of the range, I can better control how many trades it will place
- If I see that the ratio is increasing, I increase the step accordingly to reduce the max number of trades possible. At least I am only trading from either side of the median (1/2 the range).

To use this in a live environment, you should understand what it is doing or it will blow up the account quickly. Sometimes, I take it off pairs that have cleaned the table to allow room for other pairs to trade. Especially when number of total trades has exceeded 100.

BTW: I watched the Hybrid NN trade offline today. It's cool. Great job Magix. Please which time frame median did you implement for both Rev NN and Hybrid NN?

Timeframe=M1.

If on anything other than that, you most likely will never see a trade.

There is no restrictions and no expiry on anything released in this thread.

If you are still having issues, post your log files and I can check shit out from this side.

2 New Trade Explorers added, taken away the 2 busted ones.


Going to start these off with the Magic 7 pairs, M1, Default Settings.

Difference?

One has an additional trigger built into the Hybrid EA.

Though I am not totally convinced this is required, it may help in the cost averaging and time in trade.

Exit logic is the same...so, we should have a true comparison for the value of a supplemental trigger instead of just the hybrid NN logic.

I will be away for most of next week...doing a little traveling, some out of town business and commitments.

Will be checking in, but limited in any sort of analysis...

Have a Great Week!

Good Luck in your trading and testing!

Magix many thanks for the new Hybrid NN version.

I was reading and catching up with the developments, am I correct in saying that Hybrid NN is an amalgamation of the logic Evolution and Revolution along with the No Nedge code added?

I'll be setting up Rev NN 1.2 TE this week, and I'll load up Hybrid NN as soon as I can. I've got some tests finishing up soon so I'll have some more space on my VPS, and I can give these a spin.

The TF for Rev filtering is off the M5.

For Hybrid, it's off of the H1.

Both view the Homeostatic median range off the weekly.

Issue is, due to signal triggering, Revolution will not trade in anything other than an M1...Hybrid will trade off of the M5, but triggering logic on the initial release may leave some large holes in spacing...

I do want a longer version of the NN logic, but think that this will still ultimately need be applied to the faster frames. Reason?

We can extend triggering, like with your analysis for Market Range and I can always extend the exit logic to different ideas of profitability and dynamic ranges...

Applied to the longer frames, it's a lot more difficult to apply pattern recognition.