Index futures automation

[hilmy83]

I reran the backtest again of my algo on MT5. It's a continous contract backtest. The timig is a bit off (MT5 doesn't adjust daylight savings time automatically). It's also non-compounded.

It's definitely not a super accurate backtest , but just wanted to see the results nonetheless.

2020-present

RED - algo coudln't finish the day in the green/breakeven.
ORANGE- incorrect entry due to daylight saving time error (dismiss data)
BLUE - Live trading start point.

So there was one time the algo didn't get to profit for the day. It triggered the max loss I set for the algo, and took about 4 months to recover to new equity high. The actual DD is 25%.

This is what I consider my risk of ruin scenario. If it happens say once every 1-1.5 years, I'm good. If it happens back to back...I'm fucked.

I designed the system where it will continue trading the session up to the max loss sequence. If I start messing around the sequence to have "smaller" losses, it starts to throw the whole thing off.

Reduced "risk"

More reduced risk

The more I continue trading this, the more anxious I get. I haven't experienced that 'black swan' event yet since the live forward run.

 

[traider]

I think you just need to reduce risk

 

[hilmy83]

Algo is fucked again. Somehow it left open an order and I got stuck with it. My wife notified me while I was at work it was still trading! So had to log in and closed it manually.

Again, it somehow found its way on the right side of the market when it messes up!

Someone mentioned that I may have accidentally discovered a code bug where it actually able to profit when it encounters this issue. I WISH, but this algo is royally fucked. I need to do some solo debugging. THe dev is has been MIA..

 

[NoahA]

Again, it somehow found its way on the right side of the market when it messes up!

Its a sign from the trading gods! Maybe you should forget the consistent daily green closes for the account, and focus more on riding a wining trade. I mean if you finish +$20, +20, +20, each day, is this not worse than $100, -$50, +$100, etc.... overall?

But don't listen to me, I don't know anything... just enjoying your journal.

 

[hilmy83]

I have to undo the recent changes I added, cause this definitely is not what it's supposed to do!

Its a sign from the trading gods! Maybe you should forget the consistent daily green closes for the account, and focus more on riding a wining trade. I mean if you finish +$20, +20, +20, each day, is this not worse than $100, -$50, +$100, etc.... overall?

But don't listen to me, I don't know anything... just enjoying your journal.

You can invest for growth or you can invest for yield. I'm a yield's guy...it's in my nature.

 

[NoahA]

You can invest for growth or you can invest for yield. I'm a yield's guy...it's in my nature.

Totally understand, and I also love the smooth equity curve. But I just figured that if the algo is doing the trading and I wasn't personally having to place the trades, a slightly less smooth curve but higher yielding profit would be nicer to look at at the end of the week. Yield vs. growth is kind of a crutch to make the mind feel better, but at the end of the week and month, total profit is perhaps the best metric.

(Ps. I can almost guess your entry if I extend out the line that connects the exit and entry price... )

 

[hilmy83]

I feel like the stop and reverse process is overcompicating things. If I had just treat each order with its own s/l, then place a new stop order for re-entry (reverse). This would treat each order as individuals rather than this mesh of stop and reverses and partials that can can confusing..

 

[NoahA]

I feel like the stop and reverse process is overcompicating things. If I had just treat each order with its own s/l, then place a new stop order for re-entry (reverse). This would treat each order as individuals rather than this mesh of stop and reverses and partials that can can confusing..

The funny thing is that whatever you're doing is simply incredible! Nobody on here has shown results this consistent. So even though you think the algo is crap, and perhaps it can be tweaked to eliminate these errors, you still have a system that is miles ahead from results published here. (unless of course the guys doing better just don't talk)

 

[hilmy83]

The funny thing is that whatever you're doing is simply incredible! Nobody on here has shown results this consistent. So even though you think the algo is crap, and perhaps it can be tweaked to eliminate these errors, you still have a system that is miles ahead from results published here. (unless of course the guys doing better just don't talk)

Don't jinx me man. Right now I have a strategy with a broken algo. Imagine having a decent strategy but too clueless to make it work automatically

 

[rb7]

Don't jinx me man. Right now I have a strategy with a broken algo. Imagine having a decent strategy but too clueless to make it work automatically

If I may, your (trading) algo is not broken, your code is. No code is perfect. That bug will not be your last one.