Index futures automation

[NoahA]

Its amazing to see this progress, especially after your initial great runs at manual trading and then subsequent drawdowns. That equity curve looks pretty damn nice, an the fact that you have equal number of short and longs I guess means this algo can work in any market condition. (but perhaps not low volatility??)

Anyway, thanks for sharing the progress.

 

[Overnight]

...
View attachment 297459...

Well, I think it is awesome, but here's the one niggly bit that would give me pause. You say it would take you 50K for it to run well. But see that dip in the profit curve at trade #155? God forbid you implement the algo right at that top of the curve before it dips down like that.

If it does, your strat is stopped dead in it's tracks.

So you will need probably a few more 50K tranches as a backup to re-initialize it were that to happen.

Only yourself and Rickshawman have shown autostrats that have posted good returns thus far. Please make it work. We need some good news around here, man.

 

[hilmy83]

Going through a tough few days where prices zig zagging without clear direction. algo taking false breakouts and micro commissions are eating through the profits. Add the slippage to that and it seems like theere is equity ceiling. Unless there is clear directional move with high volatility, this algo is just churning out commissions.

Quantanalyzer doesnt really give a good view of what takes place intraday. it's more summary of end of day performance.

but when i pull up mt5 report, it's clear how choppy this has been

 

[hilmy83]

Going through extended drawdown. Main culprit is MES. I literally haven't made a dime from the MESZ22 contract. The best performer is M2K,MNQ followed by MYM,FDXS.
It's really interesting how the same strategy can work across other indeces but MES just coudn't have enough volatility for it to close in profit.

If Monday performs normally, it will be a decent month, otherwise I see a b/e or even slight loss. I don't know if MES is going throug something or if this is just a trait of ths market. It is the biggest of the 5, so maybe it's just more mean reverting.

 

[hilmy83]

Today is one of those "better learn now then later" kind of day.

First, I was completely fucking clueless at the fact that euro follows DST and they swithed on different day. Today was the switch. So coming into today's session, my algo on FDXS basically took breakout range off pre market noise. Which meant it traded off literal blip range and my algo took a fucking massive size based on set % risk. It wouldn't have been bad if it was a straight shot to TP, but it reversed a couple of times, and of course size and commission added up which made this into decent a shit day.

Then after that happened, I decided today is the best fucking time to add new markets into the mix. I was studying the MCL (micro oil) for a while and feel like it's the right candidate that fits my breakout style algo. Well I thought, why not add MGC (micro gold) as well. Didn't even bother to look at the average range for MGC or price action for it. Just assume, they're commodities they should behave similarly. The MGC trade turned out to be shit because it had such narrow hange. WHen TP hit, I lost from slippage/commission. Now that I look back at MGC, it's definitely not a 'breakout' type of market like MCL is.

Anyway, what fucked up day. If it wasn't for DST error and MGC (unplanned), it would've been a decent day to close out a decent month. But closing out the month around 3-4% gain

Performance since 9/1 (start of final algo version)

I'm pretty much resetting my risk back to baseline. It was supposed to be a recovery period, but since i'm so fucking annoyed at the mistakes, I'm just going to reset. Going forward will trade MES,MNQ,MYM,M2K, FDXS, MCL

 

[hilmy83]

Oops got a warning from AMP. I adjusted the algo to trade on two separate times (i'm still experimenting possible multiple sessions where breakouts are ideal). Algo has internal check for current pnl before taking trades, so that it doesnt' continue taking trades after hitting TP. I forgot about that check when I loaded it on two separate session today.

Basically the algo had wash trades for MCL since it kept cancelling orders after checking it had already profited. Did that multiple times, good thing i was on and realize the error early and cancelled the Algo. Just need to tweak the max trades code.

 

[Overnight]

Well I thought, why not add MGC (micro gold) as well. Didn't even bother to look at the average range for MGC or price action for it. Just assume, they're commodities they should behave similarly. The MGC trade turned out to be shit because it had such narrow hange. WHen TP hit, I lost from slippage/commission. Now that I look back at MGC, it's definitely not a 'breakout' type of market like MCL is...

They call NG the widowmaker. They call GC the insane asylum. Just ask @themickey GC is it's own animal and cannot be tamed. Keep it out of your portfolio and you live longer.

 

[hilmy83]

Decided to cut fdxs again, for good. A month ago it was decent, but seeing low liquidity price gaps, I dont' want to bother trying to scalp that and get screwed with slippage.

So now it's it's just the 4 indeces and MCL, running on 2 separate session range each day, on 2 separate algos so they act independently.

I'm pretty confident on the time slots. I think the PF should revert around 1.25-1.3. I think with eminis, it would be 1.4-1.5 due to improved commission.

I'm already >$1000 in commissions so far with the micros just 2 months in. Almost 30% of gross profit.

 

[hilmy83]

Did some 1000 cycle monte carlo sim based on current trade numbers

 

[lucysparabola]

Are you going to change platforms after all? Still getting an issue in QT sorted out, hoping for over the weekend. beyond that, it has been a world of difference for me.

all the best mate. PMs are always open if you need anything.