Mechanical Systems for Intraday

[Simon.B]

Hi Earik,

I remember there were lots of attempts, years ago, to create a decent Mechanical System for ES or similar for intraday. But the low daily range plus slippage and commissions made it prohibitive for intraday trading. Just curious if you or anyone else was able to create a decent intraday system since daily ranges are so much larger these days.

Thanks,
Simon

[sbank]

Funny that you bring this up Simon. I just started working on something yesterday for intraday. I still need to do some tweaking, but this is right out of the hive generator. (I normally use the hive on daily bars. But moving up to something like 2hr or 3hr bars is what I am investigating.)

Instrument: Corn (/ZC)
Time: 240m bars
Commission/Slippage per side: $12.50

300 trades under its belt is pretty good. Also corn is a very low margin commodity. (It is around 10x less than the /ES. So you can think about just adding a zero on the end for the equivalent /ES strategy.)

2023-10-17_21-43-41.png (88.17 KiB) Viewed 9405 times

2023-10-17_21-43-47.png (31.96 KiB) Viewed 9405 times

[Simon.B]

Hi sbank,

Thanks for your post. Numbers look decent already from what I can see. I was simply curious if anyone was able to come up with a profitable intra-day system for ES, like a 2-15 minute time frame, because that's where most of the effort was concentrated. The problem was a small average profit per trade, I think it was something like $40-45, and with slippage and commission - that wasn't enough. I was thinking that maybe a much larger daily range allowed for a higher average trade profit. I could be totally wrong in my thinking, so wanted to bring it up here.

Thanks,
Simon

[earik]

I've been able to come up with all sorts of things on daily bars, but anytime I get the itch to build an intraday system, I tend to get whacked. It's really an issue of the churn.

Back in my early 20's, when the big SP was still the thing to trade, I did manage to trade a semi-mechanical system with some success. (It was technically mechanical since there was no discretion, just too complicated to program in TradeStation, so I was the "computer" if you will, sitting there like a robot and phoning trades in). Anyway, the issue of churn was still a thing back then, but since the night session was zero volume and nobody could trade off their computer, you could really solve that problem by only trading the first and last hour of the day, when the volume was highest, then ignore any signals in the middle of the day when all the traders went to lunch. It's not so simple anymore with 24hour markets and all the bots running around, so the main challenge is to figure out some sort of filter that would tell you whether the markets were moving quickly or slowly, then only trade when you had the flow to make those average trades worthwhile.

Earik

PS - nice system, sbank. Corn, even!

[Simon.B]

Yeah, daily ranges got larger, but so did sort intra-day time frames. Not unusual to see 5-minute bar swings larger tnan daily ranges were 5-6 years ago.

Thanks Earik.

[lexa007]

NQ120mins built in September using Hives and GA

[lexa007]

NQ120Mins up to date. Terrible October