Is there anyone here who has a consistently profitable automated trading system?

[murray t turtle]

I ask because I heard an interview with Linda Raschke where she said she didn't know any such traders. I thought that was pretty significant as she is a very successful trader herself.

I have tried quite hard over the years to automate some winning trading rules, but with no success in the long term.

I am not asking anyone to reveal any secrets, I'm just interested to know if it is possible.

%%
YES;
but I use plenty of discretion with automation[NOT in the superstrict sense she may mean].
I'll give you part of it /
includes but not limited to Linda B Raschke quote ''holy grail = buy @ 50dma''
OF course ''not profit every trade '' LOL\even market makers dont do that.
I also allow 1% slippage on my mostly market order exits
[EDIT= i consider the discretion part much more important than the auto or system part; so just because you like the auto or system part, keep looking for more than that,especially since the 2023year is half over]

 

[Axon]

I have tried quite hard over the years to automate some winning trading rules, but with no success in the long term.

I am not asking anyone to reveal any secrets, I'm just interested to know if it is possible.

I used to have a couple of crypto bots that worked really well back in 2017/2018 until the edge got arbitraged away. I ran a 0.6 btc account up to about 50 bitcoins at the peak before it was finally over.

One bot ran on Binance specifically since the volume there was so huge. The edge was there were a few crypto coins that traded against bitcoin with really low prices. BTT in particular had a price that fluctuated between 0.00000007 and 0.00000008 BTC or so for quite a few weeks. It also traded against USDT where the price was much higher and, key to the strategy, the difference between the bid and the ask was a much lower percentage. What the bot did was it would camp out on the 0.00000007 bid on the BTT/BTC pair and about every 24 hours or so it would get to the front of the line and make a buy of about half a bitcoin worth. It would then immediately sell the BTT for USDT on the BTT/USDT pair. The trick was the equivalent USDT price for BTT was half way between the 0.0000007 and 0.00000008 BTT/BTC price. So I would buy for the lower price in the BTT/BTC market and immediately sell at 0.000000075 for USDT. Then the bot would take the USDT and immediately trade it for BTC and then finally camp back out on the BTT/BTC bid waiting for the next buy in another 24 hours. I made about 5% per round trip for a few weeks completely automated.

Another bot was a little more generic. I had a constant feed for about a dozen crypto exchanges running 24/7 and the bot would just check certain high volume crypto pairs for huge sudden drops that indicated (in theory) that a single whale was dumping a bunch of coins for some unknown reason but wasn't indicative of the actual market for the specific coin falling out (so just an instant high volume drop basically a fat finger). The bot was watching so many pairs on so many exchanges (something like 27000 unique pairs at the peak) that every day it would get several trades. Volatility and volume in crypto was so high in the 2018 bear market that a typical buy would be 80%-90% below the market price so my reward to risk was very high, much more than offsetting the occasional loss. Sometimes I'd get something for 95%-99% down. It was nuts and went on for a while.

At any rate this all worked for a while and I made about 100x the bitcoins I started with in a completely automated fashion. Then slowly it started drying up until finally the trades just stopped coming in and the ones that did lost more than they gained so I finally shut it all down taking all the winnings out of crypto and putting it in stocks with the rest of my accounts. It was an era I remember fondly and a lot of fun writing the bots, tweaking them, etc. I'd say tinkering with bots is probably more involved than just discretionary trading. I still mess around with crypto but not with automation. I use a trend following system these days that gives a signal at most once a day so I can just as easily do it by hand. Funny thing is I think one of my very first elitetrader posts was a step by step on building a crypto bot.

I used to use a program called Blockfolio to keep track of the trajectory of my account. Pic related is a random screenshot of it I took around the heyday of the bot that kind of shows the progress of the account for that period of time. Fun times.

 

[mason macgregorson]

I ask because I heard an interview with Linda Raschke where she said she didn't kno w any such traders. I thought that was pretty significant as she is a very successful trader herself.

I have tried quite hard over the years to automate some winning trading rules, but with no success in the long term.

I am not asking anyone to reveal any secrets, I'm just interested to know if it is possible.

I trade robots from targettsrading.com with the aftermarket settings from global trade titans.com. My four bots average over $1500 per day at three contracts per trade the last nine months. All forward walk real time trading . I have another set of bots again from TTP that don't trade as often but do equally well. This is my performance from yesterday and today. Admittedly these are better than average days but you get the point.

 

[murray t turtle]

Enough to pay the bills and live but personally I live below my means by a significant margin, that helps me get through significant and often long drawdowns. But I have plenty of losing months so I don't fit your criteria.

%%
LOL that'$ ok;
not every month profitable leaves a lot+ LOTS of room for profit
A budget[spending, investing, trading] plan can help also.

 

[Businessman]

I would say every month, with maybe the odd exception.

That would be a Sharpe ratio of 3.0+

Not many retail traders will be able to do that over the long run.

Somewhere between 1.0 and 2.0 should be possible for retail traders, but that probably doesn't fit your definition of 'consistently'. Going to have a lot more losing months than just the odd exception.

 

[murray t turtle]

That would be a Sharpe ratio of 3.0+

Not many retail traders will be able to do that over the long run.

Somewhere between 1.0 and 2.0 should be possible for retail traders, but that probably doesn't fit your definition of 'consistently'. Going to have a lot more losing months than just the odd exception.

%%
SPY benchmark , in no way does that ''every month'' long run;
80% cant beat SPY benchmark, average 12%depending on start date. But >18% YTD .
Not many investors would be able to do that [every month'';
even the many millionaires that bought eVery month for for 30 years.
So really have to look @ much more than ''every month''

 

[Businessman]

Rentec Medallion has something like 88% winning months over the long run, and average about 6% return per month.

 

[d08]

Things like for example classic chart patterns such as flags, wedges, H&S, or volume behaviour patterns within those chart patterns are very hard to code. Many good programmers tried, but so far the trained human eye is better. Obviously, it comes with subjectivity.

Mechanical systems are good for mean reversion, or bar by bar analysis, and stuff like that.

All this stuff is automated and the patterns can be detected well. I did this nearly two decades ago with a slow programming language and with poor code. Nearly 20 years later, you can be certain these patterns are quantified. The methods are simply not made public.

 

[PPC]

All this stuff is automated and the patterns can be detected well. I did this nearly two decades ago with a slow programming language and with poor code. Nearly 20 years later, you can be certain these patterns are quantified. The methods are simply not made public.

It is such a privilege to have here on this forum the some of the “worlds’ best” programmers and always reassuring us that it is possible to do what nobody else can.

Obviously, their methods are simply not made public, only talking about it is public. We understand, it's all secret.

PS: The real issue I have with your posts is that it gives newbies false hopes and then they can waste many months running to the West looking for a sunrise.

 

[d08]

It is such a privilege to have here on this forum the some of the “worlds’ best” programmers and always reassuring us that it is possible to do what nobody else can.

Obviously, their methods are simply not made public, only talking about it is public. We understand, it's all secret.

PS: The real issue I have with your posts is that it gives newbies false hopes and then they can waste many months running to the West looking for a sunrise.

We are not looking to solve some century old mathematical problems. Chart patterns are not insanely difficult, it hardly requires amazing programmers.