-> Risk Taking
-> Market making/arbitrage
In both cases the robot will likely not be overly simple. You will need to know a thing or two about programming.
Learning to trade is like getting a degree, learning to code is like getting a second one.
-> Risk Taking
This may very well work. I do not know of anyone that does this and is profitable over more than just a few lucky months, but does not mean it does not exist. I have heard of people that tried it, even people that told me they would let me know how it went they swore they cracked it, have not heard a word since then. Also, myfxbook is full of automated systems that end up blowing up.
If it is possible, you have to monitor the market all the time, be careful about events, look for evolving conditions.
I do not see how a fully autonomous robot - you can go to holidays and come back and find money in your account - could be programmed. Would be so complex...
Are there any profitable traders that are interested in this?
-> MM & Arb
There are plenty of hft and quant firms out there. Among other things they buy order flow from Robinhood and soon other "zero commission" brokers. Old market makers that did not want to move to algos just quit and did something new.
I see them mostly as mm or parasites. They can be so anoying. I used to market make on crypto 2 years ago, the parasite would just jump in front me all the time. Any order I place he'd hop 0.0001 higher/lower. Front running bots. It's not all front running bots of course. Soros & Jim Simons (I think) run statistical arbitrage funds. Computer farms to treat a ton of data help.
It works obviously.
For stat arb you will need: Learn the basics about the market (6 month course), an intermediate programming degree (1 year), a statistics phd (8 years). Good luck.
For market making you will need: Learn a bit more than the basics about the market (1 year, don't need the speculating course), a programming degree (1 year). I am being nice with the numbers I think.
> Fully automated robot you purchase
Pros and Cons
I don't even know how it is legal, it's pretty dangerous. They sell you some fairy dust and also you get advised to "hold on" "do not be emotional with fear just hold" "be strong during drawdowns they are normal" it's actually risible.
New traders that might be a little gullible, a little greedy too will end up with some fairy dust that is guarenteed to bring their account to zero, they do not understand the markets, AND to top that, they hear that they should not panic about losses and just persevere because drawdowns are normal and they are just experiencing fear, this is what will make them lose, and this is what the robot is here to fix. "Just trust".
What you have here is a recipe for disaster.
> Robot that executes for you/semi automatic
I see a few possible uses:
- You want to buy alot of shares of a small company long term, you set a bot to purchase some each time the stock has a strong down momentum with (fills you without moving the market)
- You want to sleep / take a few days off, you run an algo that will move your stop loss each time the price goes up, would not be too hard to say "use the smallest of 20 0.236 ATR something else, if consolidates move closer, if price == .... then do this, ifelse the price surges then move my SL closer" etc.
- You could have a feel for the market and be very , just set a bot to buy an bull break with a couple of conditions, or a pullback of a certain % but still above a certain level/MA followed by a . Sets and trails the stop loss too until you intervene because conditions have changed and you do not want to stay in or whatever reason.
- Big one: monitors correlations for you. I'd actually recommend that. A robot to warn you of your global portfolio correlation, maximum risk, and makes some recommendation (shows what are you most correlated positions making a problem, and lists stocks/currencies/futures that are uncorrelated or negatively correlated)
> Robot that alerts you
Plenty of people use that. Makes perfect sense. Something just warns you when the price does something particular or whatever. I myself mostly use trading view price alerts. I just set an alert on Bitcoin MA50 & MA200 makign a death cross, should fire up by Halloween.
I would not go overboard on this for "advanced" things, because it could get complicated, and a quick glance is even better anyway, you want to focus on what pops out "obviously", you want to see what other people are seeing on the chart.
You may want to be long on certain currencies, the bot could tell you when the price is starting a pullback...
> Indicator (sold)
There are people that dish out indicators and show people "See? Here, here, and here, my indicator fired a signal right before Bitcoin went up", and people actually get excited by it and want to purchase... And guess what? It works extremely well with crypto enthusiasts that invested in imaginary magic internet beans. Who would have thought they were gullible and dreamt of easy money fall out of the sky? The coders use the right combination of indicators, for example EMAs and , to generate great signals IN THE PAST.
I made such an indicator myself just to show it's bollocks. I think some people still used it seriously.
Black box "predict the future" nonsense ==> Obvious scam.
So watch out for those. If you are interested in something, make sure you understand WHAT IT DOES.
Example: If the indicator provides you with order book info and idk things like this that you don't have with your broker and charting service, it obviously has value...
On the other hand, it is possible to code for something both profitable and accessible in term of complexity. Let's say a weekly chart with a couple of indicators from different timeframes (use 1D to identify trend and 4h to set entries and exits for exemple) can provide a reasonable income (lag is the biggest issue in this game imo). Indicators should be adapted to a unique market (length, lag, alpha, slow/fast limits, etc.) and not to be expected being profitable on a different market. These parameters would have to be adapted regarding the state of the market (I'm talking chronological, and this works very well in crypto markets, like for exemple market behavior is very similar throughout impulse waves of one cycle) but also frequently no matter what as the market evolves (bitcoin cycles getting longer and longer).
In other word any profitable strategy should not to be expected working indefinitely (has to be adapted frequently - over the course of months) neither on other indices (a bitcoin profitable strategy would not work on ethereum without - at least - reviewing the parameters). Plus, this is a game of probability and you have to accept fake signals, a strategy >60% profitable trades is good and positions size should be weighted considering this probability parameter (on top of reasonable stop-losses, I don't use take profits).
To end I would add that, regarding crypto-assets, prices follow extremely sinusoidal patterns that are way to correlated to the initial strength/velocity of moves (indicators compiled on top of such prices as well), my explanation to this is that this market is mostly traded off automated-strategies. This is a hard concept to both illustrate and explain.
This is just my point of view with the limited and very amateur knowledge I do have over this topic (automated trading strategies).