Trading with AI: Now possible for everyone in Forex & Crypto (not ChatGPT)
Welcome to our article on how AI can already today bring your trading to the next level. This is not another ChatGPT set-up in which we let AI built us trading bot, but a very real application of AI to your trading.
You as a trader have done a couple of hundred of trades, maybe a few thousand if you are trading for a while already. Even if you analyze all your trades, that’s still a very small sample size. What if an intelligent tool would analyze hundreds of thousands or millions of trades, learns from it, and then applies this huge amount of knowledge exactly to your trades, finds your areas of improvement and is able to filter which trades of you are likely to be profitable, and which ones not?
Thinking one step further, what if that intelligence actually tells you exactly this, or even filters and calibrates your trades in an optimized way, and continues to improve this as it learns more? Well, I can tell you a lot if this is already possible for traders today, but lets go step by step.
Outline:
As a first step, let’s deep dive a little bit into the theory behind it, you’ll be surprised what is all measurable and how an intelligent tool can use those information. Second, I will give you a few real examples of how this intelligence is applied for a trader, and last but not least, I show you how you as a retail trader can actually already use this to your favor and how the whole process works.
Theory set-up & data points for the AI:
Alright, let’s get started with tiny bit of theory, but it will get very real in a minute when you see what I am going to show you.
There is a huge amount of trading relevant data an intelligent tool can use to learn from it, much more than you might initially have in mind. In an initial step, we have a look at three layers:
First: All your actions when trading and data points we can derive from this. Here are a few examples: Trade length, number of symbols during the day, or traded simulatenously, order type entry, way of exit, adjustment of Take profit or Stop Loss during the trade, position size, and much more. Just this list already show quite a lot.
In the second layer, those action can be merged with trigger points. There are some more examples in the graphic above, such as large loss trade before entry or a big win trade, your drawdown for the day, month, a reverted trade, so a trade that was in profit but turned to loss beforehand, a winning or losing streak. Again, lots of aspects, and if you imagine merging those now with all the points from the first layer quite a lot of combinations.
Now the third layer: What were the market environments, prior, during, or after each individual trade? Trade followed the short term, medium term, or long term trend, or against the trend, high or low volatility, an overbought or oversold market, which geographic trading session?
As you can see, just with these listed actions, combined triggers, and combined with the market environments, we will have millions if not billions of combinations. In those combinations, we will be able to find patterns, behavioral patterns, which show a strong correlation with the outcome of the trade. The more trading data from more traders we feed into the system, the better it can find those patterns and learn, and then apply those patterns to YOUR trading, see which apply and which have the strongest performance effect.
In fact, we are already running all of this, and we already found a lot of significant patterns although we are just starting. Seeing those patterns and outcomes, we can actually oftentimes make sense of those from our own trading experience and I’ll give you a few examples in the minute of what we found and how this can be applied to your trading. One of the major strengths of this approach is that it is not purely rule-based such as many trading systems, thereby not reproducible and attackable. BUT! If we add this behavioral aspect of your trading into the equation, the human aspect, it becomes completely customized to you, yes learned from many traders, but then targeted to your exact behavior, and thereby it’s not possible to copy. It may not be very obvious to everyone, but especially traders which had successfully running bots probably understand that this is one major upside!
Real examples of AI application to your trading:
Let’s move from the more theoretical part now to the second, more practical part and see how AI can already today significantly boost your trading performance:
As said, at hoc-trade, we already run these analyses, and we identified plenty of patterns commonly seen for traders with huge effects on their performance. We collected forty of them in hoc-trade right now already. As of now, we collect them all in a dashboard and send you an alert in case we identify a new one for you, or see that you act in one of your patterns again that more often than not makes you a loss. We are in the last step of internal testing, but will start our beta test soon. If you would like to have a completely free testing access and become one of our OGs, just join our Discord server (LINK). All communication will be there, lots of interesting things to come, that I can assure you. This AI can of course be applied to any trading market, we start off with Crypto and Forex trading.
Let’s have a look at 2 interesting examples: One of my favorites so far is what I would call revenge trading.
We found a pattern that the time between closing and opening a new trade can have a big influence on your performance of that trade, and that influence is especially strong if the previous trade was a loss trade. So what we can see here, is that a trade which is opened within 5 mins after a loss trade has a much worse performance compared to a trade in which the break was longer. In fact, this trader lost more than 1,100 USD on those trades opened within 5 mins after a loss trade, or 7.1 pips or on average 0.44% per trade, while in a trade which was opened with more than an hour break after a loss trade, the trader on average gained 0.24%.
Now thinking about it from a trader perspective and the emotions and behaviors a trader might have in such a situation, this actually makes a lot of sense. Shortly after a loss trade, you may feel disappointed, angry, annoyed, and want to make up for your loss again, therefore quickly entering a new trade. It also makes a lot of sense that this trade out of emotions is not as well planned, structured and executed compared to a trade which you enter with a more calm and clear mind. We are only at the beginning of this intelligence entering and improving our trading, but this is great example of how it can do this.
In the second example we found a really interesting pattern for traders during their loss trades.
As said, hoc-trade also merges the trading data with the price data of the trade, and it found that oftentimes your Stop Loss adjustment during a loss trade makes things much worse compared to original Stop Loss level. So here we see that a loss trade in which the Stop Loss was adjusted, the average loss is 11.4 pips, while in the original stop loss setting, the loss is only 5.2 pips per trade.
Looking at this from a traders perspective again, this also makes a lot of sense. Accepting losses is tough, is against our human nature, and oftentimes we try to “save” our trade by moving or even completely removing our Stop Loss, which actually makes it worse though. Again, all this data can be collected and analysed pretty much in real-time, and with more and more trades and traders, it gets more complete and accurate. So far, we’ve found & collected around 40 patterns, but this number will surely increase in the future. Quite a few of those probably have a direct impact on you as well. The great thing is, that you do not need a very long trading history to find those patterns for you, as the patterns are learned from a huge dataset which give plenty of insights, and then applied to your trading and see whether similar significances apply.
How does it work? Here is the process:
Ok, so how can this work and how can AI now directly improve your performance? Let me give you a bit of an overview of how we build this up.
In order for the system to find and apply those patterns, quite a lot of data is needed, but we built a fully automated process so no manual work from a trader is needed.
First, you can connect your trading account through a read-only API or in case of Forex, through your read-only investor password. No access is given to the account, just the reading of past and new data is enabled. The trading data flows into our system, and it automatically also retrieves all quote data, indicator data, news, and much more, to add this extra level of intelligence of merging your trading data with real price feeds, etcetera. As a next step, our tools searches for patterns in all the trading data, and analyses whether your trading shows patterns already previously identified from other data.
All your outputs will be automatically presented to you in your customized dashboard, you will see the effect of your personal patterns, how you improve over time, and most exciting: You will get near real-time alerts in case you act in one of your previous patterns again, a new pattern is found for you, or the risk profile of your current trade deviates significantly to your previous trades.
Thereby, this continuously improving intelligence can already today strengthen your trading performance a lot, showing you exactly what kind of trading behaviors are averaging you a loss, and warning you in case you are doing it again. I believe this is extremely powerful already, as we know that it’s those trading behaviors that make the vast majority of traders lose money, and not the missing indicator or chart pattern. Even if you didn’t know at all how to read a chart, you should still have a 50% winning probability, and not the 90% of loss traders as we see it in the market.
While this is not (yet) a fully independent AI that makes trading decisions, it is already a very big help for your trading which gives you an edge over other traders. This tool is of course only the first step in the evolution process, and we have many exciting things already planned, areas in which AI can have an even more direct influence on your trading, but this goes a bit far for now and will be part of future articles. So if you are interested, subscribe our channel to stay up-to-date for all future developments.
How to access?
Now, most important, it’s only a few days left until we set all of this live what you’ve read here in the article and it can be used by the traders in our Discord server. It’s all free, no strings attached, and of course we will also reward our first adopters, the only thing is we may have to limit the accounts to not blow up our servers initially. So if you think that’s interesting, you are very welcome to join our Discord server.
AI in trading is definitely one of the most interesting applications out there, and it doesn’t have to be fully proprietary within a hedge fund that develops smart systems to execute trades, but can have very tangible performance effects for all of us as a retail traders.
I hope you liked this small snapshot into how AI can already today improve your trading performance, and if you did, please leave us a clap (up to 50 possible :))and subscribe the channel. We are a young start-up and only getting started in this area, so lots of interesting trading, AI, and analytics content to follow.
Thanks for reading, happy trading, and stay safe!
For all updates, also follow us on Twitter!
