There are a vast majority of algo bots (especially for crypto trading) available on the internet.
I’m not here to advertise any of those bots, so I won’t mention any of them, but you can easily google them. Some crypto exchange platforms embedded them into their system, but there are a huge collection provided by “third parties” as well.
Track live crypto price of 10000+ coins!
note: Be very cautious when using third party provided bots! It means you have to trust them, give them your money, and let them do whatever they want with it. They may even scam you by showing some unfortunate losses in the transactions while no transactions happened at all.
In this article, I’m about to give you a brief introduction to what these bots do to help you choose the appropriate one for you. I think there are 3 main types of bots based on trading logic. All of them are searching for signals to trade, but each finds them in different ways. Some try to predict the future by various indicators and events. For example, a rising low RSI might indicate an uptrend, though these bots take several indicators into consideration. Others follow a strict strategy, a similar strategy to what a trader would do by themself, but this bot automates the process. For example, setting up a regular buy scheme or using a grid to determine when to sell and buy. The third kind of bot does not try to predict price movement, rather, they look for arbitrage opportunities. This means that when an arbitrage bot makes a set of transactions, you always end up holding only fiat coins (USDT, for example) and never be exposed to the risk of price movement. Of this, I’ve already posted two articles:
Price Predicting bots:
Such a bot continuously watches a wide range of the asset’s historic price, viewed in an OHCL candlestick graph. There are numerous indicators that can be drawn on such a graph (of course, the bot won’t necessarily draw these graphs, it only checks the values). These indicators are well known by every speculative trader, they often spend their whole day watching graphs and searching for patterns and signals, and they often end up seeing what they want to see rather than what they really see. These bots help you watch the graphs automatically which means such a bot might be only semi-automatic by drawing you relevant graphs with interesting indicators and even showing you buy and sell signal advice. Though, of course, the human interaction can also be skipped and let the bot trade whenever it sees a signal.
The profitability of these bots is very diverse. Some believe that these indicators do help you trade, while others think that the past does not have any effect on the future in the concept of stock/crypto exchanges. (Side note, to mention that the latter one is proven by mathematics by defining price movement as martingale, a.k.a purely random.)
In case you want to build such a bot, here’s what you’d need: You continuously have to scrape/download OHCL data of an asset, with a minimum of 200–300 data points. Decide what indicators you might use. For this, keep in mind that most of the indicators do the same, but in a slightly different way, so stacking up the “same” indicators might overweight false signals. Mainly there are three kinds of indicators: momentum-based, volume-based, and volatility-based. Make sure to evenly gather indicators from each category. Calculate every value for each indicator on each data point, and search for buy/sell signals. First of all, you also have to determine what is considered a signal. For each indicator you use, there are predefined indicators on the internet which are based on long-time experience, but you can find patterns as well. Make sure to check values dynamically, not only the static values. This means, that an RSI=30 might indicate a buy signal, but only if it is rising at the moment. In other words, if the RSI line crosses the 30 from below, it might be a buy signal, but not otherwise.
One major advice on building these kind of bots is to also determine the current trend. This will be the most difficult part, since you cannot determine it exactly. It can be an uptrend on the 5m period graph for the last 2 hours, meanwhile it is a massive downtrend watching on a 2hour period graph for the last 2 days. Moreover, even if you can clearly see a trend by eyes, it will take high efforts to teach your bot what is considered as bullish, bearish or side-walking. Though if you fail to determine the current trend, it could jeopardize your bot’s profitability since every indicator works differently in different trends. A buy signal in an uptrend might not be a signal in the downtrend.
Other than indicators, speculative traders tend to watch patterns, as well, like rising bottoms or head-and-shoulders, etc. Such patterns to be determined programatically is even more difficult than determining trend, though unluckily patterns seem to be slightly more reliable than indicators.
When you managed to check every indicator/pattern and trend for each datapoint, you then have to tune your parameter to find true signals and minimalize false positives. You can always simulate your current parameters and calculate how would it perform on different scenarios, and try to tune it by hand. Also you might even build an AI on top of your bot to tune itself.
Strategy Following bots:
This kind of bot is much simpler than the previous. Traders usually follow some simple strategy that helps mitigating the human factor in trading. Of course this mitigation can be maximized if the strategy is even executed by a robot rather than a human. There are a few base strategies, but you can also come up with anything that suites you.
The simplest strategy is the DCA — Dollar Cost Averaging. This ensures that when you decide to invest a bigger amount, you buy the asset at an average price. In other words, this protects you from buying at the wrong moment, when the price has risen temporarily. Of course, it also eliminates the opportunity to buy cheap. In fact this strategy is not profit-oriented portfolio management but rather just a tool to build your portfolio optimally. To build such a bot, you only have to predetermine the amount you want to invest and the periodicity. For example, you can set up a recurring buy of BTC every day for 50 dollars. That means by the end of the month, you’ll have a total amount of 1500 dollars invested in bitcoin, with an average cost based on the month’s price of BTC.
The second most intuitive strategy is the Grid. The logic behind this method is to always buy low and sell high. This can be achieved if you buy when the price is falling and sell when it is rising. The strategy is called Grid because technically, we are drawing (horizontal) gridlines on the price chart and buy each time when a line has been crossed downwards and sell when crossed upwards. The number of grid lines can be finite therefore, it stops if the price has reached a minimum or maximum price, or can be infinite, so the bot keeps trading in any circumstances. This bot is perfect for side-walking trends but can be profitable in bullish and bearish markets as well, but everything depends on the right tuning of the model.
These bots, as the name implies, aimed to be risk-free. Due to price fragmentation in the market, assets might be slightly overpriced or underpriced for a short period of time. If you can find such opportunities and you are quick enough, you can make either a chain of a transaction or separate, simultaneous trades, which consist of both buy and sell transactions. The main concept is that you either buy the same asset on one exchange and sell it on the other, or staying on the same exchange, buy an asset, trade it to another, and sell that one. In any case, you start with USD (or any other fiat coin), and end up in the same currency, also, the transactions are executed in a fraction of a second. For detailed descriptions, read my other posts on the topic:
Download MAXBIT Android App, Your best source of all crypto news!
Share this article: