Today’s “Long Story Short” discusses how to turn a bet into a systematic trading strategy. Kollider’s series is purely educational and not investment advice. Please do your own research.
The cryptocurrency trading boom has given everyone, not just professionals, direct access to a wide range of exchanges, products, and tools that can be used to make money. Yet, in spite of all the new developments and opportunities that crypto trading brings, most “trades” do not differ that much from “bets’’.
This begs the question, when does a “bet” become a “trade” and what sets apart a single “trade” from a “trading strategy”?
Today, we propose the concept of a minimum viable strategy (“MVS”). This framework should allow you to systematise an idea by focusing on three aspects of trading.
Entering or Increasing a Position
For the MVS, the entry decision is the most familiar and easiest to think about. In case you have examined what makes a market move, then you have already started to think about how to enter a market. This part of the framework refers to both the signal that gives you an idea of where the market will go, and the execution that turns that signal into an order, either manually or algorithmically.
When discussing entering or increasing a position, we can focus solely on the signal because the execution itself could be very simple in practice. A typical entry execution just sends a bid at the best ask when the signals suggest the market will trend upward. In contrast, it sends an ask order when the signals suggest the market will trend downward. This is actually perfectly viable, especially with the addition of some minor risk management rules, as we’ll cover shortly.
The entry signal has just one requirement. It must output a market direction.
But it can be anything underneath. For example, a bitcoin miner might construct a signal based on a relationship of bitcoin’s prices to the hash rate. A crypto analyst might create one based on the changes of transaction volume between specific addresses on the blockchain. In the Joy of Algorithmic Trading, we talked about a fictitious momentum trade that used an MACD as an entry signal.
The entry signal’s output can be continuous and always has an answer on the market direction. Or it could be sparse and only give non-neutral answers at times. The hash rate signal example above exhibits a continuous signal. It could be converted into a sparse signal (e.g., if we instead take the rate of change of the hash rate and use higher cutoff points before outputting a market direction). Similarly, we could set a threshold for the MACD signal to trigger only after it’s crossed at least X percent to force it to give less frequent, sparse outputs.
Managing the Strategy’s Risk
Once you have the entry signal, you now have to make a decision on the risk you are willing to take. Although it could be tempting to just keep executing your signals (e.g., buying whenever you receive a bullish signal), you are probably going to want to place some limits as there will always be a chance you could lose most, if not all, of your capital.
For us, the MVS should at least have two self-explanatory components: the maximum position you are willing to hold will prevent you from increasing your position even with more and more signals (the bonus here is that additional signals can be treated as confirmation of your existing position); and the order size you would send to the market each time you receive a signal. For continuous or high-frequency entry signals, you may also add a time throttle variable that forces the execution to wait when receiving back-to-back entry signals in the same direction.
There are more fine-grained optimisations that you could do (e.g., getting the margin right), but these are optional for the MVS.
Exiting or reducing a position
Knowing when to exit a position is often the most difficult part of the MVS framework, and some novice traders may not immediately realise its importance. Unfortunately, it might only become apparent when a trader consistently loses money despite an exceptionally accurate entry signal.
Just as an entry signal can be composed of anything so long as it gives a direction, an exit decision only has one mandate: it has to reduce a strategy’s holding at opportune times before it turns into a losing position.
There are many ways traders could go about this.
In the beginning, they might just use the entry signal itself. If the strategy is currently long and the entry signal gives a bearish direction, then the strategy can use this to reduce or completely close out its position. However, whether this works well depends on many factors, including the output frequency of the signal. As an example, if the signal is very sparse and only gives you a direction every few weeks, you might be forced to hold onto a position without knowing for weeks whether it is good to keep or not.
They might go with a variation of the entry signal. Depending on the signal, one possible way is to increase its frequency of signals by lowering the trigger threshold. For example, suppose we are using an MACD crossover that considers a crossover of 5% as an entry signal, we might have to use 1% on the other side as a good-enough signal to exit a position. Similarly, we may consider reducing our position if an entry signal like the MACD goes to zero.
We can also specify an expiration time for our entry signal, after which we’ll automatically reduce our positions unless a new entry signal arrives. This allows us to consider whether the reason we’re in a position has become “stale”.
You also don’t need a signal to exit your position. It might even be advantageous to use fairly static variables, especially if you have very sparse entry signals and if you are looking for events that rarely trigger. Some commonly-used ones for these are the use of take profits and stop losses with a predetermined threshold from the entry price. You might even consider trailing your stops, but both fixed and trailing versions are used in practice.
Once we’ve completed the three parts of the MVS, then it’s time to reconsider our initial question: when does a “bet” become a “trade” and what sets apart a single “trade” from a “trading strategy”?
In the absence of a strategy, every trade is just a bet, no matter how solid the conviction or signal may be. It is only by being methodological and using a framework like the one outlined above that simple ideas can be transformed into strategies. In other words, if you plan each aspect of a trade, such as trade entry, risk management, and exit strategy, you can distinguish between a “bet” and a “trade”, or an individual trade from a “trading strategy”.