Grid-trading and DCA bots
A grid bot lays a ladder of buy and sell orders over a price range and earns from every up and down. That only works as long as the price stays in the range, and fails exactly when a market runs strongly in one direction.
A grid bot automatically spreads buy and sell orders at fixed intervals across a price range you choose. An example: Bitcoin is at 60,000 dollars, and you set a grid between 55,000 and 65,000 dollars with a step every 1,000 dollars. If the price swings back and forth within it, the bot buys low and sells high, again and again, without anyone having to predict the direction.
A DCA bot, short for dollar cost averaging, works differently. It buys a fixed amount at fixed intervals, wherever the price is. At its core that's the same as a savings plan from Stage 2, just for cryptocurrencies and often at shorter intervals.
The catch with the grid bot lies in how it's built. It's made for markets moving sideways and fails when things clearly go one way. If Bitcoin rises to 80,000 dollars in our example, the bot has already sold everything by 65,000 and misses the rest. If Bitcoin falls to 40,000 dollars, the bot buys more at every step down to 55,000 and then sits on a large holding worth much less than it paid.
A word of caution about the figures grid bot providers advertise: claims like double digit returns per month with supposedly low risk almost always come from the providers themselves or from marketing copy nobody has checked independently. Unlike momentum or value, there's no independent long term study for this kind of product that you could rely on.
Technically, a grid bot is an automated form of very short term mean reversion within a fixed price range. Each completed round of buying and selling brings a small gain equal to the distance between two steps, minus fees. How many such rounds there are and how much the bot earns in total depends on how much the price swings within the range. The direction of the market doesn't matter for that.
The risk is unevenly spread and lies outside the price range. If the price breaks out downwards, the bot builds an ever larger position whose value keeps falling. Economically it's as if you automatically kept buying more with no limit on the loss. If the price breaks out upwards, you miss the difference between the top step and the high actually reached. That's no loss in your account, but economically it's still a price you pay.
Unlike momentum, value or pairs trading, there's practically no peer reviewed research on grid and DCA bots with cryptocurrencies documenting returns over long, independently collected periods. The figures that exist come almost entirely from the platforms that sell or run these bots. That's exactly the conflict of interest described in the lesson Evaluating Sources. Anyone wanting to try such a strategy should work it out themselves with old prices and realistic fees rather than taking advertised success rates on trust.
Summary
- A grid bot earns from ups and downs within a price range, whichever direction.
- If the price breaks strongly out of the range, the strategy fails, upwards and downwards in different ways.
- There's hardly any independent research on these products, only figures from the providers selling them.
Did you get it?
In which market phase does a grid bot work best?
When the market moves sideways and the price swings within the chosen range.
What happens if the price falls sharply below the grid's lower bound?
The bot buys more at every step and builds an ever larger position whose value keeps falling.
Why should advertised return figures from grid-bot providers be treated with caution?
Because they mostly come from the platforms selling the bots, with no independent study behind them.
Related
- Momentum and trend followingStrategy Types
- Mean reversionStrategy Types
- Pairs trading and statistical arbitrageStrategy Types