A backtest is the process of using the trade data from the exchange to simulate how a strategy would have performed over a given amount of time. This is often used to test the viability of a strategy by running it through these large data sets.
How Shrimpy Uses Backtests?
Shrimpy often uses backtests in order to compare the results of rebalancing against those of HODL. The number of backtests we typically ran for each portfolio size and rebalance period pair is 1000. If this is not the case for a backtest, we will often explain in writing how many we decided to use. 1000 is typically used because it was determined to be sufficiently large to produce an obvious trend. Read more about backtests or run your own.
What does a typical backtesting process look like?
Let's take a walk through a typical backtesting scenario that is used in many of the articles and discussions that we post online. First, we often have a range of values that are used for both rebalancing periods and the number of assets in the portfolio. Since they are a range, we need to iterate over the values and run a backtest for each of those values. For example, if you have a range of rebalance periods from 1 hour to 1 month, we would start by setting the rebalance period to 1 hour. In a similar way, if we have a range of number of assets we would like to evaluate. Say, between 2 and 10 assets in the portfolio, we can also iterate over the number of assets.
The result is a grid where every asset count and rebalance frequency is evaluated by pairing them together. So, we would start by evaluating 1 hour rebalances with a portfolio of 2 assets. After running that test 1,000 times, we move onto the next pair. The next pair would then be 1 hour rebalances with 4 assets. And so on, until all rebalance periods and all number of asset pairs were evaluated. (It should be noted in this example, we only evaluate the even portfolio sizes because it would take a long time to evaluate every single portfolio size)