Skip to main content
All CollectionsBacktests
How to set up Backtests
How to set up Backtests

Backtest with scanning criteria, rebalancing, and more. Advanced options like Initial Portfolio Value and Buy/Sell criteria are available.

Updated over a week ago

Backtesting your strategy

Validating your strategy is a great way to confirm how your criteria would have performed in the past. Click the Backtest button on the left navigation pane to open Scrab's Backtests.

For Backtests to work, you must have at least one scoring model. You can use one of your pre-existing scoring models, create a new one, or use one of Scrab's pre-set options.

We made it easy to use Scrab's backtesting feature, as you will only need to do the following:

  1. Define your initial scanning criteria for Scrab to Scan: Scoring Model, Companies, and the Backtest period.

  2. Tailor your backtest parameters: Rebalancing Frequency, Portfolio Size, Asset Allocation, Required Years since IPO, and Low Score and Missing Data Tolerance.

For an even more custom backtest, after calculating your backtest data, you can use the Advances setting of Initial Portfolio Value, select Reduce Portfolio Turnover, and, as an optional feature, use the Buy & Sell criteria.

Also, as you run backtests, we will keep a historical record of the backtests you've run on the left-hand side of the backtests view, labeled 'Recent Backtests.'


Initial settings

Once you're in the Backtests screen. Edit the following:

  • Scoring Model: Use the dropdown to choose which strategy you'd like Scrab to backtest.

  • Companies: Choose one of the 3 backtest variants.

  • Backtest Period: Select the time range you'd like Scrab to backtest. Our platform can scan as far back as 01/06/2006, but we suggest only going back 5-10 years.

Now that you've chosen your scanning criteria, click Calculate Data for Backtest to go to the backtest parameters.


Backtest parameters

After the historical data is fetched, you can now choose how you'd like Scrab to use that data:

  • Start date, End Date – This is displaying the date range you've chosen to analyze

  • Rebalancing Frequency – We've set the default to update your portfolio every month, but you're welcome to experiment with different times to see what works best. Usually, sticking to monthly updates is the most effective approach.

  • Portfolio Size—Scrab will sort companies in your strategy from best to worst based on our data. This option helps you decide how many top performers to include in your portfolio. Do you want the top 20 or perhaps the top 50? Aiming for about 25-30 of the best is usually a good choice.

  • Asset Allocation– Select how you would like the weights of your stocks to be determined. You can choose from four options:

    • Equal Risk - Each asset will carry the same amount of risk, determined by comparing the historical price changes and the potential for loss among all the stocks you're considering.

    • Equal Weights- All assets will have the same position size.

    • Hierarchal Risk Parity- Each asset will share the risk equally, grouping similar assets to achieve a well-balanced mix without putting too much weight on any asset.

    • Proportional to Metric—The asset weight will be directly or inversely proportional to the latest value of the selected metric.

  • Required Years since IPO—Select how many years the company has been on the market. We recommend choosing companies with a minimum of 2-3 years post-IPO to ensure reliable historical data analysis. Newer companies are not excluded but ranked lower to balance established and new market entrants, letting you decide on the quality versus quantity of data for your strategy.


Tolerance for Low Scores & Missing Data

This option helps you let our algorithms know how to treat companies with low scores or missing data. You have four options:

  • Low tolerance means that the company can't have more than 30% of missing or poor data (the data is considered poor when its total rules scoring is less than 40%)

  • Medium tolerance means the company can't have more than 45% of missing/poor data.

  • High tolerance means that no more than 75% can be missing or poor-quality data

  • Full tolerance means we'll consider all stocks regardless of the quality or lack of their data.

When backtesting a detailed strategy with many rules (more than 10) and many companies (300-400), pick a lower tolerance to focus on the highest quality scores.

For simpler strategies (1-2 rules) or fewer companies (less than 300), choose a higher tolerance so you can still identify your top picks, even with stricter criteria.

This way, you adjust the level of strictness based on your strategy's complexity and the number of companies you're considering, ensuring you get the most relevant results for your analysis.


Additional Advanced Settings

Expanding the Advanced Settings dropdown, you'll see:

Initial Portfolio Value

Enter any starting amount you'd like to simulate the potential outcomes of your investment strategy more accurately.

Reduce Portfolio Turnover

Tick this option to minimize frequent stock swaps in your portfolio, only making changes when scores differ significantly (more than 1%). This reduces unnecessary trades and maintains stability in your investment strategy.


Buy and Sell Criteria

In backtests, companies for a portfolio are typically selected based solely on the highest Total Score.

Buy Criteria

Suppose you set Buy Criteria (for example, Growth >= 50% and Value >= 70%). In that case, companies will still be selected based on the highest Total Score, but only if they also meet the conditions set in the Buy Criteria for one or more categories. This means that a company with the highest Total Score might not be included in the portfolio if it falls short in one of the specified categories. If the Buy Criteria are too strict, none, or only a few, of the companies from a large pool may meet your requirements and be included in the portfolio.

Sell Criteria

Sell Criteria is a condition that removes a company from the portfolio automatically. For example, if a company was included in the portfolio initially because it had a high Total Score and met the Buy Criteria, but after a period of rebalancing its Growth score falls below 40% (as specified in the Sell Criteria), it will be removed from the portfolio, even if its Total Score remains high. This company will be replaced by another one, provided a company meets the Buy Criteria to take its place.

Sell Criteria are a safeguard, helping you remove companies under specific conditions, like falling growth forecasts, valuations becoming too high, or other metrics deteriorating.

Both of these criteria allow you to explore different strategies, such as buying companies with attractive valuations (e.g., Value > XX% in Buy Criteria) and selling them when they become overvalued (e.g., Value < YY% in Sell Criteria) or removing companies with negative forecasts or other poor metrics.

Once you've configured your additional settings (if any), click the Run Backtest button.

From here, you can interpret the backtest results.

Did this answer your question?