All of the rules in the scoring models have something in common and a few things that are different.
These are things that are repeated in every rule:
Name & Category
Choosing the Name is up to you, as well as the Category, but we suggest using relevant names/categories so that it would be easier to read and understand the results at a later stage.
Maximum Score
This is the upper limit of how high the best company should be scored (rewarded with how many points?). The maximum score threshold doesn't matter much to the single metric, but it is a good way of overweighting or underweighting the significance of the particular metric compared to the other metric.
The pure numbers you use are relative and don't mean much since our algorithms use rankings with two decimal places anyway, but the number (maximum score) starts to mean a lot compared to other metrics' maximum scores.
Grading Scale
This is your definition of what it means to be a good or bad metric reading. In other words, this is the place to define the "Worst" and "Best" thresholds for the particular metric. When the metric reading crosses the "Best" threshold, the company receives maximum scoring; however, it receives zero points when the "Worst" threshold is crossed.
Everything between receives some points accordingly to the place of the reading on a grading scale (i.e., the closer the reading is to the "Worst" threshold, the fewer points the company receives; closer to "Best", the more points it's rewarded with up until crossing the "Best" barrier, in which case it receives the maximum points available).
Must-have Rule
This might be checked if you want the just-added rule to be treated as an extremely important one, meaning that if the company doesn't score better by least 0.01 points than the minimum threshold, it would be put at the bottom of any list of the companies that you add it to later. Consider it a forced downgrade, no matter how good the company scored in other categories.
Basically, it's better not to use this option unless you really want to definitely get rid of the companies that do not meet your criteria (and rewarding them with zero points is not a sufficient punishment).
You might want to tick the Must-have Rule checkbox in metrics like Market Cap or stock Price to eliminate the companies with low capitalization or penny stocks traded for less than 5$ per share.
Now it's time for the unique options in every rule. Let's figure it out rule by rule.
1. Metric Value in Static Range
This is the simplest rule since it requires you to provide only a few easy-to-understand parameters.
In our example, we try to add the Debt-to-Assets ratio, which means that the lower the debt level is compared to assets, the better situation of the company, hence we need to reverse the threshold (as on the screen) and tell the algorithm that the "Best" situation is when the company has 5% of debt to assets or less; and the "Worst" situation is when the company has 40% or more of debt.
If we were using a metric like Return on Invested Capital or the Operating Margin, we would oppositely set the threshold since we want the ROIC or Margin to be as high as possible. Thus, the Grading Scale for the Operating Margin would look like this: "Worst = 5%, Best = 40%".
2. Metric Value in Historical Range
This is similar to the above, except that you want the range (and grading scale) not to be fixed but rather unique for every stock, so the highs and lows (the range) are calculated automatically based on historical data.
Here, you are providing the Maximum Score available to the particular Financial Metric you want to use, and then you need to tell our algorithms a few things not mentioned in the previous rule.
First of all...
Higher Values Preferred – yes or no?
We need to know whether you'd like to give maximum points to a high or low reading of Financial Metric. In this case, we want the P/S to be as low as possible to consider a company a bargain so we won't tick this option.
But if the Financial Metric was set as the Upside, we'd like to check that option because the higher the upside, the better the opportunity to buy a stock, hence we'd want to give such a reading the maximum points available.
Years for Auto-Channel
Since range and grading scales are dynamic in this rule, our algorithms need to know how far back in history they should look for a pattern to determine a range typical to the particular company. Basically it looks like this:
The algorithm is trying to automatically measure and visualise the frequent highs and lows (based on the standard deviation from the mean), so it can assess later where the given metric currently is comparing it to its typical historical range.
In general, we need at least a few years of data to draw a proper channel, so 5-10 years is usually a good option.
Horizontal Auto-Channel
This option tells our algorithms that you want to draw only horizontal channels and not diagonal ones. This is a decent choice when assessing the metrics typically horizontally moving within their static ranges. The upside is a good example; to some extent P/S or P/E also fits that definition.
That being said, there are exceptions, and for some companies or some indicators, the reading moves diagonally, not horizontally, so you might want to deselect the Horizontal Auto-Channel option in such circumstances.
By the way, you can adjust these channels manually if you want in case some peculiar company doesn't fit the definition above. Please refer to this article: Adjusting horizontal channels manually (available soon).
3. Difference between two metrics
Use it to assess the desired metric, not within static or dynamic (historical) range, but rather comparing it to another metric.
The settings are very similar to the Metric Value in Static Range Rule, except here, you are comparing different metrics to each other. To be more precise, you are comparing the Secondary Financial Metric to the Primary Financial Metric.
In our example above, the company will be rewarded with zero points whenever the Revenue Estimates for Next 12 Months are not higher by at least 2% than the current Revenue TTM. On the other hand, if the Estimates are at least 15% higher than the Revenue TTM, we'll give such a company maximum scores (8 points in this case).
Everything in between will be graded with some points but neither none nor maximum.
4. Metric Change Over Time
This one helps to score a company by determining how much the particular metric has changed over time (how much the metric has risen or declined over time in the percentage values).
The only new thing here is the Timeframe you need to provide as a starting point to measure the metric's change over time.
In that case, we want to assess whether the Revenues TTM rose at least 10% during the last five years. If not, the company will be rewarded with zero points. If the Revenue grew by at least 65%, it would be given maximum points.
Everything in between will be graded with some points but neither none nor maximum.
5. Metric Decline Over Time
It works similarly to the Metric Change Over Time, except that we can measure only decline here, and the starting moment to gauge the decline is not defined by a point in time but by the last peak of the metric (the highest reading ever that occurred during some period).
You can configure that metric similarly to Metric Change Over Time with one exception. Since we try to measure the decline from the last top (highest reading) that occurred during the Timeframe provided, the Best result can't be higher than 0% (of decline).
So the Worst parameter on the grading scale must always be provided as a negative number, and the Best parameter can't be higher than zero.
This metric is really useful if you want to monitor the EPS or Price Target revisions.
To better understand these rules, please read the article with extended descriptions and a lot of examples: Different types of rules you can use