How to use charts Indicators and the new distribution filter

Updated in October 2026. This article was first published in December 2023, when indicators were added to the chart. The editor has changed since then: the indicators are now the statistics table beside the Forecast chart, and the distribution filter is the bars next to it. The steps below describe the current editor; the pictures further down are from 2023 and are captioned as such.

In Machinations, a Forecast plays your model many times and shows what could happen, not only one outcome. The Forecast chart summarizes all of those plays for each node you track, so you can judge the health, efficiency or progress of a part of your model at a glance. Instead of reading hundreds of lines, you read a few numbers that describe them.

For instance, you might track how many Items a crafting model produces by step 50, across 200 plays. The chart shows the typical path, how wide the spread is, and how often each result happened.

How to open the chart and read its statistics

1. In the playbar, choose how many plays to run in the number next to Forecast (for example 200), then press Forecast.
2. When the Forecast has finished, click a node’s card under the chart, for example Items. The chart of that node opens.
3. Read the table on the right. It lists seven statistics for the step marked on the chart (step 50 here):

• max: the highest value any play reached at that step.
• p99 and p95: the value that 99 and 95 of every 100 plays stayed at or below.
• med: the median, the middle value. Half of the plays ended below it and half above it. The line drawn on the chart is the median over time.
• mean: the average, the sum of all the values divided by the number of plays.
• p5: the value that only 5 of every 100 plays ended below.
• min: the lowest value any play reached at that step.

Fan chart or line chart

At the top of the chart, Trajectory switches between two views of the same plays. Fan chart draws the median line with a band for the middle 50% of plays (P25 to P75) and a lighter band for 90% of them (P5 to P95), so you see the spread at every step. Line chart draws one thin line for every play, which is useful when you want to see single plays behave differently from the rest.

How are averages calculated?

In the example above:

The Mean: represents the Mean of the plotted Node’s values at a certain Step, across all numbers of Plays executed. For the Common pool in the picture, 70 is the Mean of all the values the pool reached at pull 20, across 100 plays. The Mean is calculated as follows: the sum of all values, divided by the number of values. E.g.: The mean of a, b, c is (a+b+c)/3.

The Median: represents the median value of all values of the plotted Node, at a certain Step. For the Common pool, 70 is the median of the values at pull 20, across the 100 plays; it is “the middle” value of your set.

In your analysis, use the median in comparison to the mean, to check if your data set is not skewed by a small proportion of extremely large or small values. This comparison will provide a better representation of a “typical” value.

Min/Max Values: The minimum/maximum values in the data set of the analyzed Node, at the same Step of multiple executions. In the example above: the Min and Max values are 58 and 81: the lowest and the highest values the Common pool reached at pull 20 in any of the 100 plays.

Distribution filter:

The distribution is the set of bars beside the chart. Each bar is one possible value at the chosen step, and its length tells you how many plays ended there. The bar menu at the top right, Bars, has two settings:

• Shape: Distribution shows each value on its own; Cumulative adds the bars up as you go down the chart, so the bottom bar counts every play.
• Values: % of plays labels each bar with its share of the plays; Count labels it with the number of plays.

Clicking a bar selects the plays that ended in that range. A chip such as 39 of 200 plays selected appears above the table, the chart draws only those plays, and the table recalculates for them alone. To see all plays again, click the x on the chip.

Interprating the histogram:

Let’s read the distribution bars of the same Common pool after 100 plays of 20 pulls.

At pull 20 the bars group the final values into ranges. The most frequent result is the range 68 to 72, which 42% of the plays reached; the ranges on either side hold 22% each.

So by pull 20, a player has a 42% chance of holding between 68 and 72 Common cards.

As you can see, the distribution histogram is a powerful tool for identifying the consistency and range of outcomes in a simulation. It allows the game designer to understand not only the average outcome but also the spread and variance, which are crucial for balancing the game’s crafting mechanics.