What is the mean delta?

Good afternoon,

Ref A: coworkly | few_words/coworkly goal page
Ref B: What are the goal statistics? - Beeminder Help

Ref A: My current rate is 3 Hours per day since I started this commitment on 4 January 2022. Beeminder states I’m averaging 2.9 per day. I understand this.

I’ve read ref B.

My understanding from basic English is that “average and mean” are the same thing, and yet Beeminder is showing me significantly different numbers. This confuses me.

These numbers are very important to me.



Graph stats

  • AVG RATE 2.9 per day

Data stats

  • DATA PTS 73
  • MEAN 184.4
  • MEAN DELTA 4.5
  • 90% VARIANCE ±5.0

If someone could explain to me what the “data stats” for my Commitment are, that would be super! Please keep in mind my mathematical/statistical knowledge is largely at a secondary school level.

To be clear, I’m asking for a qualitative explanation of the significance of the “data stats”, explained at a secondary school level.

Thank you!


Oh, graph stats are telling you things about the bright red line and data stats are telling you about your actual datapoints. Does that clear up some of the confusion?

1 Like

As Danny says, the graph stats are about the bright red line and the data stats are about the datapoints.

The mean delta tells you about how much you typically do on days when you do anything toward this goal. In your example, you typically achieve 4.5 hours of progress on each doing day.

The average rate number isn’t directly comparable, because although the slope of your bright red line has averaged a steepness of 2.9 hours per day, that’s every day, not just doing days. So if almost every day is a doing day, you’ll have accumulated a hefty safety buffer, but if you take every third day off then you’d be constantly skating the edge with most days being ‘emergency’ days.

90% variance* is about how much your values vary, relative to the bright red line, day-to-day, in the almost worst case, which might be most helpful in managing other types of goals, like weight loss, that have some variability that’s beyond one’s direct control.

The mean in the data stats is a bit of a mystery as far as knowing when it might be useful. Without looking at the code, I think it’s the average of the cumulative aggregated datapoints over the life of the goal. In a decade of beeminding I’ve not once referred to that.