We need a one-page chart that shows the empirical cM values found for various relationships. We know the theoretical, or calculated values, but the randomness of DNA results in a fairly wide range in some cases – particularly for distant cousins.
The chart below shows my guess as to what a chart might look like. The x-axis is cMs on a logarithmic scale. The y-axis is % of all the values for each cousinship (the number of results at each cM value divided by the total number of results for that cousinship – which would normalize the chart for different total number of results for each cousinship). The area under each curve would be 100% of all results. The roughly normal distribution curves are “centered” on the calculated cM values for each cousinship. Based on experience we know that first cousins (1C) tend to share segments with cM values relatively close to the calculated value of 880cMs, producing a tall thin curve (I think); whereas 5C (calculated average 3.4cM) or 6C (calculated average 0.8cM) must have long cM “tails” on this chart in order for us to “see” the shared segment with Matches which are above a 7cM threshold, producing a short wide curve (I think).
Note in this hypothetical chart, the small red dots at the end of some tails were taken from the data compiled by Blaine Bettinger (who did a great service to us all by compiling and reporting this data), which can be found at:
We need this data displayed this way so we can easily enter with a shared cM value on the x-axis and see the range of cousinships possible. This would quickly show which cousinship is most probable, and how close, or far, other cousinships would be.
As I now think about it, at any cM value on the x-axis, wouldn’t the sum of the values of all the curves have to equal 100%? But to achieve that, we’d have to include all the possible curves, including siblings, half siblings, half double second cousins twice removed, etc., which is probably impractical at this point. I’d rather see the chart soon with the cousinships shown below, than wait a long time for the perfect chart.
Another thought is to blow up the part of the chart from, say, 5cM to 50cM. This would be fairly simple once the data is collected.
Still another observation is that if this chart were based on all collected data, data based on endogamous shared segments would generally be shifted a little more to the right; and data based on non-endogamous shared segments would generally be shifted a little more to the left.
BE CAREFUL – THE CHART BELOW IS A THEORETICAL GUESS (with only a few valid data points)
06B Segment-ology: Segment Size vs Cousinship Chart Needed; Jim Bartlett 20151106