Replies to post #211534 on Biotech Values

[biomaven0]

>>If the pairing is random, how can it also be one that minimizes cumulative M?

I should have been more precise. My assumption is that there are many possible sets of pairings that will produce around about the same cumulative M and it is likely computationally impractical to find the one that minimizes cumulative M. So instead randomly choose a subject, match that subject with the closest possible pair, and then randomly choose another subject and match with the closest possible remaining pair. If two possible matches are close enough you can choose randomly between them.

It might seem like cheating, but I suspect you can repeat this process several times and mash all the results together, thus reducing the chances that you got some fluke matching.

Peter