While that is true, the tradeoff there involves throwing 1,000 state of the art at the enemy, vs. throwing 5,000 cheap and adequate. (not sure if the multiplier is 5x but something like that) NASA does not have exactly the same situation.

There was a time when they were going to try to do more, cheaper, faster missions, which IIRC lasted until the first one failed, and then that policy was abandoned.

What if part of the experiment was to test the viability of COTS parts?

They used off-the-shelf components because they were low priority, low cost side projects that were not expected to last long anyway.

Similarly, the video (and associated audio which malfunctioned) of the skycrane-deployed landing was done by COTS components which otherwise would not have been present at all.

Also, by COTS, they're using components which would otherwise be meant for industrial use, so they're still likely to be more rugged than what might come to mind when thinking of off the shelf components.