[KrazyCarl92]

Imagine a robot that grabs two containers off the step every single time in autonomous faster than any other team in the world (or division), then does nothing else the rest of the match.

A number one seeded team at world champs is likely putting up 2-4 full stacks on their own every single match. I would guess that a team capable of putting up 2 full stacks could potentially be around as a 24th selection, and the 25th selection will be just about as effective.

There is a tremendous incentive for the number 1 seed to select the "grabs containers-does nothing" robot because they can probably still put up 5 or 6 full stacks as an alliance even with a "do nothing in teleop" robot because they can put up 2-4 stacks and the other alliance partner can put up 2 stacks. And with 4 team alliances, they can sub in the really fast auto RC grabber when they need them and sub the team out when they can safely get the cans otherwise and could use a "contribute something in teleop" robot with a slightly slower (but still sufficiently fast to win the race) step grabber.

Now consider that any team with the resources to be the fastest team in the world at grabbing RC's from the step in auto is likely to also be able to do something during teleop to contribute to the alliance. Now change your "grabs containers-does nothing robot" to a "grabs containers-holds their own" robot, and you can see the scales are tipped even more in their favor.

Now the assumptions about the abilities of the "200th best team" at the world championship and there being an auto RC grabber which beats all other RC grabbers in the division cut and dry are probably inaccurate to certain degrees. But I still cannot see how this singular trait would not overwhelm all other considerations for alliance selection criteria given that there are two more members of the alliance that could be effective robots.