Quote:
Originally Posted by Siri
Back in 2010 (our first year of swerve drive), if the goal was to perform well on the field that year, our complexity-based failures probably "cost you more than the added features could have possibly given you over a simpler design". But I doubt you could find anyone wouldn't do swerve that year if given another chance.
|
This is what offseason is for, though. If you have any doubts at all about your ability to implement a difficult drive system reliably, then the proper time to experiment with it is when there are neither strict time constraints nor serious costs to failure. Only once you have enough experience within working team memory to do it in a way which does not compromise reliability should you put it on the table for build season.
Quote:
|
This year, ok, we've had a couple in-match failures, maybe one even cost us a match. But I seriously doubt we would have been on Einstein without the swerve--it was just so integral to our strategy/alliance. There were of course other strategies which were very successful (and 6 that were more), but I doubt we could have implemented them to better effect than the one we chose, in part building off that under-performance in 2010. In short, there are big-risk-big-reward drive features that really are worth it, even if there's a risk of "if at any point it fails during a match". It's just that in some cases, you have to be willing walk the longer arc of history.
|
As far as my design philosophy goes, if your "big-risk-big-reward" drive truly qualifies as "big risk" (for any reasonable definition of "big"), then you probably shouldn't do it. Drive is far too crucial to baseline ability to play the game to be gambling with. From what you describe, it sounds like you now have enough experience with swerve that the risk is not significantly above what less-capable teams would experience with a much more trivial drive system.
"If at any point" was intended heuristically and is hyperbole, and perhaps I should soften it: the loss of drive ability in a match is a crippling blow, and sacrificing anything other than very small increases in its probability for added functionality is very likely going to have negative utility. For the vast majority of situations, "do not sacrifice reliability for features" is going to give you a reasonably optimized decision.
Re: intentional underbuilding, bumper supports are one thing, and the actual drive is another; I'm not sure I'd personally be comfortable with bumper supports that I didn't know would stand up to FRC impacts, but I could understand the justification for doing so. But I am very sure I would never put anything in the drive train if I doubted that it would last.