Something interesting about this year’s skills challenge is the way your score increases as you score more balls. From 1-5 balls interacted with, the score skyrockets linearly with the number of blue rows blocked off. From 5-10 balls, the score increases polynomially, as you begin to score red rows. Finally, from 10-30 balls interacted, the score gradually increases from 106 to 126
The graph below shows this trend, with trendlines added for visual support:
Here’s the interesting part: If you define the complexity of a skills run based on how many balls the robot interacts with, then a 126 point skills is three times more complex than a 106 point skills, with only a slight increase in score.
Wow. I’m glad someone took the time to do this analysis.
Be interesting to see how alliances are chosen based off of skills rankings. Not a huge difference between moderate to complex hood bots, and much simpler ones.
Prolly gonna be three tiers of scores in that case…
Thanks for doing this!
it shows basically how if you don’t have time to do anything, it’s most important to just put one ball in every goal, in order to kill all the blue rows and form all the red rows.
Yes. Don’t forget that you also have to descore one from the middle tower though. Also some robots may be able to descore a blue ball and score a red ball at the same time.
I think that this mainly shows that movement is crucial. This sticks with the theme of previous autonomous skills portions, but this is marginally more unique for driver control. Going to 6 different goals will make a higher score, but the complexity of the autonomous makes it a more difficult proposition than scoring fully in 3 goals. I think this means you focus on movement algorithms and try to get around the field as much as you can, rather than cycling balls.
Pushbots do weird things every game. In this game, a pushbot with a long pointy stick could theoretically get 40% of all possible points in programming just by driving back and forth, knocking out the balls in the center tower.
Besides that, I don’t think that driver skills is going to be a fair metric at all at judging efficiency of a robot, since all robots should be able to at least get to every tower within the minute at this point, since after 105, assuming they’re a hoodbot, it doesn’t really matter that much. However, driver skills will become a fair metric again if people finally max it out and rely on the Skills Time tiebreaker, because that will more clearly show efficiency.
