Cycle Time Predictions At Kickoff, a Mostly Murky Look

With kickoff tomorrow, I was thinking about a recent conversation I had with another mentor. It concerned predictions for cycle times that we make at our kickoff meetings, and what reality produces. Did we predict that teams would get up to 10 cycles? 15? “Maybe not at week 1, but at champs we’ll definitely see 15 cycles!”

I did a lot of week-to-week number crunching, made graphs, and I found out… cycle times go down.

image
[here blue is quals, and orange is playoffs]

Duh.
I already knew that in my gut, but it’s nice seeing it on paper. So now the real question: Does cycle time go down predictably as the season progresses?

Since games vary year-to-year, looking at exact numbers of seconds isn’t going to help. I wanted to look at the cycle times each week in comparison to Week 1. I also didn’t care about auto or endgame scores as there is usually some protection or no defense. So I pulled some numbers from The Blue Alliance (someone more savvy may have found an API tool for this, but I just clicked and scrolled):

My (very imperfect) calculation:
T = average teleop game piece points per match
P = average points per game piece
Cycle Time = (135s) / (T/P)

The same downward trend for cycle times in the last three years.

What about each week as a percent of week 1?


It looks like there is a pretty consistent drop to about 65-75% cycle times in quals before District Champs start. Then it drops to 50% or lower at Worlds.
And the trend is a lot more similar for playoffs than I would have thought.

Implications?
Double your kickoff predictions!
I think we (my team) always makes predictions for cycle times that are very optimistic. Usually they are based on a robot moving unobstructed, without dropping anything, and no user errors. You know, like world champion teams. I feel pretty comfortable, based on these data, to take those predictions and double them (at least) for the average week 1 match.

Even if you think YOUR team is going to be a short cycle superstar, remember that you may be allied with two teams who don’t score (much). And it’s the alliance score that counts. :grin:

This also varies with game type. The 2022 field was much more flexible in terms of scoring. In 2019 and 2023 the scoring was less flexible and there were multiple game pieces.

My numbers and stuff
Google Sheet
Feel free to play with the numbers. Or give yourself a headache and try to add 2018 or 2017!

46 Likes

I absolutely love this analysis! Thanks so much for sharing, it’s very cool to see the impact of the iterative process in action, across all of FRC.

Best,
-Mike

7 Likes

Even closer in post-bag years than before. I wonder if this represents the effect of playoffs teams pre-bag being more likely to have a second bot to practice with.

Thanks. My gut also felt this way. Even knowing this, I struggle with how to design and plan. If I know my teams bot will be a slow cycler then I would have the team focus on defense. (They never want to focus on defense.) If I think we can cycle fast, then I may be just fooling myself as most of us seem to do. I plan to double check my optimism with these numbers.

I should look at 2013 cycle times to compare to this new game!
Edit: never mind, that data doesn’t exist on TBA.

How does it compare if you use teams predictions at kickoff on CD?

1 Like

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.