Aerial Assist Score Simulator

[BigJ]

Hi all,

We've spent the last 2 days indoors here in southeastern Wisconsin, so I thought I'd share some of my work. I threw this together while the team was discussing how to play the game this weekend.

https://docs.google.com/spreadsheet/...VE&usp=sharing

It is a teleoperated score simulator. The base assumption of the sheet in general is that cycles are starting at the buzzer between auton and teleop. It is read-only but you should be able to make a copy of it and edit it to your heart's content.

The cells you want to modify are the average times for each task and the percentage chance of the scoring tasks happening. (Use logic for the % chances and your estimations, as some of the existing "cycle types" logic can fall apart if you aren't realistic)

Extending the spreadsheet is pretty easy. If your team has a different idea of how to divide the timing of different tasks, you can add or modify those rows. If you have another cycle style, you can add a row down there, and the only formulas you should have to manually do are the total time and points/cycle.

The sheet shows both an average calculation of cycles and points, as well as one with the number of cycles rounded down.

The sheet also has an area that automatically sorts cycle types based on their performance from the numbers you enter.

The major issue I can currently see on the sheet is that it does not count truss/catch points earned "before the buzzer" but not before the end of a cycle. It should still be a decent estimation, though.

It helped us a lot during brainstorming and strategy discussion, and we are excited about the possibility of even using it at competition, if we can more accurately estimate task times, to see if certain cycle types might be more or less viable depending on the alliance makeup.

Feel free to ask me any questions about the sheet.

I don't intend for this topic to be a debate about the current numbers in the sheet, so please take those as example numbers.

[Strategist257]

I think that i've noticed a mistake in the points per cycle. in row 31 column C the total number of points for that cycle should be 51, not 46. Actually now that i look at it, a lot of the points per cycle are wrong. don't forget that the points are 10 points for the high goal, 1 points for the low goal, 10 points for 2 assists, 30 points for 3 assists, 10 points for the truss and 10 points for the catch

and what is the 1 assist truss high mult in cell 20a?

[BigJ]

Originally Posted by Strategist257

I think that i've noticed a mistake in the points per cycle. in row 31 column C the total number of points for that cycle should be 51, not 46. Actually now that i look at it, a lot of the points per cycle are wrong. don't forget that the points are 10 points for the high goal, 1 points for the low goal, 10 points for 2 assists, 30 points for 3 assists, 10 points for the truss and 10 points for the catch

and what is the 1 assist truss high mult in cell 20a?

The current spreadsheet has Catch at 50% success rate -- that's why those average cycle scores are 5 less.

The 1 assist cycles are imagines as being one robot doing repeatedly, which each have a time penalty for driving back to recieve a ball compared to 1 truss high mult, which assumes that another robot is ready and waiting for a new ball to be introduced.

[Strategist257]

ok, thanks, i forgot about the 50% catching.

[Bill_B]

Downloaded to open office form and lost the auto sorting of the scoring leaderboards. not a big deal, I guess the xlsx has a macro that does it and it would have been blocked anyway.

Good to have the probability constants table that reflects how likely the points are to be scored. Trouble comes on the timing of the failed shots. We need variables that estimate the time to recover from missed shots both high and low. Then some formula magic is needed that can incorporate those times into the various cycle-time sums. (reciprocal of the probabilities?) It's been well calculated that the max cycle score is 60, but without any ideas of how likely that score is to happen or time required to do it, the 60 is a curiosity only. Something akin to how much you could win on a scratch lottery card.

Thanks for the program though. I will keep busy for a while with it and it will keep my mind off our team's crazy design decisions in the process.

[BigJ]

OpenOffice might have an equivalent sorting function. There is definitely a problem when trying to do shot miss rate -- could be solved by adding a new category with a time estimate for retrying, and adding that time multiplied by (1 - P(goal scored)) to the average cycle time.