AWS'60 Results Analysis

All things antweight

Moderators: BeligerAnt, petec, administrator

User avatar
jes
Posts: 36
Joined: Fri Jan 03, 2020 12:16 pm
Location: Bristol
Contact:

AWS'60 Results Analysis

Post by jes »

I wasn't sure what type of robot would be best to design, so today I watched the AWS'60 video ( https://www.youtube.com/watch?v=Yd6b21LqA3M ) and recorded the result of each fight, with the type of robot that won, type of robot that lost, and reason for the result.

Here's the data in case it's useful to anyone. The error bars show the upper and lower bound of the Wilson confidence interval at a 95% threshold, i.e. we're 95% sure the true value lies within the error bar (unless I misunderstand). Interestingly, every single error bar includes the 50% level, so this data would not disprove the hypothesis that every single fight is decided at random.

The datapoints with large error bars come from a very small number of fights, so they can be taken with a pinch of salt. I had to manually label each fight so it's possible I got some of the results or robot types wrong. I also know that I missed a couple out because it wasn't clear to me which one had won. In total I recorded 223 fight results.

This shows the win percentage for each weapon type I identified:

Image

And this shows the proportion of fights that ended with different outcomes:

Image

"Pitted by self" includes accidentally driving yourself into the pit and being flung into the pit by energy from your own weapon. "Pitted by spinner" always means you were flung into the pit by your opponent's spinner.

Just from watching videos of fights, I would have said that the spinners are a totally bad idea due to their propensity to throw themselves into the pit, but from the data it looks like the vertical spinners at least might perform better than average.
User avatar
MarkR
Posts: 375
Joined: Mon Dec 18, 2017 12:46 pm
Location: Reading Hackspace
Contact:

Re: AWS'60 Results Analysis

Post by MarkR »

Excellent, my decision to build an Axe bot is vindicated, as Axes seem to fair worse than "nothing" :)

Mark
Robots: Betsie - RaspberryPi controlled flipper bot with gyro stablisation - too clever for her own good?
Stacie - tidy flipper; 4wd driven by hair bands
User avatar
EddieJ
Posts: 126
Joined: Sun Apr 08, 2018 2:11 pm
Location: Dorset

Re: AWS'60 Results Analysis

Post by EddieJ »

MarkR wrote: Fri Jan 10, 2020 9:45 am Excellent, my decision to build an Axe bot is vindicated, as Axes seem to fair worse than "nothing" :)

Mark
Not just that! They are more CONSISTENT at being worse, than nothing :shock:
You need to bring that average up! ;)
#MakeAxesGreatAgain

Great stats! Do you plan to sample data from the other 59 AWSes also? :P
Bots:
DisinfectANT (Antweight), Cilit BANG (WIP Antweight), Kinetic Disassembly (WIP Antweight)
User avatar
MarkR
Posts: 375
Joined: Mon Dec 18, 2017 12:46 pm
Location: Reading Hackspace
Contact:

Re: AWS'60 Results Analysis

Post by MarkR »

I suspect the "Vertical spinners perform better than average" is mostly because they're generally built by the more competent builders / drivers. If I made one, it would suck.
Robots: Betsie - RaspberryPi controlled flipper bot with gyro stablisation - too clever for her own good?
Stacie - tidy flipper; 4wd driven by hair bands
User avatar
LimaHotel
Posts: 258
Joined: Fri Mar 16, 2018 10:30 am
Location: West Devon

Re: AWS'60 Results Analysis

Post by LimaHotel »

wow WOW AAAAAA STOP THINKING YOU'RE STRESSING ME OUT!

I kid, that's some dedicated analysis! But seriously - I've never so much as even sketched a robot before cutting my first piece of polycarb! If you come up with some super statistically justified the numbers are with me bot, then I look forwards to facing it in the arena!

Also - watch the prior AWS's. Not all obviously, but we've had a good number of different winners lately. AWS58 being something of a real surprise! It might adjust your statistics a nats

We also have a few different arenas that get used for AWS, and the arena can have a real impact on what kinda design is more competative. Pete's arena is relatively large without being the biggest, and 2 of the 4 drop offs are walled. This makes it significantly friendlier to spinners than an arena like the Ant Freeze or RoboShowdown one. It also gives Flippers a bit more of an advantage over Pushers. Meanwhile (AAAAA names please I'm so sorry)'s arena that we used in AWS58 has the drop offs on the end walls, not on the corners. It's also got the largest floor space of any of the arenas I've fought in at any rate. CHRIS! It was Chris. Right? Chris hosted AWS58?

As an aside, welcome! Are you gonna be at Ant Freeze?
And build a Grabber! Then I can challenge you to a whiteboard with Jiggy! :p

You're also missing one VERY important weapon categories there: Thwackbots :p. Okay, so maybe they aren't super influential these days :lol: . Also whether the bot was 4WD or 2WD or a walker. In my opinion, the only weapon distinction that matters is "Spinner" or "Non-Spinner". Beyond that it all comes down to construction quality, drive, and driver; with Weapon being the least important of those 4 categories :p
A grabber? I CHALLENGE IT WITH JIGGY!
User avatar
LimaHotel
Posts: 258
Joined: Fri Mar 16, 2018 10:30 am
Location: West Devon

Re: AWS'60 Results Analysis

Post by LimaHotel »

MarkR wrote: Fri Jan 10, 2020 10:43 am If I made one, it would suck.
I don't believe that for a second!
Now MY spinners are gonna suck. Mostly because I'm insisting on using parts I found in skips (BLDCs from old CD drives, the bar itself is gonna be kitchen knife steel, etc), but also because I'm planning to have the spinner off-center (He Span It Left Hand, innit) which will just make the gyro forces utterly intolerable and I'M OKAY WITH THIS
A grabber? I CHALLENGE IT WITH JIGGY!
User avatar
Kyro
Posts: 412
Joined: Fri Jan 18, 2019 12:36 am

Re: AWS'60 Results Analysis

Post by Kyro »

to me, he span it left hand means anti-clockwise on a horizontal spinner...

if you are going for a vert, just have a 100rpm motor on the left and a 1000rpm motor on the right... there will be plenty of left hand spinning ;)
Team Rocket
Trappist 1(4wd grab 'n' lift)
Ton 618 (4wd expanding bot)
Io(4wd flipper)
User avatar
jes
Posts: 36
Joined: Fri Jan 03, 2020 12:16 pm
Location: Bristol
Contact:

Re: AWS'60 Results Analysis

Post by jes »

EddieJ wrote: Fri Jan 10, 2020 10:35 am Great stats! Do you plan to sample data from the other 59 AWSes also? :P
Hmm, taking a look at AWS'59 might be worth doing just to see how consistent the results are. If they both agree on the rough ordering of the different types, we can have more confidence in it.

Don't think I can quite be bothered to do all 60!
MarkR wrote: Fri Jan 10, 2020 10:43 am I suspect the "Vertical spinners perform better than average" is mostly because they're generally built by the more competent builders / drivers.
Good point. I think this is absolutely right. I can't think of a good way to factor this out unfortunately. Is there a series of fights for newbies only or something that we could look at?
LimaHotel wrote: Fri Jan 10, 2020 10:59 am We also have a few different arenas that get used for AWS, and the arena can have a real impact on what kinda design is more competative.
This is another good point I hadn't considered!
LimaHotel wrote: Fri Jan 10, 2020 10:59 am As an aside, welcome! Are you gonna be at Ant Freeze?
It's unlikely. Hinckley is quite a long way for me and I probably won't have built anything I'm happy with by then. I was planning to start with the BBB Antweight event in April.
LimaHotel wrote: Fri Jan 10, 2020 10:59 am Also whether the bot was 4WD or 2WD or a walker. In my opinion, the only weapon distinction that matters is "Spinner" or "Non-Spinner". Beyond that it all comes down to construction quality, drive, and driver; with Weapon being the least important of those 4 categories :p
Fair points. These are much harder to classify just by watching the video unfortunately :)
MySolderIsOlder
Posts: 190
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent

Re: AWS'60 Results Analysis

Post by MySolderIsOlder »

Proud to have been 50% of the flail spinners at AWS60, though I only won one fight IIRC - and that was mainly through pushing.
Stuart (Tony's dad)
MySolderIsOlder
Posts: 190
Joined: Mon Jul 30, 2018 1:22 pm
Location: Kent

Re: AWS'60 Results Analysis

Post by MySolderIsOlder »

But yes - the one thing I've learnt in my limited time roboteering is that driver skill is far more important than weapon design. I should probably spend less time tinkering and more time driving, though i doubt it would do me any good!
Stuart (Tony's dad)
Post Reply