Ball map averages for Palin, Clapon, Masty, UFO, and Goose

[Sarah Palin]

I wanted to look at players who usually fly just one kind of plane, and see how they compare across maps.

Here are their lifetime ball ladder ratings. Higher means more likely to win that map.



GOOSE usually plays explodet. He has the lowest standard deviation of these players, that is, his win chance is the LEAST affected by what map is being played. His performance is consistent across all maps.

MANA usually plays miranda. His win chances are the MOST affected by the map chosen. Funnelpark and Lost City are maps where he excels; whereas he is the worst of the group in Darkwar and Cross.

UFO usually plays miranda. Interestingly, despite playing the same plane as Mana, their scores don't line up at all.

Palin usually plays acid loopy. I am the second most map-dependent player in this group after Mana. Open maps like Grotto and Labyrinth are terrible for me, while I excel at Darkwar and Asteroids.

Clapon usually plays bomber. He is the second most-consistent player after Goose; his worst map is Grotto and his best is Cross.

[Sarah Palin]

Who should I add to this chart? Remember, I'm looking for players who almost always fly the same plane.

[argonide]

UFO usually plays miranda. Interestingly, despite playing the same plane as Mana, their scores don't line up at all.

Lol

Picking people with the most games played might help. Also I know apathy always plays explodet.

[nobodyhome]

fun stuff. as an aside, if you know some programming, i can give you a data dump of the tables if you want and you can perform some more of this kind of stuff if you want.

and here's my other suggested players:

miranda:
ssd
axelrose
xx2

explodet:
ryebone

bomber:
nikon

biplane:
sunaku
ojm

loopy:
brillo

[CCN]

oh ****, I rememeber getting these dumps.


><

[Sarah Palin]

Originally Posted by nobodyhome

fun stuff. as an aside, if you know some programming, i can give you a data dump of the tables if you want and you can perform some more of this kind of stuff if you want.

I'm doing all of this in Excel - I just looked at my most recent match on Cave, then looked at my Map Avg rating, then did the same for my most recent match on each other map etc.

[andy]

very nice idea, but i think you should use players with approximately the same number of ball matches, and i suggest close ratings (all top 50). Also be sure to pick players that always play the same plane.

suggested list:
Loopy: Brillo and tgleaf (check out acid and df difference)
Biplane: sunaku/esoteric
Bomber: Nikon
Explodet: ryebone
Miranda: Tominator

everyone on this list has approximately 300 matches played (thats why i didnt include apathy, goose and clapon) the only one is sunaku with more than 600 but the other bip on the top 50 (eso) has 200 games played.

[Kuja900]

Im fairly sure ryebone frequently uses other planes, as does eso. Apathy would be a good choice.

[Evan20000]

I'm honestly curious how a TBD version of this would look.

[andy]

Originally Posted by Kuja900

Im fairly sure ryebone frequently uses other planes, as does eso. Apathy would be a good choice.

i think rye uses only explo in ball, at least i always see him use that, apa has more than 1000 matches played.

[Sarah Palin]

The number of matches shouldn't matter as long as the player has enough (200+) that their average on each map has shaken out.

The shortlist:

LOOPY
Palin
Brillo
Tgleaf

BOMBER
Clapon
Nikon

BIPLANE
Sunaku
OJM
Esoteric

MIRANDA
Mana
UFO
SSD
Hmmmm

(I'm fairly sure I've seen Axl and XX2 using Loopy often.)

EXPLODET
Goose
Ryebone
Apa

[nesnl]

Use a histogram. A line graph can be misleading because if you aren't looking at it the right way you will think there is a relationship between the different scores (due to the line and the slope). But the line means nothing.

[drunkguava]

don't use ssd. he switches between randa and explo depending on the map and his team's composition.

[ryebone]

Actually I have over 500 games played. Heh.

Since there seems to be some contention, I do solely use explodet in ball ladder. The only times I've switched to loopy is if my ENTIRE team is too dumbassed to switch to light planes, but that's happened maybe 5 times maximum.

[Sarah Palin]

(SSD removed from shortlist)

Keep suggesting players, I have this chart now in a format where it's easy to add players.

[mlopes]

Awwwwww, no biplane

[Sarah Palin]

Here is the chart with all the players nominated so far.



As you can see, Loopy is the most map-dependent plane and Explodet is the least. I would like to get more Loopy data.

For Loopy: Cave, Darkwar, Mayhem and Asteroids are excellent maps, while Funnelpark, Labyrinth, Core and Maze are apparently equally bad.

For Miranda, Grotto and Mayhem are their best maps, with everything else being pretty average.

For Explodet, the map almost doesn't matter, although Grotto and Maze stand out with their tight goal areas.

For Biplane, Cross and Grotto are the best maps and Snow and Darkwar are the worst.

[drunkguava]

spark - loopy
don't follow - explodet
anghell - bomber
tex - explo
val - loopy
daxtor - randa
d4rt - loopy

[andy]

pick people high on ladder or you get very low map averages.

[nostrich]

I'm just going to leave this here. Graphs and charts are kind of my "thing" so I couldn't resist. As you can see, there's not really enough data for it to be useful, but it's pretty!

I couldn't be bothered to label it, since I'm just kicking the tires, but down the side you've got Loopy, Randa, Bomber, Biplane, Explodet, and along the top is the different maps, same as Sarah Palin's x-axis. The bigger and bluer the dot is, the more likely it is that plane will do well on that map, the bigger and redder it is, the more likely it will do terribly. Blank spaces are even odds (ie, +/-0.)

I'd say this -- and Sarah Palin's charts -- would be much more interesting with a fuller dataset -- I don't really believe the Loopy's performance is that much more variable than the other planes. I'm kind of curious about what nobodyhome can get his hands on.

Also, great idea Sarah Palin.

[poofighter24]

using me for this would definitely not be a good idea, lol

for the first 400 games or so I was basically exclusively TA randa, then switched to 50-50 miranda-explo until like game 1000, and currently am running at about 20-80 miranda-explo (hence the vast increase in rating). waaaay too streaky to get good single-plane results.

for miranda, I would recommend armada x since all he plays is an utterly useless miranda, as I would like to see which maps his complete lack of defense and kill support can be made up for by his being able to warp more than mirandas that actually bother to shoot people.

brillo for loopy, eso and hurripilot for biplane. sunaku plays too much whale to get accurate biplane metrics imo.

[Jayfourke]

I always use DF-loopy, if that helps.

[gemigemi]

I always play whale in ladder (less than 5 games with something else).

[Sunaku]

I actually switch between bip and explo depending on the composition. So stop looking at my stats, perverts.

[drunkguava]

Originally Posted by gemigemi

I always play whale in ladder (less than 5 games with something else).

you play ladder?

[Sarah Palin]

Originally Posted by andy

pick people high on ladder or you get very low map averages.

I already accounted for this; on each map, each plane's average is divided by the plane's average for all maps. So if I added a bunch of bad Explodets to the stats it won't pull all the Explo numbers down.

This is ok since we're not comparing Explodets to Loopies, we're looking for data about "Explodet is better on Map X than Map Y."

The assumption is even adding a bad Explodet to the stats is useful because even a bad whale does better on a good-whale-map than a bad-whale-map.

I'd say this -- and Sarah Palin's charts -- would be much more interesting with a fuller dataset

Great chart nostrich!

A better dataset would be nice, if ladder could track what planes people play.

I know this is possible because the people who run COOP were able to do it:

http://alticoop.erasum.com/index.php?display=stats

If ladder initiated this kind of tracking, within a week we would have a massivvvve pile of data to play with

We would also be able to do things like "A team with one Loopy wins Cave x% of the time, but when there are two Loopies they win y% of the time."

I don't really believe the Loopy's performance is that much more variable than the other planes.

Well we can make a friendly bet about it. I have 4 more Loopy nominees to add to the three already in the chart. Will add them tonight and see how the numbers move around...

[nobodyhome]

Tracking what planes people play is somewhat different on ladder than it is on coop. In coop it's 1 life only, so you can say that "in this run of coop, so and so player played biplane". Whereas in ladder, you can respawn with different planes several times, so you can't really say that "this game was a game in which so and so player played explodet" because that changes throughout the game.

That said the basic idea of tracking what planes people spawn as CAN be done, and this is part of the feature set we are working on for next ladder season (which, again, will be out when it's ready heh).

[Sarah Palin]

Originally Posted by nobodyhome

That said the basic idea of tracking what planes people spawn as CAN be done, and this is part of the feature set we are working on for next ladder season (which, again, will be out when it's ready heh).

Awesummmm

[Sarah Palin]

Three more loopies added to the data.

Loopy still looks pretty swingy considering there are twice as many more data points as any other plane:



But have a look at the individual data. I don't see a pattern, do you?



Here is the same chart but with everybody normalized so we each average zero:



Still no pattern that I can see (other than Cave being an awesome Loopy map, and that I ROCK TEH DARKWARZ).

It seems individual skill level and map familiarity trumps the map's effect on your win %. There is no map that is uniformly bad for us 6 loopies except - weirdly enough - Core.

[banana]

I've played explodet for 99% of my ladder matches if you're tracking explodets. I'm Flyingbanana btw.

[Evan20000]

You could add me for Miranda, although I don't play like most Mirandas in ball....

[Greekjr14]

Me biplane plz

[blln4lyf]

I tend to play miranda nearly all of the time as well