Public Statistics

[DankNuggets]

:thumbsup: :w00t: :groovy:

was worth the wait!

[Maverick]

The 23rd Scorereset period just closed!
TABARNAK!!! has received alot of achievements again. Someone should tell him

[Static Burn]

[B]The 47th Scorereset period

Score resets happen every two weeks, meaning there's only about 26 each year, so this can't possibly be the 47th of 2008

Nice job though

[Maverick]

Heh I counted the weeks wrong when I tried to establish the first scorereset period number. Thanks for the comment.
I've fixed the scorereset period numbers, they should be correct now.

[Soup du Jour]

I'm looking at the statistics for the 23th scorereset period of 2008 one thing that confuses me. How did TABARNAK!! get more wins and losses in the shark section than he did in the overall section?

Edit: Similarly for TABARNAK!! and the points columns.

Also, maybe there's a mathematical reason for why this is happening but shouldn't the winner of highest overall rating also win highest rating in one of the ships (since the highest overall is higher than any individual).

[Maverick]

How did TABARNAK!! get more wins and losses in the shark section than he did in the overall section?

I can't really explain why. It's as if the bot has seen him make more kills in the shark then he really did.

Edit: Similarly for TABARNAK!! and the points columns.

Point difference can be explained; the bots don't know how many points you gain on each kill, so they estimate the number of points by the bounty of the killed player.
After a short while the server updates the bot with the actual number of overall wins/losses/points. Since the ship statistics is cumulative from the overall scores, the bot can't modify the ship statistics just from the score update. Instead the bot tries to keep the difference non-negative (actual scores vs ship scores) by skipping points.

Also, maybe there's a mathematical reason for why this is happening but shouldn't the winner of highest overall rating also win highest rating in one of the ships (since the highest overall is higher than any individual).

Not necessarily, a player may have scores from a number of ships where overall he has achieved a high number of kill points and wins while on each ship individually he hasn't achieved a record in kill points and wins. (see my 2nd post for the rating formula)

[bounty]

Great work on this Mav and others involved! Amazing!> all the cool features and such must have been alot of work. Thanks for your love and dedication to our game.

[Kontrolz]

A bit of information about the calculation of Rating and Average (if you didn't know yet):

Rating
Rating is calculated as follows:
Rating = (kill points * 10 + (wins - losses) * 100) / (wins + 100)


Averge
Average is calculated as follows:
Average = (kill points) / kills

Average is basically the average number of points you get from a kill.

Are you using the correct formula for rating? The formula the game uses is:
R = (kill points x 10 + (wins - losses) x 100) / (wins + 100) kills

You posted:
R = (kill points * 10 + (wins - losses) * 100) / (wins + 100)

You missed the additional multiplication at the end. The reason I ask is because I've noticed that the ratings on the site and the ratings in the game are nowhere NEAR each other. I wasn't sure if this was a genuine mistake, or a change to the formula just for the site, or just a typo. But again, the ratings on the site are not even close to the ratings in the game and they seem unfair.

[Maverick]

I've noticed the same but couldn't find the correct formula anywhere, so I assumed that the formula I have was the correct one. I've also done a few tests in-game, comparing the rating result in-game and the result of my formula. They did match.

Where did you find another formula where you need to add "x wins" at the end?

Notice that in-game you can't see the killpoints. You only see the total points which is the sum of flagpoints and killpoints.

I'm noticing now that the number of wins / losses in-game differs from the public statistics while it shouldn't differ that much. I will investigate this issue.

[Viruk]

Great work on the public information.

But why is it also displaying PRIVATE information?

What if I don't want people making assumptions about me based on nationality and usage? Do I want people deliberately lurking outside my normal resolution and taking pot shots at me.

It should not be displaying any player-specific info that a normal player couldn't find out through bot commands or looking at scores.

[Maverick]

It's not showing any private information. The information showed is already on public display in most online games like Counter Strike or other Steam titles, Gunbound, Guild Wars, World of Warcraft, and others.

[Kontrolz]

I had the formula in a ?log

Kontrolz> what is the formula for rating anyway?
BH822> let me check
Kontrolz> ok
BH822> R = (points from kills x 10 + (wins - losses) x 100) / (wins + 100) kills

I don't know where bh got the formula but that's what I've thought it was for years now. I did the math though and it doesn't seem right with extra multiplication of kills. Still though, something is off with the statistics because they are nowhere near in-game statistics.

[Raver420]

[roxxkatt]

yeah, i saw a shark with like 100-400 rec get #1 in rating for this period...




also, when are you going to fix the horizontal graphs, like resolutions, banners,

and when are you going to add numbers to the highest usages page?

[Maverick]

also, when are you going to fix the horizontal graphs, like resolutions, banners,

Not going to be done as it's not of any importance.

and when are you going to add numbers to the highest usages page?

See above