Galcon 2: 49 players online!
:: Community ::
ForumsMessagesModsChat
 

Forums :: [OLD] iPhone :: Strategy :: Official Stat Warriors Leader Board

You must sign in to post.

Page 5 / 22 1   2   3   4   5   6   7   8   9   10   11   12   13   14   15   16   17   18   19   20   21   22 

Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 4:48am

Cabin Boy zmar01

Joined: Jun 6, 2009
Posts: 656

(2v2)
I still dont like how players like:
jas
fart force
seichewarn
goozok
are in the top 30 while players like

matty
56
periwink
dejo
toomacho

are all lower than rank 70

It seems like you base too much of the rank on how many games you play against ranks, rather than how much of the games they win. (win%)



Mastery = 100 * (Number of Opponents * Games Played) / Total Game Count At This Rank

Seems like the less games you play at a lower rank and the more you play higher rank the higher score. Nothing to do with any winning.
and your score goes down when people play 1 game and leave

Field = Opponents Played Against / Total Opponents Possible

whats the total opponents possible got to do with rank?

there are alot less people playing in the hours of about 8:00-14:00 gmt

If thats the only time you can get on then your score is going to be less.

So really, a higher Field rating is MUCH better to have. 


Seems a bit unfair.


Skill = Field * (Mastery / 100) * (Win Percent / 100)

field and mastery which seem like weird things to base skill on


Score = Win Percent * Win Percent * Rank * Rank * Skill 


Why are you squaring rank? and what rank is that? phil's rank on your profile?

Isnt that just like saying that if you play 200 games and get to GA thats better than someone who has played 200 games over a longer period but not got better than 1 stripe.



I personally dont like the whole system, im not saying i could make better as i cant program. And i dont even know if i can make a better ranking system.

the results for your system look alright mostly.

1v1 ranks look nice.

And i respect the work you have done, but i dont like the formula.
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 4:56am

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

I think I understand why the 4v4 and 3v3 ranks look so different then we are used to.

They don't contain any 1v1 games in them.

Traditionally, 1v1 games are included in the FFA game ranks.  I believe this might be the first time the 1v1 games are excluded from them instead.

I think this is why currently we see sir_earl and bbsmylie2 above Fifty6 and im.a.girl.  

Because the players we normally see up high in ranks in the FFA charts here at galcon.com include those names we currently see in JUST the new 1v1 Leader Board....  and those same people, at least for the last month, haven't played a whole lot of 4v4 games.  Which leaves them out of the 4v4 leader board.

I think as time goes on and games are played, the vast difference in scores between bbsmylie2, sir_earl and those players we know to be better skilled, will eventually even out to a large degree.  This is because those better skilled players will be able to maintain their higher winning percentage and still accumulate, over time, a larger Field rating.  While on the flip side, the lower skilled players will be unable to raise their winning percentage due to the quantities of games played, which makes it harder and harder to USE their higher Field numbers to their advantage.

So all you highly skilled 1v1'ers out there...  play against each other with 4 total players in a room instead and show these people that quality does in fact trump quantity.  :)
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 5:07am

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

Thanks for the feedback Zmar.  I really do appreciate it.

(2v2)
I still dont like how players like:
jas
fart force
seichewarn
goozok
are in the top 30 while players like

matty
56
periwink
dejo
toomacho

are all lower than rank 70

It seems like you base too much of the rank on how many games you play against ranks, rather than how much of the games they win. (win%)



I agree.  

But I'm not sure of thats due to the limited sample size (less then 1 month) or if thats due to an error in the formula.  If the sample size were 6 months, for example, the Field numbers for everyone will have evened out a bit I think.

Also, once I convert this ranking to use its internal scoring rather then basing it off Phils in-game ranks, I think the numbers will shift a bit away from quantity played towards quality played.
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 5:29am

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician


Mastery = 100 * (Number of Opponents * Games Played) / Total Game Count At This Rank

Seems like the less games you play at a lower rank and the more you play higher rank the higher score. Nothing to do with any winning.
and your score goes down when people play 1 game and leave


Correct. Whats wrong with that?  Mastery % isn't about who won, its about who and how much you played against.  Who actually won is calculated later.


Field = Opponents Played Against / Total Opponents Possible

whats the total opponents possible got to do with rank?


a Ton!  It determines the sample size you are dealing with.  A player who can beat 6 people out of 10 may not be able to beat 60 out of 100 or 600 out of 1000.  The size of the sample set is hugely important.


there are alot less people playing in the hours of about 8:00-14:00 gmt

If thats the only time you can get on then your score is going to be less.


Yes.  Their score will be less.  Are you suggesting we artificially boost peoples scores because of the time of day they play?  

How then are you going to compare people who can only play around midnight to those who can only play around noon?


So really, a higher Field rating is MUCH better to have. 


Seems a bit unfair.


This isn't about being fair, its about assigning rank to a much larger set of players then the current 10 day rotation allows for.  


Skill = Field * (Mastery / 100) * (Win Percent / 100)

field and mastery which seem like weird things to base skill on


Hmm  maybe I should have called it "Score Modifier" instead of Skill.


Score = Win Percent * Win Percent * Rank * Rank * Skill 


Why are you squaring rank? and what rank is that? phil's rank on your profile?

Isnt that just like saying that if you play 200 games and get to GA thats better than someone who has played 200 games over a longer period but not got better than 1 stripe.


I square the rank number so that the points awarded goes up, the higher up in rank you go.  Winning against a GA should give you more points then winning against a pink.


I personally dont like the whole system, im not saying i could make better as i cant program. And i dont even know if i can make a better ranking system.

the results for your system look alright mostly.

1v1 ranks look nice.

And i respect the work you have done, but i dont like the formula.


That's fair.  I never set out to please everyone and I was bound to NOT make everyone happy by doing this.  I only set out to achieve the best results I possibly could using the tools at my disposal.  I personally feel that the resulting formula, while not perfect, does a fairly good job of things.
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 5:43am

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

To be fair, I am trying to see if there is a way I can reduce the Quantity effect on the scores a bit.  If there is a way to do that without messing the whole thing up, I'll implement it.
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 5:51am

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

Take note of this:

Total Points for Bbsmylie2 in 4v4: 1571056
Total Points for Matty in 1v1: 4271812

If I were to combine the 1v1 and 4v4 ranks, Matty would be #1 hands down.

This tells me the model IS actualy working.


edit: I'll see if I can actually combine them into 1...  I'd like to see it myself!  :)
post updated on Mar 10, 2011 @ 5:56am
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 7:40am

Cabin Boy zmar01

Joined: Jun 6, 2009
Posts: 656

To be fair, I am trying to see if there is a way I can reduce the Quantity effect on the scores a bit.  If there is a way to do that without messing the whole thing up, I'll implement it.


Everything to do with quantity X 0.9

Which means it is now worth 10% less of what it was.

If if did make up 50% of the overall score it now makes up 45%.




Mastery = 100 * ((Number of Opponents * Games Played) / Total Game Count At This Rank) X 0.9

Field = (Opponents Played Against / Total Opponents Possible) X 0.9

for example.

But be careful not to go around tweaking and then checking the leaderboard and tweaking more ect.

Its better to find your formula first, then see what it comes up with.

0.9 was just a number i pulled out of no where.


But it might just be better to get rid of phil's ranks out of your ranks first.
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 8:13am

Cabin Boy grims99

Joined: Feb 12, 2009
Posts: 773
Location: London, England

hehe well I for one defo prefer the offical net ranks leader board ;)

grims99 - pro galcon player
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 9:15am

Cabin Boy ionizedfire

Clan: iElite
Joined: May 25, 2009
Posts: 178

Yay another question.  Lol sorry jazz.  What can I say, I like looking at my stats on there.

Mastery = 100 * (Number of Opponents * Games Played) / Total Game Count At This Rank 

Against 1 stripes in 2v2:

Me
60 Opponents
453 Games
65% Mastery

Bonehammer
57 Opponents
440 Games
89% Mastery

^? 


Edit: Oh I just had a thought. Does your game data update faster than your mastery, field, skill, and score stats update?
post updated on Mar 10, 2011 @ 9:20am
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 10:35am

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician


Edit: Oh I just had a thought. Does your game data update faster than your mastery, field, skill, and score stats update?


Yup  ;)  From 12:30am to 8:30am its importing games.  From 8:30 to 9:30am its updating the stats.  Then the process starts again at 12:30pm.

But I'll check to make sure thats what this is, just in case.
post updated on Mar 10, 2011 @ 10:36am
Re: Official Stat Warriors Leader Board :: Mar 10, 2011 @ 8:09pm

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

Well...  my home computer is doing the importing, and it did an auto-update and rebooted itself while I was away.  :(  

Time to figure out a way to get the shared hosting company to run this instead.
Re: Official Stat Warriors Leader Board :: Mar 11, 2011 @ 1:47pm

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

Update:

I have the hosting company now polling the game dump file every hour.  Looks like it takes roughly 30 minutes to import the last hours games into the database.  And then about 5-15 minutes after that, the Leader Board stats finish its update process. 

So at the top of every hour, you can expect to see current standings.   =)

Much better then twice a day.
post updated on Mar 11, 2011 @ 1:49pm
Re: Official Stat Warriors Leader Board :: Mar 11, 2011 @ 2:09pm

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

Question:

I can now add in the 2v2_3p and 2v2_2p game types, AS WELL as an overall FFA, Teams, and Labs Leader Boards which combine all the gametypes for each into 1.  

Would you rather see a row of all gametypes when you look up a particular alias, like you see now, or would you all rather see just the cumulative rank flag for FFA, Teams and Labs?
post updated on Mar 11, 2011 @ 2:16pm
Re: Official Stat Warriors Leader Board :: Mar 11, 2011 @ 2:59pm

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

New Leader Boards!

http://toomanywheres.com/galcon/LeaderBoard.aspx?type=teams 

http://toomanywheres.com/galcon/LeaderBoard.aspx?type=ffa 

http://toomanywheres.com/galcon/LeaderBoard.aspx?type=labs 


These take the BEST score from each individual game type within Teams, FFA or Labs and uses that to place the individual player in the ranking.
Re: Official Stat Warriors Leader Board :: Mar 12, 2011 @ 10:23pm

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

IonizedFire - Nice Job!!  You are ranked #3 in FFA and #4 in Teams.  

I also like that you have achieved these ranks against a lot less opponents then most of the other players.   Well done!

http://toomanywheres.com/galcon/LeaderBoard.aspx?alias=IonizedFire
post updated on Mar 12, 2011 @ 10:24pm
Re: Official Stat Warriors Leader Board :: Mar 12, 2011 @ 11:32pm

Cabin Boy ionizedfire

Clan: iElite
Joined: May 25, 2009
Posts: 178

Sweetness. Thanks :)
Re: Official Stat Warriors Leader Board :: Mar 13, 2011 @ 3:28pm

Cabin Boy zmar01

Joined: Jun 6, 2009
Posts: 656

I'll see about providing an excel data dump like Phil does...


*Reminder*

oh not like phil. The whole reason is I can't copy the stat dump into excell. 

Who played who. Who won and who with. 

2v2 prefabaly
or 1v1 but not FFA. 


I will have to make a new formula for FFA. 

Like 

Skill= longest time without movement
Mastery=how many higher ranked players you prevented winning by suciding. 
Field= number of planets taken just before standoff.
Score=the most amount of ships you had in 1 planet

what do you think?
Re: Official Stat Warriors Leader Board :: Mar 13, 2011 @ 4:58pm

Cabin Boy esparano

Clan: iElite
Joined: Nov 5, 2009
Posts: 3471

hahaha.. the goal is to make an equation that favors rank whores and noobs less :P
post updated on Mar 13, 2011 @ 11:48pm
Re: Official Stat Warriors Leader Board :: Mar 13, 2011 @ 6:00pm

Cabin Boy socratic

Joined: Feb 10, 2010
Posts: 1947
Location: =3

2057 games for just my Socratic Account.... how many total on my 110+ aliases combined O_o
Re: Official Stat Warriors Leader Board :: Mar 14, 2011 @ 11:37pm

Cabin Boy jazz

Joined: Feb 20, 2009
Posts: 534
Location: aka The_Musician

zmar01 said ...
I'll see about providing an excel data dump like Phil does...


*Reminder*



http://toomanywheres.com/galcon/DataExport.aspx?type=2v2&date=03/12/2011 

You must specify the Type of the game data being retrieved, and the Date for which you want.

Date should be formatted as mm/dd/yyyy and the available Types are:

1v1
3v3
4v4
2v2
2v2_3p
2v2_2p
BOU
CRA
STE
ASS

Page 5 / 22 1   2   3   4   5   6   7   8   9   10   11   12   13   14   15   16   17   18   19   20   21   22 

You must sign in to post.