Galcon 2: 36 players online!
:: Community ::
ForumsMessagesModsChat
 

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

You must sign in to post.

Page 11 / 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 28, 2011 @ 6:13pm

Cabin Boy no.yes

Clan: iElite
Joined: Mar 7, 2010
Posts: 120

but I'm not sure its the best solution yet...  cuz several of you wont like these results.  ;)


well, as it was written some time ago in forums that ranks are important part of galcon and each normal player cares about his rank. so each pro player switches from time to time but there are those that switch more often than others. Also people switch TO best players so they will get negative SI

But implementing this system will even up differences from switching for there is whole spectrum of bigger and smaller switchers. In addition, the game will be much more enjoyable if you know that you dont have to concentrate on switching and instead on game.
post updated on Mar 28, 2011 @ 7:00pm
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 6:18pm

Cabin Boy esparano

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

To add to what no.yes said, I'm not sure if using the numbers 1-10 to represent the ranks is a good idea, because they are just arbitrary numbers and don't really represent the "skill of a player."  Like, a GA should be able to beat a pink flag by much more than 10-1, or beat a red flag by more than 10-5.  Multiplying by these numbers or dividing doesn't really make sense since the numbers don't mean anything, they're random. Plus, an opponents to partners ratio such as 3:2 would affect the score MUCH more than 10-9, even though they're only a difference of one rank. 
    However, subtracting to find the rank difference can be much more useful, which is why I think no.yes's new % multiplier is a better idea. But I have to say, I'm not sure that 40% is correct..
You have to think of it this way:
If your opponents to partner's flag ratio is 5:6, then your partners are only 1 flag higher than your opponents.  Do you really win 40% more games because of this? Do you actually win more or less than that? The % that is chosen has to reflect the ratio of games that two flags will win against one another. Plus, win% in jazz's equation is multiplied three times (the last time I checked) so that affects your score a LOOOOOT more.  The problem is that we have to agree on a good %, otherwise it could hurt people or help people too much

I say that the % should go up a lot more with each flag difference.  A flag difference of 1 doesn't really say all that much, but a flag difference of 3 means you're a dirty team switcher.  

AOR - APR . . | . . score multiplier
-------------------------------------
-5. . . . . . . . . . . . 20%
-4. . . . . . . . . . . . 28%
-3. . . . . . . . . . . . 42%
-2. . . . . . . . . . . . 63%
-1. . . . . . . . . . . . 83%
0 . . . . . . . . . . . . 100%
1 . . . . . . . . . . . . 120%                          
2 . . . . . . . . . . . . 160%
3 . . . . . . . . . . . . 240%
4 . . . . . . . . . . . . 360%
5 . . . . . . . . . . . . 500%   
(the score multipliers increase exponentially to reflect the multiplied effect that win% has on your score)

What do you guys think?
post updated on Mar 28, 2011 @ 6:24pm
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 6:24pm

Cabin Boy no.yes

Clan: iElite
Joined: Mar 7, 2010
Posts: 120

The interesting thing is that Ionized has almost the same Switching Index as steve.jobs but the underlying reason is probably much different. 

Ionized - people switch TO him so even if he dont do anything better ranks try to switch TO him. Hence he has negative SI

Steve - he actively switches to better ranks.

for this reason maybe it shouldnt be called Switching Index but something else? Cant think of good name. Unequality index - guys propose somethin...

Also, it shows that measure of desire to switch is not possible to find from stats available because of reason above (people try switch TO best players even if they dont do anything and they will get negative index)

anyway the UNEQUALITY INDEX we use here (AOR minus APR) could be very useful to isolate skills of single player. it still can be very useful to make some adjustments. lets see what will happen.
post updated on Mar 28, 2011 @ 6:33pm
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 6:27pm

Cabin Boy esparano

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

The desire to switch doesn't affect somebody's skill^ 
Like being unsportmanlike doesn't really make you any better or worse at the game, it just makes you a big fat jerk.  (Like Zen_Power17 being one of the best fusion 1v1ers). Ranks try to isolate somebody's skill only, so unfortunately ionizedfire's ranks may suffer too.  Even if he doesn't want to be paired with good players, good players switch to his team, so his stats still get boosted either way.  (I'm not implying that this is actually the case, I'm just giving an example lol)
post updated on Mar 28, 2011 @ 6:30pm
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 6:31pm

Cabin Boy jazz

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

I gotta say... You guys are what make this project fun to work on. Thanks! =)
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 6:49pm

Cabin Boy no.yes

Clan: iElite
Joined: Mar 7, 2010
Posts: 120


AOR - APR . . | . . score multiplier
-------------------------------------
-5. . . . . . . . . . . . 20%
-4. . . . . . . . . . . . 28%
-3. . . . . . . . . . . . 42%
-2. . . . . . . . . . . . 63%
-1. . . . . . . . . . . . 83%
0 . . . . . . . . . . . . 100%
1 . . . . . . . . . . . . 120%                          
2 . . . . . . . . . . . . 160%
3 . . . . . . . . . . . . 240%
4 . . . . . . . . . . . . 360%
5 . . . . . . . . . . . . 500%   
(the score multipliers increase exponentially to reflect the multiplied effect that win% has on your score)

What do you guys think?


thanks for this analysis esparano - this is what i lacked, i think we could start with these percentages and see how it goes
post updated on Mar 28, 2011 @ 7:02pm
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 7:06pm

Cabin Boy no.yes

Clan: iElite
Joined: Mar 7, 2010
Posts: 120

the funny think is that we players can solve switching problem just by manipulating stats. who would think that this could be solved without Phil interaction :)
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 7:28pm

Ensign diebkstabber

Joined: Dec 2, 2009
Posts: 701
Location: under my bed

I think if we just break Steve.jobs galcon finger, everything will be solved!!

We could have like a galcon mafia, that goes to your house and adjusts you attitude. May not be cost effective considering players are all over the world, I'm still working out the details.
post updated on Mar 28, 2011 @ 7:30pm
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 7:34pm

Cabin Boy quantum

Clan: Singularity
Joined: May 19, 2010
Posts: 454
Location: Every(No)where @ once.

I think you have a great idea diebkstabber, maybe we could implement a thing that will remote wipe or brick your phone when you teamswitch..?
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 7:42pm

Cabin Boy ionizedfire

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

Ionized - people switch TO him so even if he dont do anything better ranks try to switch TO him. Hence he has negative SI

Ranks try to isolate somebody's skill only, so unfortunately ionizedfire's ranks may suffer too.


It's ok! I thought of a new solution for this.  I will anti-team switch! So whenever I join a server, I will switch the better players with each other.  So now when I go to chat after playing a bunch of games, people will be like 'Ahhh Ion you anti-switchin galconite!' It could work.

Anyways, that does happen though.  People switch to my team all the time...but it does still give me an advantage whether or not I want it, so it makes sense to include rank partner vs rank opponent factors to the scores.
post updated on Mar 28, 2011 @ 7:43pm
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 7:42pm

Cabin Boy esparano

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

:D

Lol I'm sure you'll still be first in every category ;)
post updated on Mar 28, 2011 @ 7:49pm
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 8:49pm

Ensign diebkstabber

Joined: Dec 2, 2009
Posts: 701
Location: under my bed

Ok you've fixed the " goozok problem" your working on the " Steve.jobs problem" then can you tackle the " diebkstabber problem" which is why I have been so ****ty lately!! Jazz please help do you have a magic formula??
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 8:51pm

Cabin Boy jazz

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

yea...  stop playing that account and go anonymous for a while.  :)

Keep in mind that these stats will hang around for more then 10 days, so its OK if your account goes pink for a few.
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 8:52pm

Cabin Boy jazz

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

Your my hero Ion!  Takin one for the team!!  Literally!  :)
Re: Official Stat Warriors Leader Board :: Mar 28, 2011 @ 8:55pm

Cabin Boy jazz

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

I'll try to get this added to the ranks by this weekend.  

What I like most about this change is that it works for ALL team games and not just 2v2.  It will work for Fusion and Desktop just as well.  =)  Now that is Cool!

Hats off to you all for coming up with the idea!
Re: Official Stat Warriors Leader Board :: Mar 29, 2011 @ 1:40am

Cabin Boy no.yes

Clan: iElite
Joined: Mar 7, 2010
Posts: 120

Just another thought. MAYBE WE SHOULD IMPLEMENT THESE CORRECTIONS SOMEHOW AT WIN% LEVEL, NOT AT SCORE LEVEL ? it would go through win% correction toward score as well

This way we wont have problem to find these ideal percent corrections (see below). Also exponential thing mentioned by esparano wont be necessary.

corrected win% doesnt have to be shown in tables (it would cause too much confusion). Everything would be calculated in background.

first lets define one thing (I use SI for index name for consistency)

+1 SI game - a game where your partner rank is one lower than opponents rank (averaged). Eg. your partner is red (captain) and you play against two blue (admirals).

it would be possible to calculate what is win% for all +1 SI games (using stats data)

lets say overall win% for +1 SI games is 40%. 

overall win % for zero SI games should be around 50%

Then lets say your win % for +1 SI games is 48%.

You are 8% above the average. To reflect this we would just need to increase win% by 10% for all +1 SI games. Then your win% that would go to calculations will be 58%.
post updated on Mar 29, 2011 @ 2:31am
Re: Official Stat Warriors Leader Board :: Mar 29, 2011 @ 1:51am

Cabin Boy no.yes

Clan: iElite
Joined: Mar 7, 2010
Posts: 120

On another thought this would probably require too much work too implement
post updated on Mar 29, 2011 @ 2:35am
Re: Official Stat Warriors Leader Board :: Mar 29, 2011 @ 2:22am

Cabin Boy quantum

Clan: Singularity
Joined: May 19, 2010
Posts: 454
Location: Every(No)where @ once.

I really like the "all team games" stats! So much more awesome then Phils ranks we have currently. 

I wonder if we took into account and really dissected the whole points for a win in 3 player 2v2 games. Like if I lose my teammate while in intermission, and I beat the two people I was playing against before, it only counts as one win, (additionally, I think this is why Phils ranks exclude 3player 2v2 games from ranks) it seems like a bump in the road, because to be fair, you beat 2 players, and so it seems like the ranks should be re equated for another variable for the odds difference in winning against 2 other players.  I see it as a 1v1 should equal 2 wins or 2 losses... Or a 3 p 2v2  the solo player gets to win the equivalent of twice since he played as 2 players simultaneous ly...

When it's just 1v1 in a 2v2 round It would be 50/50 but...
When ur against 2 other players, what the advantage or disadvantage? And how should that be scored when included with 2v2 games?
Does that make any sense?
post updated on Mar 29, 2011 @ 2:25am
Re: Official Stat Warriors Leader Board :: Mar 29, 2011 @ 2:37am

Cabin Boy zmar01

Joined: Jun 6, 2009
Posts: 656

This all seems good and all but arent you forgetting something?

This discourages ranking up,

and the less games you play the higher your SI score will be or whatever you want to call it.

As you say APR

If you stay a blue flag and dont play other blue flags the next rank you can get up to is GA

So staying blue lowers your APR if you would have been a 3 stripe.

8 compared to 5

meaning that your average is going to be at least 1 less.
And this will not effect your stats in any other way, much, if you just play less games over a longer period of time.

Here lies another problem with basing your rank system on Phil's


To add to what no.yes said, I'm not sure if using the numbers 1-10 to represent the ranks is a good idea, because they are just arbitrary numbers and don't really represent the "skill of a player."  Like, a GA should be able to beat a pink flag by much more than 10-1, or beat a red flag by more than 10-5.  Multiplying by these numbers or dividing doesn't really make sense since the numbers don't mean anything, they're random.



I like the summery.



I dont know if its APR/AOR or AOR/APR

either way its based on phil's ranks.

And i said before the only reason why it works in the mathmatical equation, is because phil's ranks are based on the higher ranks have high win% against people with high win%
Weight is irrelevant at stats over several months as it was designed for 10 days.
Or are you still keeping the 10 day stat ranks?
Im not sure.



I dont think i can realistically implement my idea on a spread sheet.

As i need to make a score for every game and find some way of adding that score for each game onto each player.
I might be able.

but the reason why i made the spreadsheet was to test out ideas.

I should work on it more.
Re: Official Stat Warriors Leader Board :: Mar 29, 2011 @ 9:58am

Cabin Boy no.yes

Clan: iElite
Joined: Mar 7, 2010
Posts: 120

APR is not dependent on your rank at all. Its only dependend on rank of your partners.

The INDEX (SI) that we agreed on to use is AOR-APR (neither APR/AOR or AOR/APR - these were just first propositions) This index will be used to adjust score by some percentages. Its just the matter of finding correct percentage adjustments for different values of index. Here we have occasion to adjust fact that numbers assigned to rank are arbitrary (as esparano mentioned) and other things.

The data that is used for calculation of AOR and APR is based on data for more than 10 days - the all time data that jazz is collecting on his server.

However, the ranks shown in source data are based on Phil's 10 days ranks. So, if lets say ionized doesnt play for some time he becomes pink and then boosting INDEX for his partners although he in reality is 3stripe/GA level player BUT:

In longer time it doesnt matter because everything is even up. In addition, I noticed that jazz stats may give regular players incentive to NOT switch user names, so we will see such situations less and less.
post updated on Mar 29, 2011 @ 1:03pm

Page 11 / 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.