Experience is really weird right now. 

I look here - http://pantheon.demigodthegame.com/rankings/overall/page/6/

and see I am at rank 104, experience 1,072, and have 10 wins and 7 losses (of course, when I click on my name I see 23 wins and 4 losses?)

Then I look 3 ranks up to Excerion, rank 101, experience 1,077, who has 3 wins and 4 losses.  His average favor per game is the only thing under-the-hood that would give him a higher experience number.  If that is indeed the criteria then a new player who gets max favor in 10 games will be ranked higher than an experienced player with 100 wins but lower average favor?   

I got the feeling that games against the AI didn't count towards experience in Pantheon.  This was because my first bunch of games all were against the AI only and my experience never changed (stayed at 1000).  Once I started playing players though my experience started changing.

This means that your wins/losses doesn't necessarily correlate to your experience since the majority of your games don't count towards experience.  I could be totally wrong on all this though, it was only derived through a few glimpses at my ranking.

So I did three custom games of 5v1 and just let them run until I won.  Mostly I just did it for giggles.  Here is how my games list records them...

No players are listed, it says I earned 0 favor points and it says that I lost them.  It was all AI and me.  All I did was buy citadel upgrades occasionally while I stood in the base.  I did earn favor points after each one.

I, of course, don't care about not getting my stats for these games as I did nothing in them.  Just wanted to point out that they were recorded wrong.

We will continue to look at why games are not being recorded properly, specifically whenever AI players are involved or when someone disconnects.  I have the data I need to show that a problem exists, now it's just a matter of tracing it and resolving it.

The current Pantheon site has a few issues due to some changes in the backend.  Please ignore any problems for now.  I am still working on the new design (whilst juggling a few other things) and that will resolve all of the major, design-related issues as well as make things a little easier to navigate.

Experience is what we use to rank players.  You gain it by playing games against other players (and possibly AI, but we have yet to decide).  We use experience instead of favor points because we want experience to be based on player skill, not how many games you play.  As far as a discussion with Brad, that's on his terms.  Keep in mind the rating system is still very early and will change as we go along.

Thank you for the link.  I'll pass this along to T-Man, who is the one in charge of the rating system.

The rating system is not based around favor points as far as I know.  It is based more on how well you do overall in each individual game.  If you have any suggestions or problems in the way the system has ranked you, feel free to bring them up.  At the moment the system is somewhat simple, but we are planning on making it more complex and thorough.

Bara

Interesting.  My understanding of how players are ranked in individual games is via a series of achievements like: most kills, least deaths, most gold spent on citadel, etc. for which they are awarded favor points.  If there are other criteria please let me know.

If you do just use "most" to rank the players then in team games you run into the following problem - the ingame achievements are awarded across teams.  So an amazing player on a losing team is almost never going to be ranked highly compared to average players on a winning team, and as long as we are playing pub games amazing players can lose ^^

For example - Ultimeh is probably one of the top 20 players right now, but in a game he played against me on March 15th he earned only the minimum favor points possible as a participant.  My team was very strong, so we monopolized the achievements.  You see this across the top players - most of the time they are on a weaker team they gain minimum favor, Even when they play well they can't get as many kills or destroy as many buildings as someone who is playing ok but has the advantage of momentum.  

So maybe do some under-the-hood or overt ranking within teams to supplement overall performance across teams?  I dunno...

Heh, don't worry.  The AI will have an experience rating just like everyone else.  So an AI that plays better will have a higher experience rating versus one that plays horrible.  Most AI may end up having less experience than most players, but then we'll have a reason to add "ultra hard AI" sometime down the line (this is all my speculation, and not something we or GPG will actually do).  Regardless, because AI will have an experience rating, you will gain experience proportionate to their experience, so if an AI is easy to beat, you won't get a lot of points.

Bara

I am finding that the stats are coming up wrong, currently I have a 18 game losing streak, though since this last beta my teams haven't lost a game yet.  When I win the losing streak continues though the win streak stays at 3.  It also is now showing a loss in a custom game last night and considering I was the one hosting the game and the one who destroyed their base, it shouldn't have been a loss, but a win.  The other 3 people who were on my team received a win, but I received the loss, which makes it even stranger.

I'm sure it will be fixed by release, just wanted to make you aware in case it was something not mentioned.

[BUG / ERROR]

When I am looking at my player profile, and I click on the tab for custom games, I get an error. 

This is a link to the page that has the link:  http://pantheon.demigodthegame.com/player/12220/overallstats

This is the broken link:  http://pantheon.demigodthegame.com/player/12220/customgamestats/

I'm not sure what's causing the streak problem, but it might be related to the "everyone is on forces of light" problem we are currently having.  I'll take a further look at the numbers and see if I can point out anything to the devs.

Yes, I know that the page is currently broken.  Monday morning I will be pushing an update to the Pantheon site with an entirely new design.  Assuming no problems occur with the new update, that specific error should be fixed.

Bara

I know you are probably aware of it but if you take any other indicators than "winning" to compute the experience you risk that people will exploit the system and will not play to win but to gain most experience. Even if "winning" is the major component it can lead to such unpleasent things like "kill stealing" (like a snipe on a doomed foe who would have died anyways), cowardly play (lets say you are already in a huge advantage but getting not killed even once gives a experience bonus you play totally passive so the game takes another 15 minutes instead of the 5 minutes it would have taken if you played just "to win" - no fun for the losing team) or intentional playing around with your enemy with no intention to win fast but to kill them often as long kill streaks give bonus points etc. which is no fun to the losing team either.

Just came here to ask in full release will rankings be decided by who has the most experience? This is, imo, a really innacurate way of deciding the rankings as support characters etc. may not get as much experience. Win/Loss ratio or favour points is still a better way than experience imo.

And just read TBO's post, agreed.

From what it sounds like experience will be given for wins.  So the support character on a winning team will get roughly the same as other characters.  This is their version of ELO, correct me if I am wrong.

another game I wasn't in at all that I'm seeing on my stats and in my games list:

http://pantheon.demigodthegame.com/games/player/11146/

Looks like the new site is up, though games still aren't showing up properly (I have a game listing that I'm not in and several games listed where the players aren't showing up on the correct teams, etc.)

Should we start reporting bugs for the new site as if they are new bugs or is what launched this morning just a UI upgrade (no backend code changed)?