Ladders in 1.0

Discussion in 'Planetary Annihilation General Discussion' started by neutrino, July 23, 2014.

  1. jamiem

    jamiem Active Member

    Messages:
    115
    Likes Received:
    89
    The ladder for ranking/competition doesn't seem to be immediately necessary, but having a way for people playing public matches to see each others approximate skill/experience in lobby may help make the multiplayer side of the game more fun and accessible to new and existing players, and not just those interested in competitive play.
    cdrkf and lokiCML like this.
  2. tatsujb

    tatsujb Post Master General

    Messages:
    12,902
    Likes Received:
    5,385
    https://forums.uberent.com/threads/uncanny-parallels-hawken-an-early-access-game-with-issues.62162/
    there you go
    [​IMG]
    steam charts image.
  3. stuart98

    stuart98 Post Master General

    Messages:
    6,009
    Likes Received:
    3,888
    I think that might be the wrong link there bro. ;)
    Raevn and squishypon3 like this.
  4. tatsujb

    tatsujb Post Master General

    Messages:
    12,902
    Likes Received:
    5,385
    nope. Out of decency I didn't link the steam charts image straight-up.

    but that's what mattered to me.
  5. mgmetal13

    mgmetal13 Active Member

    Messages:
    203
    Likes Received:
    151
    Why not just add PA Stats to the released game for now until the planned competitive feature update can be finished? That way there is a ladder in the released feature set.
    cdrkf likes this.
  6. neutrino

    neutrino low mass particle Uber Employee

    Messages:
    3,123
    Likes Received:
    2,687
    It's definitely something to consider.
  7. Raevn

    Raevn Moderator Alumni

    Messages:
    4,226
    Likes Received:
    4,324
    Except it's a PA steamchart. We were talking about FA in that particular instance. So yeah, wrong link ;)

    FAF has a similar population to Steam, both of which are probably eclipsed by those who aren't playing via either client, which makes it less than 1/3 of the playerbase.
  8. tatsujb

    tatsujb Post Master General

    Messages:
    12,902
    Likes Received:
    5,385
    dude i'm exhausted and drunk, gimme a break I'm only up because the forum needs me kinda like gotham needs batman right now.
    Screenshot from 2014-07-25 02:29:48.png

    multiply FA's stats times a million to get FAF's stats.
    [​IMG]
  9. cola_colin

    cola_colin Moderator Alumni

    Messages:
    12,074
    Likes Received:
    16,221
    It's a really stupid idea to compare those charts, as none of those games forces the player to even use steam.
    Quitch and Raevn like this.
  10. tatsujb

    tatsujb Post Master General

    Messages:
    12,902
    Likes Received:
    5,385
    the hawken thread comes in handy here, 'cuz it informs us that steam chart for PA is inclusive of uberstadalone version. (wich makes sence since both players play together)

    whereas your argument is valid for FA. steamFA and FAF aren't a common player base. and steam is not aware of FAF, it does not include it in it's stats.

    As a result and by base mathematics the above charts show us that FA+FAF > PA even if FAF = 0
  11. beniesk

    beniesk New Member

    Messages:
    19
    Likes Received:
    5
    Disclaimer: my math may be completely wrong... believe these numbers on your own risk.

    This chart shows your chance of playing at least one game with an opponent of similar strength in one week, depending on the number of players (of your strength), and depending on the time you and your opponents are willing to wait.

    For example: if there are 1000 players of your strength, and each player logs in once a week and waits up to 10 minutes to get a game (or 2 times a week, waiting up to 5 minutes...), then your chance to get a game in one week is around 85%.

    This chart completely and utterly ignores peeks at weekends or evenings, or the fact that you might want to play more than one game in a week.

    matchmaking.png
    What can we learn from this? A matchmaker requires several 1000 players to work.


    How do I get these numbers?
    - A week has 10080 minutes, let's call this "d"
    - The average player waits exactly "m" minutes per week (e.g. a player cancels the matchmaker after 5 minutes of waiting, and he attempts to play 2 games in a week -> he waits 10 minutes per week)
    - "w = m/d" is the relative amount of time a player is willing to wait. "w" is between 0 and 1.
    - If there are only two players in PA, then the chance they meet and get a game is "1 - (1-2w)**2 / (1-w)**2" (see http://math.stackexchange.com/questions/103015/chance-of-meeting-in-a-bar for an explanation). Let's call the probability of two players meeting "p"
    - This means the probability of two players not getting a game is "1-p"
    - If there is a player P1, and "n" other players, then P1 does not get a game with a probability of "(1-p)**n"
    - Meaning if there are "n+1" players, then one player finds one opponent with a probability of "1-(1-p)**n".
    - And that last formula was used to paint the lines in the chart.

    You find the excel sheet that created this chart attached to the post.

    Attached Files:

    knub23, Abaddon1 and tatsujb like this.
  12. tatsujb

    tatsujb Post Master General

    Messages:
    12,902
    Likes Received:
    5,385
    I spared you the gif this time but bear witness that I was reeally tempted to put another Andy Samberg
  13. Abaddon1

    Abaddon1 Active Member

    Messages:
    209
    Likes Received:
    169
    Tatsu, you do realize that the steamcharts are of number of players currently playing, and the FAF stats are of unique players to the servers. You can't use the steamcharts info to get a very accurate view of exactly how large the game population is. Also how would steams statistics be able to know how many players are playing the game not-through steam. If this is true, PA is the only game I know of that steam can magically tell a total population value for.
    byte01 likes this.
  14. cola_colin

    cola_colin Moderator Alumni

    Messages:
    12,074
    Likes Received:
    16,221
    Most players play much more than once a week.
    Also you can push down those numbers by making it visible to all players who are logged in that there is somebody searching for a ranked game. So a person who is searching a game is not hidden, but the game shows "There is somebody who wants to play a game on your skill level right now".
    Really there is no need for all this math. FAF has exact numbers on how many players it has and how many ranked games come to be. What's so hard in just reading that?
    cptconundrum likes this.
  15. tatsujb

    tatsujb Post Master General

    Messages:
    12,902
    Likes Received:
    5,385
    how the hell would steam players and uber players be able to play together otherwise?

    they play together but their mutual game only counts for one side's count? what?

    it's pretty easy for steam to do this from my perspective. i'd be harder for them to filter, non steam users out of their stats.

    anywho we can get FAF stats if we need. you can log into FAF at anytime and the first thing you see is global chat (which shows you the online count in the search user bar) so take samples, make an average and presto.

    I said earlier what numbers I'd picked up personally : https://forums.uberent.com/threads/ladders-in-1-0.62168/page-12#post-966052
  16. Abaddon1

    Abaddon1 Active Member

    Messages:
    209
    Likes Received:
    169
    Um no it wouldn't. The steam stats are based on the number of steam accounts currently listed as "playing Planetary Annihilation" that are smoothed averages of peak players. It doesn't know what other people are in games steam people are playing, only Ubernet would actually know the total number of players in game at any time, and I don't think they've made those numbers public.
    stuart98 and Raevn like this.
  17. Raevn

    Raevn Moderator Alumni

    Messages:
    4,226
    Likes Received:
    4,324
    Because they both use the backend Uber servers. But steam doesn't use that to determine player count, it simply looks at when a player starts PA via steam, and when it closes it. Steam player counts do not include the standalone launcher counts.

    Edit: Ninja'd
    stuart98 and Abaddon1 like this.
  18. mered4

    mered4 Post Master General

    Messages:
    4,083
    Likes Received:
    3,149
    [​IMG]
  19. beniesk

    beniesk New Member

    Messages:
    19
    Likes Received:
    5
    Not so certain about your first statement, a minority is playing every day. But what about the average player.

    What FAF does with its "someone of your skill level is searching a ranked game" is increase the "m" parameter in my formula - meaning FAF requires less players to get to the same high probabilities. It was a brilliant idea of ZeP to include that button.

    FAF provides one datapoint, it does not tell us what the minimum amount of players required for a matchmaker is. My formula does neither (it's far to simplistic for that, and we do not know the input parameters), but at least it gives a feeling how these numbers add up. See it as "yet another tool" in the discussion.
  20. tatsujb

    tatsujb Post Master General

    Messages:
    12,902
    Likes Received:
    5,385
    take it up with this guy, not me:
    you're just buying yourself time. first of I believe @Shwyx second we all know here on the forum that PA's playerbase is majoritarily steam. are you going to deny that?

Share This Page