DemiCom DemiCom

opengames.asp stats wrong ???

opengames.asp stats wrong ???

lets look at these stats: http://www.demigodthegame.com/opengames.asp

in this moment there are :

3200 players online

73 games in progress

10 games waiting for launch

-> 4-10 players / game

-> use simple mathematics and you know that there are only around 500 people involved in DG games !

-> so what are the other 2700 doing ?

-> are they waiting for better weather ?

-> or is "3200" the number of people with launched Impulse ????

134,093 views 67 replies
Reply #51 Top

Quoting Lumifly, reply 25
It's all a conspiracy.  You guys, who are clearly staticians, mathematicians, software engineers, computer scientists, and all manner of other relevant occupations, clearly have the experience to say the numbers are fudged or even "off a bit."  It is clear that your superior reasoning and logic chain, which is so elaborate that you cannot even put it down on paper for us to analyze, is much more reasonable than simple programatic functionality to track these things.

No, indeed, you are right.  It is a conspiracy contrived by Stardock and GPG to make their game seem like it has only a modestly small population instead of a smallish population.

Please also tell me about how 9/11 was done by the lizardmen of the US government and how THAT VAN is ALWAYS falling you and you JUST KNOW they are watching you.

In other words: back up your arguments with reasonable logic.  There is absolutely no reason to think they would be fudging numbers, let alone fudging numbers in such a way that doesn't even make them look good.  Let's not forget that if they wanted to fudge numbers, they could EASILY add games to the game list to actually make them look far more active.  But no no, you guys are right...it is definitely a conspiracy.

Edit: And yes, GPG, please put in a "Players in game" stat anyway.

There are a lot of reasonable people posting here who arent saying that Stardock is lying.  They are saying that there appears to be something wrong or misleading about the stats.  You cant conflate "something looks funny with the stats" and "Stardock is fudging the stats".  Those are two very different things.  Reasonable people know that its extremely unlikely that Stardock would lie, and if they did lie, they would do a better job (ie not have an obvious disparity between games and population).  Only the nutty trolls think that SD is lying.

People have presented comparisons to other games and brought up similar relevant facts to back up their claims.  Something IS funny with the stats - why is 90%+ of the DG population pinging at least every 30 seconds but not playing.  That fundamentally doesnt make sense.  Shadow_Avenger is the only one with a reasonable explanation, and even that wont really explain things IMO.

I personally think that something is bugged in their total users stats.  But thats just a guess.

Reply #52 Top

What about pantheon matches? I don't see them listed under Games In Progress so they probably are not counted under Current Games.

Reply #53 Top

I think Skirmish and Pantheon games are marked as "Instant Action Match".

Reply #54 Top

There are no pantheon matches in the list because there is no pantheon running at this moment.  When there was one, the matches were listed correctly, as were people waiting in the que and people in pantheon games that have not yet started.

In other words: back up your arguments with reasonable logic.  There is absolutely no reason to think they would be fudging numbers, let alone fudging numbers in such a way that doesn't even make them look good.  Let's not forget that if they wanted to fudge numbers, they could EASILY add games to the game list to actually make them look far more active.  But no no, you guys are right...it is definitely a conspiracy.

Lumifly, I don't think a single person in this thread ever wrote that SD is lying.  As for logic - both Frogboy and Istari gave their logic of other games/systems having similar stats.  However when we did compare the numbers (just simple math), the result was (just going to repeat and simplify for you so you can understand):

  • Demigod: 8-15% of online players are in games (There is not much to do with demigod active except play a game)
  • Steam:  13% of online players are in games (Steam logs you in without even starting any games and has a multitude of additional functionalities.  It would be like counting people with Impulse started and comparing that number to players playing)
  • Supreme Commander: 55-60% of online players are in games (GPGNet has chat and the vault library which require you to be online)

Basically in Demigod, on average, for every 3-5 hours spent with game active and logged in to server people spend 30 minutes playing.  I personally think this is pretty uncommon behaivor and one that I'm interested in learning more about.  Hence the discussion.

Now, go troll elsewhere...

Reply #55 Top

My server is querying the opengames page every hour, for purposes of later graphing the data. You can access the data here: http://coma.norikumi.com/demigod-stats/

demigod-stats.txt is a summary, but the page from every hour is also saved. The filenames are formatted in this manner: <hour> <date>.html.

<hour> is a number from 0 to 23, where 0 is 00:00 and 23 is 23:00. Like I said, I'm doing this to graph the data, but I thought you guys might be interested in watching the trends right now.

Reply #56 Top

Right now it says that over 26000 palyers are online 8O

Reply #57 Top

We think the server work yesterday/this morning has caused it to not remove old games/logouts, we're looking into it.

Reply #58 Top

It's now 48k.

http://coma.norikumi.com/demigod-stats/demigod-stats.txt - some stats. It never, ever dropped below 1.3k, and has been increasing since 3PM UTC.

Reply #59 Top

Btw. I am working on a visual representation too ;).

I am parsing the opengames.asp with a PHP HTML DOM parser and putting the values into a database. Currently I am visualizing it with JpGraph, which works quite well so far. Originally I wanted to host it on my webspace, but it only got PHP4 and the parser requires PHP5 (and I don't really want to exchange that parser).

Though the best solution would be rrdtool, just like on supcom.db/stats, since rrdtool makes a lot things easier, at least in principle and creates much nicer graphs.

Reply #60 Top

Thanks for the time you put into this Spooky! k1

Reply #61 Top

Really nice work spooky.  It would be interesting if you could parse the players in games too (human players column). :)

Reply #62 Top

Quoting Misfortune, reply 11
Really nice work spooky.  It would be interesting if you could parse the players in games too (human players column).
Yes, I am planning on doing that, shouldn't be that hard to do. I will prepare the database for it, but initially only the 3 graphs will be available for now. It should go online tonight.

Reply #63 Top

Well, implementing 'players in games' (I also count players in games that are not started yet, hence all "active" players) wasn't that hard, see updated image in my post above. Still waiting for webspace access on insidesupcom.de, where it will be hosted.

Reply #64 Top

I still don't know if single player games are counted, the semantics are a bit confusing to me. But yes I pretty much only play single player games and I'm sure I'm not the only one. Don't forget that all the players playing SP only are not likely to be active on this forum. You'll get a distorted image.

Reply #65 Top

I guess I shouldn't bother (graphing) then.

Reply #66 Top

Quoting comanoodle, reply 15
I guess I shouldn't bother (graphing) then.
Well you still can if you have the motivation :). There is nothing wrong with redundancy, in case I screwed something up in my dirty php code or so ;).

Reply #67 Top

I still don't know if single player games are counted, the semantics are a bit confusing to me. But yes I pretty much only play single player games and I'm sure I'm not the only one. Don't forget that all the players playing SP only are not likely to be active on this forum. You'll get a distorted image.

According to SD, single player players should not be counted as online as they (most likely) never log in to the server.  And even if they are counted, the data is still relevant.