Author Topic: Global Statistics Experiment  (Read 5914 times)

loial21

  • Autococker
  • Posts: 2807
Re: Global Statistics Experiment
« Reply #20 on: August 12, 2005, 02:06:20 AM »
You should also put something in there that shows if you're running and jumping because it takes less skill to run and jump then it does to play while just walking. Also, while you're at it you should put into effect if they're crouching or not.
WHat you make sense :)

b00nlander

  • Autococker
  • Posts: 784
Re: Global Statistics Experiment
« Reply #21 on: August 12, 2005, 02:16:38 AM »
*cough* http://www.planetquake.com/gibstats/ *cough*

good idea, i had that system running but its a lot of work with all the newbies and all the players changing their names constantly, thus only making a rating on a dedicated community matchserver would be good.

Blitz

  • 68 Carbine
  • Posts: 329
Re: Global Statistics Experiment
« Reply #22 on: August 12, 2005, 02:51:47 AM »
..or if there was a global nameserver/etc.

XtremeBain

  • Developer
  • Autococker
  • Posts: 1470
Re: Global Statistics Experiment
« Reply #23 on: August 12, 2005, 03:13:45 PM »
Gibstats is more geared towards a Q2DM or variant style of game.  It provides good statistical analysis of statistics after the maps have been played.  It doesn't provide a very practical way of automatically updating or providing real-time statistics.  Also it doesn't output the statistics out in a fashion that suits the game.  Another problem is the funnames within the game that aren't very human-friendly when logged by pball.

I agree lots with what Dirtytaco had to say.  As well I definitely don't think global stats should be implemented initially.  It would be more suitable if bundle it with a server release that does an overhaul of the game (for example - restricting public server play of "non-compliant" maps).  Right now there is full stats support of the game.  There has been since before I started playing, which is longer than most who frequent this forum or play in the public servers.

It's too bad that there haven't been as many development efforts within the community to create a better stats program but that comes with the size of the community and the interest of stats within the community.
On that note, I have a work stats system that requires two experienced perl developers to overview and catch most of the performance downfalls of my code (one to review/change and another to agree ;).  This would be ideal and not very realistic and there is still a lot of work required to make it deployable (especially in the Win32 environment).
Fortunately the system is entirely database driven and can be configured for local/remote databases easily, there is a simple quickly put together front-end that made its way into the eR33t web page previously, but finding someone to create a front-end is much easier.

Blitz's suggestions are very similar to the calculations in my present system, but that's because most were as result of our previous collaborations when I was designing the system.  There are still rooms for improvement, for example yesterday we did identify some issues while we were using a small sample of previously captured log data, so there are still some room for tweaks.

jitspoe

  • Administrator
  • Autococker
  • Posts: 18802
Re: Global Statistics Experiment
« Reply #24 on: August 12, 2005, 03:54:40 PM »
Right.  Stats are only accurate if you have some kind of global ID system.

Having stats only for matches is an interesting idea.  It would certainly be easier to regulate, and might promote more clan play, but stats might not be the best incentive for clan play.  By that I mean it wouldn't be good to have people playing in clans solely because they're motivated by stats.

One thing that might be interesting is to have different skill categories instead of one generic "skill number."  For example, shooting people near your flag and shooting flag carriers would count toward defensive points and grabs, caps, and shooting people near their base would be offensive points.  And you could have accuracy, efficiency, etc. as well.

Eiii

  • Autococker
  • Posts: 4595
Re: Global Statistics Experiment
« Reply #25 on: August 12, 2005, 06:43:25 PM »
Right. Record what direction they're facing too, and their Z axis value. It's easier to jump when you're lower than 0, so it's easier to kill.


:P

ZeroLegend

  • VM-68
  • Posts: 161
Re: Global Statistics Experiment
« Reply #26 on: August 12, 2005, 09:25:20 PM »
if your are going to do it in matches people might decide to match the noob clans to build up stats so they dont look bad. So the skill rating as first thoughtout would play a role in the matches. So people would be a little more motivated to play good clans and try their best if they care a lot for their rating. Pub is a bad idea anyway because i myself pgp the whole map sometimes cause im bored. I tend not to play well when im bored...Pub is boring...

Dirty_Taco

  • Map Committee
  • Autococker
  • Posts: 1630
_
« Reply #27 on: August 13, 2005, 03:07:47 AM »
Post removed
« Last Edit: July 26, 2010, 01:15:08 AM by Dirty_Taco »

Dirty_Taco

  • Map Committee
  • Autococker
  • Posts: 1630
_
« Reply #28 on: August 13, 2005, 03:14:33 AM »
Post removed
« Last Edit: July 26, 2010, 01:15:00 AM by Dirty_Taco »

ZeroLegend

  • VM-68
  • Posts: 161
Re: Global Statistics Experiment
« Reply #29 on: August 13, 2005, 10:33:04 AM »
lol

Eiii

  • Autococker
  • Posts: 4595
Re: Global Statistics Experiment
« Reply #30 on: August 14, 2005, 02:04:23 AM »
This is random, but you'll notice that the timestamp on jits' post is AFTER mine, even though in comes before in order. Weird.

agent-hwk

  • VM-68
  • Posts: 178
Re: Global Statistics Experiment
« Reply #31 on: August 14, 2005, 12:57:35 PM »
lol

jitspoe

  • Administrator
  • Autococker
  • Posts: 18802
Re: Global Statistics Experiment
« Reply #32 on: August 14, 2005, 04:53:10 PM »
eiii: not for me...

ZeroLegend:  The whole point of having statistics set up this way is so that the ranking of who you're matching determines your change ranking.  If you match a low-ranking newbie clan, you probably won't be able to increase your ranking much if at all, however, if you lose to them, it will significantly decrease your ranking and increase theirs.  Well, maybe not significantly, but more significantly than matching a clan of equal or higher ranking.

agent-hwk

  • VM-68
  • Posts: 178
Re: Global Statistics Experiment
« Reply #33 on: August 14, 2005, 07:45:55 PM »
thats sounds nice

Termin8oR

  • Autococker
  • Posts: 1023
Re: Global Statistics Experiment
« Reply #34 on: August 17, 2005, 05:53:17 PM »
Wow... This is a great idea...

-Termin8oR