Interesting, Actual Win/loss Beta4a Statistics

Funka · May 2004

<div class="IPBDescription">as parsed from server logs</div> originally posted at the OldF clan boards here : <a href='http://nsoldfogies.dyndns.org/phpBB2/viewtopic.php?t=1752&start=15' target='_blank'>http://nsoldfogies.dyndns.org/phpBB2/viewt...t=1752&start=15</a> in response to discussion that the 10-minute CO timer felt too short, etc., so i ran some stats that are summarized below. i might even be more curious to further break these down by mapname, to see what maps work best for what team, but i'm not quite that dedicated! there are links at the bottom to the source files generated for these stats, in case anyone wants to start running their own RegExps against them...

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> ("Funka!")</td></tr><tr><td id='QUOTE'>

first four days of B4a (May 1 through May 5)
62 marine victories.
82 alien victories.
44 games were 10 minute-games.

four day period, Thursday through Sunday, May 6-May 9:
138 complete games of CO played on PJ.
58 marine victories (42.03% of the wins)
80 alien victories (57.97% of the wins)
42 of these were 10-minute games (30.43% of all games played on server, which accounts for just over half of all alien victories).

same four day period (Thursday through Sunday) but one week later -- this last weekend -- May 13 through 16:
69 complete games of CO played on PJ. (Classic has been playing mostly)
23 marine victories (33.3333333% of the wins)
46 alien victories (66.6666666% of the wins)
25 of these were 10-minute games.
36.2% of ALL games (accounting for 54.35% of the alien victories) are won by the 10-minute timer elapsing.
Either marines have been slipping recently, or not enough CO is being played on the server, or it's a fluke, or a trend, ... i don't know.

SUMMARIZING ALL stats, here is from late night May 01 (release date of 4a) to May 16 (last night, roughly midnight) --

450 complete games of CO played on PJ.
191 marine victories (42.44% of the wins)
259 alien victories (57.55% of the wins)
136 of these were 10-minute games.
30.2% of ALL games (just over half of all alien victories) are won by the 10-minute timer.

while i had all this open, i put together some more interesting stats:
Classic NS, May 13 through 16 --
150 games of classic NS played.
36 games lasted from 0:01 to 09:59
65 games lasted from 10:00 to 19:59
26 games lasted from 20:00 to 29:59
23 games lasted for more than half an hour
65 marine victories (43.333%)
85 alien victories (56.666%)

Classic NS, beta 4a only, May 1 through May 16 --
487 games of classic NS beta 4a played.
110 games lasted from 0:01 to 09:59
200 games lasted from 10:00 to 19:59
104 games lasted from 20:00 to 29:59
44 games lasted from 30:00 to 39:59
29 games lasted for more than 40 minutes
226 marine victories (46.41%)
261 alien victories (53.59%)

source files are all here:
<a href='http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0501-L0505.TXT' target='_blank'>http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0501-L0505.TXT</a>
<a href='http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0501-L0505.TXT' target='_blank'>http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0501-L0505.TXT</a>
<a href='http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0501-L0516.TXT' target='_blank'>http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0501-L0516.TXT</a>
<a href='http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0506-L0509.TXT' target='_blank'>http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0506-L0509.TXT</a>
<a href='http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0513-L0516.TXT' target='_blank'>http://pj.oldf.net/nsdumps/CO_B4A_VICTORY_L0513-L0516.TXT</a>
<a href='http://pj.oldf.net/nsdumps/NS_B4A_VICTORY_L0501-L0516.TXT' target='_blank'>http://pj.oldf.net/nsdumps/NS_B4A_VICTORY_L0501-L0516.TXT</a>
<a href='http://pj.oldf.net/nsdumps/NS_B4A_VICTORY_L0513-L0516.TXT' target='_blank'>http://pj.oldf.net/nsdumps/NS_B4A_VICTORY_L0513-L0516.TXT</a>

-f!</td></tr></table><div class='postcolor'>

coris · May 2004

The only reason marines doesn't win every time is because pubbers aim sucks. If they learn how to aim, then aliens will get spanked. Big time.

Eidolan · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (coris @ May 18 2004, 08:07 PM)</td></tr><tr><td id='QUOTE'> The only reason marines doesn't win every time is because pubbers aim sucks. If they learn how to aim, then aliens will get spanked. Big time.  </td></tr></table><div class='postcolor'> 
A good alien can be just as scary.

Olmy · May 2004

Not particularly true, no matter how good a skulk may be, there's ever possibility that a marine will jump and be effected by the huge knockback, giving precious time to fire of some more rounds when bit and kill a skulk. (just one example).

matso · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (coris @ May 18 2004, 08:07 PM)</td></tr><tr><td id='QUOTE'> The only reason marines doesn't win every time is because pubbers aim sucks. If they learn how to aim, then aliens will get spanked. Big time.  </td></tr></table><div class='postcolor'> 
Nah. A sucky commander will do wonders for alien win statistics. And while some alien players are good, there is just so much you can do as a skulk.

coris · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (matso42 @ May 19 2004, 09:54 AM)</td></tr><tr><td id='QUOTE'> </div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (coris @ May 18 2004, 08:07 PM)</td></tr><tr><td id='QUOTE'> The only reason marines doesn't win every time is because pubbers aim sucks. If they learn how to aim, then aliens will get spanked. Big time. </td></tr></table><div class='postcolor'>
Nah. A sucky commander will do wonders for alien win statistics. And while some alien players are good, there is just so much you can do as a skulk.  </td></tr></table><div class='postcolor'> 
Not if you're facing good marines.

Ahnteis · May 2004

You're forgetting stats like "players per team at beginning of game", "players per team average per game" and "average team stacking per game".

SaltzBad · May 2004

Ahnteis, the point is - with a half-decent commander and good marines, they'll win. But a fair percentage of what would normally be a marine win is racked up as an Alien win due to really horrible commanding, and not entirely retarded aliens. Hence saying that actual win/loss statistics from the amount of games that ever take off is probably alot more biased towards 'rines.

To some degree it doesn't matter how many leet marines they have, if they're not getting med/ammo/phasegates/mines or even Armor1.

Nemesis_Zero · May 2004

All said and done, this is more 'down to earth' than what anybody else could so far produce. Thanks for your input, Funka.
Now please note that no single server can give us a wholesome picture of the games balance. What it can do is being representative of a certain part of the gaming demographic - and as I and Mr. Baseballbat tell the PTs time after time, <i>every</i> demographic has its place around here.

In so far, it'd maybe be more fruitful to discuss what these statistics mean for pubplay than to boldly proclaim that you won't trust any statistic you didn't forge yourselves.

Haunted · May 2004

I play on PJ's mostly. For myself, I notice that, yeah, if no one wants to comm, or the comm is bad, aliens will win. But if the comm is ontop of things, the aliens have a hard time getting a toe hold on the map thanks to phases, and elec rt's, unless there is a good fade.

Also, my biggest pet peeve, there are a bunch of wankers playing ns that refuse to play alien. So the only reason aliens win, is that the decent folks, after winning a marine game, go alien to diversify, or because they decide to go through the random (which becomes the alien) team door. So then, because the marines are full of wankers, the aliens win.

But I guess numbers don't lie, but for me, it's not as fun to play as alien, and it's much harder to make an comeback. That's how I feel, maybe the game is balanced, but it's losing its fun.

Snidely · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (Haunted @ May 19 2004, 06:31 AM)</td></tr><tr><td id='QUOTE'> Also, my biggest pet peeve, there are a bunch of wankers playing ns that refuse to play alien.  </td></tr></table><div class='postcolor'> 
What's wrong with prefering one side over the other? I don't like playing aliens anymore - I find it to be a chore. Why should I "have" to play kharaa if I don't want to?

Haunted · May 2004

First off, playing aliens makes you a better marine. Secondly, when the game starts with 5 marines vs 3 marines, but there are 4 people in ready room taking up server slots, but refusing to play, well that sucks. Basically, if you get comfortable playing both sides, it's easier to balance a game as you will have no qualms going random team.

BobTheJanitor · May 2004

Marine stackage happens, but it's not that terrible. New players will go marines because it's simply easier to learn. Experienced players will go marines because it's a lot easier to be a lone player on marines as long as someone is willing to get in the CC and occasionally drop you something. If you're the only one making an effort to win on aliens this means you usually end up dropping the only res node and then watching it go down a minute later, then spend the rest of the game as a skulk with 5 res until someone finally goes onos and gets shot down in ten seconds... It's disheartening. At least on marines if the whole team is playing the game wrong, you can still go to waypoints and shoot a few skulks, build an RT or two.

The point being that teams will usually balance themselves out, with the experienced players going marine for those reason, or going alien because marines are already stacked. I haven't seen nearly as much problematic marine stackage lately as there once was.

BattleTech · May 2004

The only reason the marine's didn't win is much is because alien's get a free victory for sitting around for 10 minute's.

BobTheJanitor · May 2004

You mean a free victory because the marines sit around for ten minutes, surely?

Salvation · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (coris @ May 18 2004, 08:07 PM)</td></tr><tr><td id='QUOTE'> The only reason marines doesn't win every time is because pubbers aim sucks. If they learn how to aim, then aliens will get spanked. Big time.  </td></tr></table><div class='postcolor'> 
that and commanders

...well every well know pub problem

CommunistWithAGun · May 2004

Marines are overpowered in scrims/matches
Marines are underpowered in pub play.

Why? In pubs the skill level is generally ALOT lower. <img src='http://www.unknownworlds.com/forums/html//emoticons/sad.gif' border='0' style='vertical-align:middle' alt='sad.gif' />

Kwil · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (CommunistWithAGun @ May 19 2004, 10:31 AM)</td></tr><tr><td id='QUOTE'> Marines are overpowered in scrims/matches
Marines are underpowered in pub play.

Why? In pubs the skill level is generally ALOT lower. <img src='http://www.unknownworlds.com/forums/html//emoticons/sad.gif' border='0' style='vertical-align:middle' alt='sad.gif' />  </td></tr></table><div class='postcolor'> 
Eh? I thought it was the other way around, and that's why the aliens keep seeing nerf after nerf. The very skilled folks found that any boost you give to the aliens made them dominate, so they always complain if us pubbers suggest alien boosts or marine nerfs.

SaltzBad · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (CommunistWithAGun @ May 19 2004, 11:31 AM)</td></tr><tr><td id='QUOTE'> Marines are overpowered in scrims/matches
Marines are underpowered in pub play.

Why? In pubs the skill level is generally ALOT lower. <img src='http://www.unknownworlds.com/forums/html//emoticons/sad.gif' border='0' style='vertical-align:middle' alt='sad.gif' />  </td></tr></table><div class='postcolor'> 
Its a bit different CWAG. In Pubs with a semi-intelligent commander, marines are overpowered out the **** - even pretty shoddy teams can win. In scrims/matches they're mildly overpowered :/

Wirhe · May 2004

75% of people play on pubs/pugs/bugs, so why should we even care of clan-players? IMO, they should just aim it for the public, not compromise it for that minority.

SaltzBad · May 2004

Because the people that care about balance are competitive. The people that care about gameplay are pubbers/uncompetitive. Saying one has to rule the other out is nonsense - gameplay is the foundation of any game becoming popular, and balance is just a pretty word for long-term depth.

The_Noid · May 2004

Those statistics miss one very important aspect... A cutdown by number of players in a game.

Hydra · May 2004

What NS really needs is a universal stat system that would record who won each game, time played that round, number of players, kill:death ratio average for each team, number of players, and I am probably overlooking something else that would be useful.
In short, concrete data would help balance a lot imo, and if more servers posted stats from their logs like this it would help probably.

BobTheJanitor · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (TrueChaos @ May 19 2004, 10:49 PM)</td></tr><tr><td id='QUOTE'> What NS really needs is a universal stat system that would record who won each game, time played that round, number of players, kill:death ratio average for each team, number of players, and I am probably overlooking something else that would be useful.
In short, concrete data would help balance a lot imo, and if more servers posted stats from their logs like this it would help probably.  </td></tr></table><div class='postcolor'> 
OMG BUS!

Mintman · May 2004

Yep, that's what my money is on for what the BUS is.

bLuISh · May 2004

we CANNOT determine anything from those statistics...

here is my reason:
alot of pubbers tend to go to the same server again and again yes?

meaning player A and player B will probably play on the same server...so lets put this together..

Player A is really good at ns, and player B isnt quite as good. Player A will win most games agaisnt player B...now...

Knowing that people like to come back to the same server (i know i do)
we can expect player A and player B to come back again and agian and play over and over...what does this mean?

we get a statistic based on the same matchup of players which proves player A is clearly better than player B....but does it mean aliens are better at the moment? No...

if you really wanted to have accurate statistics, id get logs from MULITIPLE servers and combine the results completely, only then will you have a good estimate with wins/losses

<img src='http://www.unknownworlds.com/forums/html//emoticons/tounge.gif' border='0' style='vertical-align:middle' alt='tounge.gif' />

Mintman · May 2004

Yes, this does have a limited scope, but it's much better than going "omg aliens win <pick random number from the air>% of games, NERF NERF NERF!". I'd also argue that it does provide a reasonably fair insight as it is over quite a long period of time which makes it more likely that the teams were of similar skill on average.

Edit: I've also been working on something for parsing these types of logs so that you could throw multiple files at it. It's written in java. Is there any point in my continuing my work and if anyone is interested what sort of outputs and ways of analysing these logs would you like to see.

I've come up with: All logs together, By log file, By map. Giving stats like the number of alien wins, the number of those which were due to a timeout and the average round time.

Niteowl · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (Mintman @ May 20 2004, 09:47 PM)</td></tr><tr><td id='QUOTE'>
Edit: I've also been working on something for parsing these types of logs so that you could throw multiple files at it. It's written in java. Is there any point in my continuing my work and if anyone is interested what sort of outputs and ways of analysing these logs would you like to see.

I've come up with: All logs together, By log file, By map. Giving stats like the number of alien wins, the number of those which were due to a timeout and the average round time.  </td></tr></table><div class='postcolor'> 
you have looked at PsychoStats right? I think it does pretty much everything you are desiging, atm. sorry to be a rain on yer parade and all.

funka: nice work <img src='http://www.unknownworlds.com/forums/html//emoticons/smile.gif' border='0' style='vertical-align:middle' alt='smile.gif' />

zzSleeper · May 2004

this wasn't mentioned in Funka's original post, but there is some extra important data available for the stats. the server is configured to go play ns_ maps if there are 12 or more players and co_ maps if there are less than 12. this means that 95% of the ns_ games played should have between 12 and 18 players, so those stats should be pretty reliable.

Ahnteis · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (zzSleeper @ May 22 2004, 08:46 AM)</td></tr><tr><td id='QUOTE'> this wasn't mentioned in Funka's original post, but there is some extra important data available for the stats. the server is configured to go play ns_ maps if there are 12 or more players and co_ maps if there are less than 12. this means that 95% of the ns_ games played should have between 12 and 18 players, so those stats should be pretty reliable.  </td></tr></table><div class='postcolor'> 
No, that means that those games STARTED with that many players. How many players dropped within the first minute (because they joined the server to play NS:C not NS) is not accounted for.

zzSleeper · May 2004

</div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (Ahnteis @ May 22 2004, 04:06 AM)</td></tr><tr><td id='QUOTE'> No, that means that those games STARTED with that many players. How many players dropped within the first minute (because they joined the server to play NS:C not NS) is not accounted for.  </td></tr></table><div class='postcolor'> 
as i just stated, "this means that 95% of the ns_ games played should have between 12 and 18 players". this takes into account dropping - thus only 95%. when the server is warming up in the afternoon, it will be on co_. once there are enough players it will switch to ns_ and stick to ns_ until it winds down late in the evening. the first round or two may have less than twelve, but the server quickly fills up.

Interesting, Actual Win/loss Beta4a Statistics

Comments