I have a feeling the girlfriend statistics page would be very empty indeed

so was your mom when she gave birth to you

The rest of us are just pretending that we don't know eri is xou's mother.

ed1 or.. are we?
ed2 oh yeah i forgot, eri in lower case
ed3 correcting ed 2 (with -> in)
ed4 link to girlfriend section?
ed5 girlfriend's -> girlfriend
ed6 shit, xou in lower case too.
ed7 TBD

Yeah , i have several more offspring on here, unfortunately in most cases

And no that's not a personal dig but rly dude, while this is i guess mildly amusing do you really not have better things to do :o

;)

i know that this is quite unnecessary and a bit over the top but it was far more fun than doing what i should have done... studying for hours... and i actually always wanted to try out/learn to use the open office calculator so it fit perfectly. my inner nerd is satisfied.

The old modern society paradox, people choose to study one thing then spend their time learning another thing.

extrinsic motivation vs. intrinsic motivation

another victim of gaming

some great procrastination

This is really not that interesting. Stats about quad runs or different weapons could be though.

statistics about powerups would be awesome, but there is no manageable way to get the data for that :/

kewl man

Tonight's xt vs iM's dm14 game will overwrite it, 87-65. : )

hehe yeah, by far the lowest scoring dm14 win in all major competitions so far

? are you kidding me ? what about ta's win over 4K there with 1 frag difference of last cup if i remember well ? :p

this? http://static.clanbase.com/CB/screenshots/10540829_1.jpg
not that 97:94 isnt a low score or close, but this was definitely lower scoring

Heh, remember this yawnfest?
http://static.clanbase.com/CB/screenshots/178972_4.jpg

Hello. He wasnt even talking about lowest gap score. :))))))))

Ha yeah. I see what he meant.

It was late.

and dark :x.


edit: but

lowest scoring win

can be misinterpreted tho :P

I have played in slower feeling games. I was surprised at the score in the end.

same, i felt like it was much more action packed. probably best dm14 i've played :D

analysis demonstrated shit

and by that you mean what?
im saying that there is not too much to be read into the stats but i tried to point out the few things that can be seen

I guess u translated the data the way u want to see it.

Someone could translate it very differently. For example, the big average score difference could be an indication of how flawed is a map. The maps with big ones could mean that these maps are unbalanced and favour the team in control. A map with red mega and quad at the same room could probably lead to average score difference of 150+. That won’t be an indication that the map is good, but the exact opposite.

I am not a tdm player, but I am sure u can interpret the results in many ways and you probably choose the one that u like and/or came first into ur mind.

That being said, very nice work once again.

The maps with big ones could mean that these maps are unbalanced and favour the team in control.

They should favour the team in control..

yes, but not too much like in my example

Why not? Even if the team in control gets 5x the frags, control will shift and the other team can rape.

lol, just not

actually having a larger score difference could also indicate a "better" map. It could determine that either teamplay or "skill" is of more importance and so theres room for teams to seperate themselves from the rest. Where a close scoring map could mean the map is impossible to be anything other then regular on.

hiddenfortress ftw :>

i am very open to other ways to see the data, but his post wasnt exactly adding anything to the interpretation^^
i mean i even encourage people to have a look at the data and see what they come up with, ou could compare the development between tournaments or so aswell...

what you described is pretty much good old duel skepticism about maps being too easy to control but it doesnt really apply to tdm like this imho.

nice work nukm :D

Can you turn this data into an online app where one can upload some result scores, and the app tells if it is pure ownage or if for that map it is normal to have such scores difference?

that would be cool, and although i can code a bit i know shit about web stuff and this works just with what excel/open office calc offer

it would be awesome to have a really open fancy db where you can add all kinds of match results and details and everyone could do custom db queries

google docs comes with excel alike things and a form creator to automaticalls post data into the excel. That could be an easy way.

MORE STATISTICS WITH SPSS! Is the difference even statistically firm? :)

it would be interesting to see what numbers have to say about "starting control" or should i say "control flow"... in games where opponents are evenly skilled the control will definitely shift and im guessing that its highly probable that someone has indeed better chances of winning because of control flow, and who starts in control... player A who is in control from the start has better chances of winning than player B who starts on the back foot. Maybe its the other way around and player B is in a better position in a 10min game if the control switches 3 times and he is in control in the last minutes of the game when it really matters...

Starting spawns, that are unfair lots of times, decide who starts in control and control flow statistics could show that its really a bigger influence on who wins the game then is thought of now...

I like this. Call it "Frag Actuarials." nj Nukm.

you should work for aperture science!

Nice stuff, nukm!

Statistics are a source of endless waste of time.

Here the output of a game from aaa, long time ago: http://xx.xx.xx.xx/aaa-ta-39.txt

I wish ql had something like this, ofc in a nicer outfit :)


/blau


[Edit: Server moved: http://78.47.168.130/aaa-ta-39.txt]

That is brilliant, just needs a more pleasing/clear way of displaying it.

well first of all we need ql to give us the info at all ;)

Exactly. No more OSP server logs with ql :-(
Every health/armor change got logged.

/blau

yeah, originally that was meant as a backend, but no one made a UI for it :)

/blau

this has little meaning if any at all, your analysis is biased and flawed. still a good idea in principle.

0,17 - Campgrounds vs. 0,39 - Hidden Fortress

= more people know how to play pro6, in comparison lots of people still don't know enough/care about dm20.

0,35 - Deep Inside vs. 0,24 - Grim Dungeons

= some maps end up being more blowout friendly than others, but you aren't actually trying to tell us that osp5 is superior as a map to dm14tmp because of that?

there's another problem with those stats overall, some tdm games end up being decided minutes before the buzzer goes, the team in the lead will let loose and the score gap closes (also depending on how open a map is). that does not mean anything in terms of how 'good' a map is though.

0,17 - Campgrounds vs. 0,39 - Hidden Fortress

= more people know how to play pro6, in comparison lots of people still don't know enough/care about dm20.

yes, thats what i mentioned aswell
"and adding the sheer lack of practice for teams on the map the lead is quite questionable"

some maps end up being more blowout friendly than others, but you aren't actually trying to tell us that osp5 is superior as a map to dm14tmp because of that?

well the idea behind the ranking is, that on a good map, the better team should get a "clear win". if a map does not allow the better team to separate each other from the worse team then it is a problem. (its not the only criteria when judging a map but imho it is one, dont you think so?)
i dont want to say that dm14 is a worse map, but in this regard it seems to be worse than osp5, yes

there's another problem with those stats overall, some tdm games end up being decided minutes before the buzzer goes, the team in the lead will let loose and the score gap closes (also depending on how open a map is). that does not mean anything in terms of how 'good' a map is though.

if a map is decided, the score gap closes? would rather say that the losing team usually gives up then and the score gab gets bigger then under normal circumstances. in any way that is only a problem for stats if it happens more often on some maps then on other (that might very well be the case, but i dont think its that big of a deal), cause otherwise the statistical error should cancel itself out

well the idea behind the ranking is, that on a good map, the better team should get a "clear win". if a map does not allow the better team to separate each other from the worse team then it is a problem. (its not the only criteria when judging a map but imho it is one, dont you think so?)
i dont want to say that dm14 is a worse map, but in this regard it seems to be worse than osp5, yes

But your ranking does not determine whether the better team won, it only shows whether victories are determined by a large margin or a small one. Imagine a map where there are two rooms, one has all the spawn points and the other has every weapon and item. Provided one team manages to get full control at some point, the other team will likely lose and one can expect the scores to be something ridiculous like 300-30. Would this really be a good map?

You're not determining if the better team wins on a map, you're not determining why a team wins on a map. What excactly are you determining that is of any usefulness in gauging how good a map is?

Now that I thought about it, you could actually use the statistics you gathered and analyse whether better teams have maps that they tend to win against worse teams more consistently than other maps. It would make sense that these more consistent maps are "better" in terms of skill. (But you should pool more than one top team into the comparison, because a team might just be better on some maps due to practising it more)

first of all i wanna mention that i did the ranking because rankings are fun and it usually brings some discussion. i was and am aware that it has its flaws but i am defending it here cause i dont think they make the ranking completely invalid.

But your ranking does not determine whether the better team won, it only shows whether victories are determined by a large margin or a small one. Imagine a map where there are two rooms, one has all the spawn points and the other has every weapon and item. Provided one team manages to get full control at some point, the other team will likely lose and one can expect the scores to be something ridiculous like 300-30. Would this really be a good map?

certainly not, but the question is: does this apply to the maps in the pool?
the way i see it, every map that is getting played needs a really strong setup to hold the favorable area and/or the other half of the map offers enough resources to at last compete for the next round of powerups. in duel for example i think this would be way bigger problem.

You're not determining if the better team wins on a map

hmm... fair point, i am doing it the other way, defining the winning team as the better team (the team that played better on the map) thats not always true but in statistical sense it should be fine

Now that I thought about it, you could actually use the statistics you gathered and analyse whether better teams have maps that they tend to win against worse teams more consistently than other maps. It would make sense that these more consistent maps are "better" in terms of skill. (But you should pool more than one top team into the comparison, because a team might just be better on some maps due to practising it more)

interesting idea, but seeing how differently well teams have mastered certain maps i am not sure if you will get more meaningful results in the end. but i am thinking about linking clans to the scores sometime, might have a look at this then.

hmm... fair point, i am doing it the other way, defining the winning team as the better team (the team that played better on the map) thats not always true but in statistical sense it should be fine

This is not correct reasoning, even in a statistical sense. Imagine two teams (be they equally skilled or not) playing a map 20 times that is imbalanced to the point where the first spawns always decide the game. Statistically speaking, you should get both teams winning 10 maps by a large margin. Your methodology would result in showing that this is a good map, when it's actually biased by luck. You're making a _HUGE_ assumption that things like this do not happen and I have doubts that it's even valid.

interesting idea, but seeing how differently well teams have mastered certain maps i am not sure if you will get more meaningful results in the end. but i am thinking about linking clans to the scores sometime, might have a look at this then.

Yes, that's why I said pool the data. Assume you exogenously determine whether a clan is, for instance better than average or worse than average (by perhaps looking at how the clan does in final standings). Then if you compare e.g. all better clans vs all worse clans, you should arrive at a general win rate for better clans vs worse clans, for example 60%. It stands to reason that any map that better clans win more than 60% of the time is better than the other half of the maps. The only way I could see results being biased is if the top clans practise some maps (e.g. new ones) relatively less than lower skilled clans. This might actually be true, but it's a much smaller assumption than "there are no blowout maps that are biased by luck in pool" and the effects are more limited. You could also use statistical confidence tests to see whether your results have statistical merit.

(edit) gah hit reply by accident before I had finished

I've been doing the econometrics part for my master's thesis for the past few weeks, so my brain is on statistics mode atm :P

I concur with what has been said earlier, you cannot tell how good a map is by the scorelines. A more qualitative analysis would be possible if you were able to extract all the individual item pickups and frags in a time series to see how control is related to frags and how easy it is to keep/lose control, but this analysis off summarized data doesn't really show anything.

I think this would be really nice; would be really interesting to compare these stats of different teams, and during key periods of a game as well. I presume this is the sort of thing which could be programmed into wolfcam?

You must refer to a quantitative analysis, with all that statistics you are suggesting.

In my humble opinion, you need to look at this topic in a more qualitative perspective, and simply ask people what map they enjoy to play, and think is good.

No matter how much statistics you gather, you cant decide that this is better because x > y, way to many variables.

I've always been of the opinion that a collective of players will reach a much better opinion of how good a map is than someone computing statistics, because emotions can't really be quantified.

I'm just saying if you want to get quantitative and deduce something from data, you should make sure your methodology is correct in the first place, otherwise it's just garbage in-garbage out while sounding vaguely legitimate.

This is a pathetically blatant attempt to get pluses...

[+] if you agree

ROFL fail for ya own plus!!!!!!!!!!! @ 50