Server Statistics
Server Statistics
I took myself to the bathroom, pulled a face much like dize in his avatar and I birthed some server statistics for the V1 server. Starting from yesterday and updating every 24h, behold the Server Statistics
Last edited by wonkey on 22 May 2007 17:22, edited 1 time in total.
Anybody wondering how the skill column is calculated, look no further!
It is a calculation based on who you've killed or who killed you. I believe everyone starts with a skill of 1000.
MACHINEGUN = 1.0
SHOTGUN = 1.0
GRENADE = 1.5
GRENADE_SPLASH = 1.0
ROCKET = 1.25
ROCKET_SPLASH = 1.0
LIGHTNING = 1.25
PLASMA = 1.25
PLASMA_SPLASH = 1.0
RAILGUN = 1.5
Another more realistic example for anyone still reading ...
It is a calculation based on who you've killed or who killed you. I believe everyone starts with a skill of 1000.
GAUNTLET = 1.5PlayerA kills PlayerB with a rocket launcher
Take the victim, PlayerB and work out a 1/1000 of his current skill level. First ever kill, so it is 1000 / (1/1000) = 1
PlayerB loses that skill bit, so becomes 1000 - 1 = 999
PlayerA gains that skill bit multiplied by a weapon skill factor (see below). Rockets have a factor of 1.25, so 1000 + (1 * 1.25) = 1001.25
MACHINEGUN = 1.0
SHOTGUN = 1.0
GRENADE = 1.5
GRENADE_SPLASH = 1.0
ROCKET = 1.25
ROCKET_SPLASH = 1.0
LIGHTNING = 1.25
PLASMA = 1.25
PLASMA_SPLASH = 1.0
RAILGUN = 1.5
Another more realistic example for anyone still reading ...
The main problem with this is that unless you are a shit mofo who dies almost all the time you will be gaining some skill points. So it unfairly rewards players who play hundreds of games compared to those who only play say 100. IMO it's not an accurate judge of skill unless we tweak it in relation to how many games you've played.Wonk Skill = 1257
PJ Skill = 1029
PJ was pummelled to death by Wonk
1029 / (1/1000) = 1.029
PJ Skill = 1029 - 1.029 = 1027.71
Wonk Skill = 1257 + (1.029 * 1.5) = 1,258.5435
OK, will be doing some tweaking of the stats over the next weeks to try and remove the advantage gained by playing loads of games. More bias is placed on getting a good kill/death ratio as stuff like kills per game and kill streak is very dependent on how busy the server is when you're playing on it.The main problem with this is that unless you are a shit mofo who dies almost all the time you will be gaining some skill points. So it unfairly rewards players who play hundreds of games compared to those who only play say 100. IMO it's not an accurate judge of skill unless we tweak it in relation to how many games you've played.
Well arent the stats for people who play regular anyways, rather the people who play a few games a week.
If you made your changes w o n k e y wouldnt you get people with good kill death ratios and effitancy at the top , when they have only played a few games against a server full of noobs or whne the server is less busy.
Just looking at the stats now i see you have changed them.
Now fabri is top despite only plaing 28 games and if you look at the games hes played most of them was when no one was on the server.
Its much harder getting a good kill death ratio when you play durring the peak times between 6-10 pm.
The only people in the top 5 who deserve to be there imo are yellow and gravi who i see regulary play durring the peak times. Now the stats favor people who play less and at less busy times.
More tweaking needed i feel.
If you made your changes w o n k e y wouldnt you get people with good kill death ratios and effitancy at the top , when they have only played a few games against a server full of noobs or whne the server is less busy.
Just looking at the stats now i see you have changed them.
Now fabri is top despite only plaing 28 games and if you look at the games hes played most of them was when no one was on the server.
Its much harder getting a good kill death ratio when you play durring the peak times between 6-10 pm.
The only people in the top 5 who deserve to be there imo are yellow and gravi who i see regulary play durring the peak times. Now the stats favor people who play less and at less busy times.
More tweaking needed i feel.
keep the old stats! They are good, yes you gain score by playing a lot but this will stop climbing after a while and stop at a certain level.
Also think that to much emphacies on kill/death ratio is not good since if you play few games against noobs you will have very high score.
to make it possible to compare with players who played different number of games maybe this formula:
add all players you killed muktiplied by their score and divide this with the "skillsum" of the players who killed you
the problem here it is that the skill value difference betwwen players should be adequate, i believe top players before had 1500 in score and noobs 1000, this makes only 50% skill difference which is not true. Earning more points for every kill would solve this but also increase influence of number of games played.
as arnott wrote, the stats are for those who plays a lot, keep the old stats, the efficieny column is still there.
Also think that to much emphacies on kill/death ratio is not good since if you play few games against noobs you will have very high score.
to make it possible to compare with players who played different number of games maybe this formula:
add all players you killed muktiplied by their score and divide this with the "skillsum" of the players who killed you
the problem here it is that the skill value difference betwwen players should be adequate, i believe top players before had 1500 in score and noobs 1000, this makes only 50% skill difference which is not true. Earning more points for every kill would solve this but also increase influence of number of games played.
as arnott wrote, the stats are for those who plays a lot, keep the old stats, the efficieny column is still there.
VSP stats does all the adding of skill rating to killer and subtracting from victim when it parses the game log files. I have not and will not be changing that, that's how it works period.
Yes, the higher your skill the more you lose when someone kills you. Again, that's how it works and I'm not touching it. Your skill can always increase though because depending on what weapon you kill them with your skill rating increases by 1.0 - 1.5 the amount you steal off the victim.
The reason I'm tweaking things is mainly because I appeared so high in the list. I had a below 50% efficiency and was ranked 13 purely because I'd played twice as many games as some others. Now that's not right, is it? I only play with one name so assuming I kept playing at double the rate of everyone else it wouldn't be long before I end up in the top five.
There is no right solution though, hence the reason for experimenting (maybe we'll go back to default after). He's what I've done so far ...
Instead of just getting skill from the database it's tweaked as follows and only shows people with 20 or more games.
I think the current board is quite representative of a players skills. Fabri is at the top, but look at his record, he deserves to be there. As more games are played I think the cut off period should increase though, so unless Fabri keeps playing he will soon not be displayed at all because he's not played enough.
Yes, the higher your skill the more you lose when someone kills you. Again, that's how it works and I'm not touching it. Your skill can always increase though because depending on what weapon you kill them with your skill rating increases by 1.0 - 1.5 the amount you steal off the victim.
The reason I'm tweaking things is mainly because I appeared so high in the list. I had a below 50% efficiency and was ranked 13 purely because I'd played twice as many games as some others. Now that's not right, is it? I only play with one name so assuming I kept playing at double the rate of everyone else it wouldn't be long before I end up in the top five.
There is no right solution though, hence the reason for experimenting (maybe we'll go back to default after). He's what I've done so far ...
Instead of just getting skill from the database it's tweaked as follows and only shows people with 20 or more games.
Code: Select all
(1000 + ((skill-1000)*100/games)*(kills/deaths))
Well i still dissagree. Gabri is top cuase he played no one in top 10 other then me and he struggled against me. Hes up there cuase hes played off peak times.
Like i said be4 you have punished people who play more and rewarded people who play less. If u look i'm the only person even close to 100 games in top 10 is me.
Most v play durring peak times where its usally at least 4v4 or 5v5 Pj for example only really see him playing durring them times.
During them times you know your self is hard to get a effiancy above 50% for most players and your chnaces of dying are greatly increased
So all people have to do now to get to top of the stats
1) avoid playing durring the busy time
2) aviod playing on small cramped maps to reduce chances of been killed by spam. Unless noobs on the server
3) When server has noobs on it keep playing and keep winning
4) Constantly try to use rail gun to kill people for more skill.
Like i said be4 you have punished people who play more and rewarded people who play less. If u look i'm the only person even close to 100 games in top 10 is me.
Most v play durring peak times where its usally at least 4v4 or 5v5 Pj for example only really see him playing durring them times.
During them times you know your self is hard to get a effiancy above 50% for most players and your chnaces of dying are greatly increased
So all people have to do now to get to top of the stats
1) avoid playing durring the busy time
2) aviod playing on small cramped maps to reduce chances of been killed by spam. Unless noobs on the server
3) When server has noobs on it keep playing and keep winning
4) Constantly try to use rail gun to kill people for more skill.
Yes, that's stats for you. They don't care when you play or who you played against, they just record numbers.Arnott wrote:1) avoid playing durring the busy time
2) aviod playing on small cramped maps to reduce chances of been killed by spam. Unless noobs on the server
3) When server has noobs on it keep playing and keep winning
4) Constantly try to use rail gun to kill people for more skill.
The stats have now been reverted to default, but in a months time I'm sure we'll be seeing the trend for those who play many games being overrated.
Now that the stats how been reverted i see what you mean now about the stats. I guess both ways have there flaws.
Still like default way better though cause i am top !!!
But seriosuly though i just didnt feel it was right a guy who play 20 or so games been top.
Is there no way to like mix beteen the 2.
But suppose at the end of the day theres always the seprate collums that can be clicked to see where you rank in each section
Still like default way better though cause i am top !!!
But seriosuly though i just didnt feel it was right a guy who play 20 or so games been top.
Is there no way to like mix beteen the 2.
But suppose at the end of the day theres always the seprate collums that can be clicked to see where you rank in each section
- YeLLoW
- Site Admin
- Posts: 976
- Joined: 18 May 2007 10:37
- Location: Bratislava, Slovak Republic
- Contact:
wonk, how the stats deal with ppl, who disconnect before map's end, or connect few round after beginning? Counting that as a 1 game? or you have to play whole map?
And what about spectating?does it counts?
And the last question. If i rejoin from blue team to red team, ingame stats at the end shows me only numbers after i rejoined. not the whole map. its the same here? Probably not if the stats are taken from logs, they will be accurate in this way. am i wrong?
And what about spectating?does it counts?
And the last question. If i rejoin from blue team to red team, ingame stats at the end shows me only numbers after i rejoined. not the whole map. its the same here? Probably not if the stats are taken from logs, they will be accurate in this way. am i wrong?
- YeLLoW
- Site Admin
- Posts: 976
- Joined: 18 May 2007 10:37
- Location: Bratislava, Slovak Republic
- Contact:
hmm that sucks a bit. coz that makes the stats inaccurate / is there any way to fix it?wonkey wrote:If you join a map (as spectator or player) I think that counts as you have played one game. Even if you disconnect 10 seconds later.
If you change team part way through the stats will still take into account the previous team and the new team.
Last edited by YeLLoW on 29 Jun 2007 15:36, edited 1 time in total.
ICEEEEEEEEEEEEEEEEEEEEEEEEEE!!!!!!!!!!
hi all!! can I to you add ?? pley mostly on server Freze Crewt ip 85.131.247.150:26960 greet all!!