new ladder formula
Moderator: Verement
new ladder formula
I think the ladder scoring should be revised. As I recall, there is a control for the percentage of your score that gets thrown into the pot for others to capture. In the very least, that percentage could be lowered. The ladder would change more slowly. If I was to make this change, I would zero out the ladder at ish so everyone starts at the same place. I think I can zero the ladder without removing/changing the rounds/matches won. Maybe the ladder should be zero'd on a monthly basis and the winners at the end of the month posted on the stats pages.
I would appreciate everyone's thoughts on this. I'll look at the settings to see what can be controlled with the configs. Would starting the ladder fresh every month be something you want? Each month's winner would be logged and posted on the stats pages. Thoughts?
I would appreciate everyone's thoughts on this. I'll look at the settings to see what can be controlled with the configs. Would starting the ladder fresh every month be something you want? Each month's winner would be logged and posted on the stats pages. Thoughts?
Re: new ladder formula
I agree that it should be modified, and I'm going to look into this over the weekend.ishAdmin wrote:I think the ladder scoring should be revised. As I recall, there is a control for the percentage of your score that gets thrown into the pot for others to capture. In the very least, that percentage could be lowered.
Yes, you can modify the ladder without affecting the other stats. You can do this manually, or if you're in linux, you can write a bash shell script to delete the old ladder file and "touch" a new ladder file and schedule this script to run automatically.If I was to make this change, I would zero out the ladder at ish so everyone starts at the same place. I think I can zero the ladder without removing/changing the rounds/matches won.
This is a good idea. If we change the ladder system so that it's harder to move up and down, then there is a possibility that some elite players will gain so many points that there would be no hope of catching up. By clearing the ladder monthly, new players to the game would have a shot at climbing the ladder.Maybe the ladder should be zero'd on a monthly basis and the winners at the end of the month posted on the stats pages.
Will deleteing and touching the ladder file cause a problem if you do not also stop and start the game? You certainly shouldn't do that to http logs without stopping the apache daemon. If it's necessary to stop and start the game, I guess I'd have to grep the pid's. Not a biggie, just wondering if it's necessary.
Two ladders huh. Interesting. Maybe instead of armagetron keeping track of two ladders, maybe they could be combined externally using a bash script. It could be as simple as averaging all the monthly ladders.
Two ladders huh. Interesting. Maybe instead of armagetron keeping track of two ladders, maybe they could be combined externally using a bash script. It could be as simple as averaging all the monthly ladders.
It shouldn't, only because the calculation of the ladder takes a fraction of a second and you're only switching ladder files once per month. If it were more often, then there would be a problem. However, it's possible that in the brief moment the replacement occurs Armagetron could create its own ladder file between the time the ladder file is deleted and the new one is created. Armagetron will naturally create a new ladder file if none exists. In fact, you could probably just delete the old one and not have to worry about touching a new one. Let Armagetron take care of it.Will deleteing and touching the ladder file cause a problem if you do not also stop and start the game?
This is a good idea. =)Two ladders huh. Interesting. Maybe instead of armagetron keeping track of two ladders, maybe they could be combined externally using a bash script. It could be as simple as averaging all the monthly ladders.
- Tank Program
- Forum & Project Admin, PhD
- Posts: 6711
- Joined: Thu Dec 18, 2003 7:03 pm
OK, I'd like some opinions on this. The basic controls are as follows:
LADDER_PERCENT_BET x percentage of your score to be put in the pot
LADDER_MIN_BET x minimum credits to be put in the pot
>>So, a 'what if' - what if the percentage thing was changed to 0, the minimum bet was the only amount put in the pot? Better players would not risk more than lesser players. This would cause those who played often to rise higher if they played well. A less frequent player who wins over the best would not gain very much. Not sure if I like it this way, but I'd like to know what you think.
LADDER_TAX x percentage the IRS takes from the pot
>>Do we need any ladder tax? I don't think so. It's not like it represents money to me!
LADDER_LOSE_PERCENT_ON_LOAD x you loose this percentage of your score every time the server is restarted
LADDER_LOSE_MIN_ON_LOAD x but minimum this value
>>This is sort of the thing 2 was asking for. If I understand it correctly, restarting the server daily would lower everyone's score daily. Correct me if I'm wrong. Would this be a good thing? If the stats are reset every month, there is less of a need for this in my opinion. If the stats are continuous (forever), then this could be used to stop stat whores from filling up the top spots and then dropping that name.
LADDER_GAIN_EXTRA x the winner gets his ping+ping charity (in seconds) times this value extra
>>I personally think this is a good idea, but it can be easily abused. For instance, it's possible to jack your ping just when this calculation is going to happen. You would get extra ladder points by cheating. Making it zero would stop that, but then higher ping players aren't compensated for their unfortunate circumstances. Low ping players have an obvious advantage over everyone else, should this used to even the ladder out a little?
What does everyone think?
LADDER_PERCENT_BET x percentage of your score to be put in the pot
LADDER_MIN_BET x minimum credits to be put in the pot
>>So, a 'what if' - what if the percentage thing was changed to 0, the minimum bet was the only amount put in the pot? Better players would not risk more than lesser players. This would cause those who played often to rise higher if they played well. A less frequent player who wins over the best would not gain very much. Not sure if I like it this way, but I'd like to know what you think.
LADDER_TAX x percentage the IRS takes from the pot
>>Do we need any ladder tax? I don't think so. It's not like it represents money to me!
LADDER_LOSE_PERCENT_ON_LOAD x you loose this percentage of your score every time the server is restarted
LADDER_LOSE_MIN_ON_LOAD x but minimum this value
>>This is sort of the thing 2 was asking for. If I understand it correctly, restarting the server daily would lower everyone's score daily. Correct me if I'm wrong. Would this be a good thing? If the stats are reset every month, there is less of a need for this in my opinion. If the stats are continuous (forever), then this could be used to stop stat whores from filling up the top spots and then dropping that name.
LADDER_GAIN_EXTRA x the winner gets his ping+ping charity (in seconds) times this value extra
>>I personally think this is a good idea, but it can be easily abused. For instance, it's possible to jack your ping just when this calculation is going to happen. You would get extra ladder points by cheating. Making it zero would stop that, but then higher ping players aren't compensated for their unfortunate circumstances. Low ping players have an obvious advantage over everyone else, should this used to even the ladder out a little?
What does everyone think?
I like the idea of making it harder to move up and down the ladder, but there is one thing that I'm still not crazy about. It seems the current ladder formula relies only on round wins.
So given this scenario:
Player 1 and Player 2 are tied for 20th place on the ladder and each play separate rounds.
Player 1's round
3 players in the round: Player 1, Player 3 (rank 30), and Player 4 (rank 40)
Player 1 wins the round, but gets no kills
Player 2's round
8 players in the round Player 2 and Player 5-11
Player 2 loses the round, but gets 3 kills, Player 5 (rank 5), Player 6 (rank 6),
and Player 7 (rank 7)
Player 1 gains points in the ladder and Player 2 loses points. In my opinion, this is unfair.
Let's look at stats in a larger scope. 100 rounds are played.
Player 1
60 round wins
30 kills
Player 2
10 round wins
120 kills
It's obvious that these 2 players have different playing styles. Player 1 is almost definitely going to be much higher in the ladder if they started at the same position,
but I think that there is an chance that either is going to be "better" than the other. If they played 1v1, I would put my money on Player 2.
Solution?
Kills should hold some sort of weight (especially since matches are determined by kills not round wins) and in my opinion, should hold more weight than round wins.
LADDER_PERCENT_BET could be split into:
LADDER_KILL_PERCENT_BET
LADDER_ROUND_PERCENT_BET
Each kill would add to your ladder score based on the level of the player that you dumped.
I'm not sure about suicide, maybe a tax based on your current ladder score ... oh, I guess you would just lose your bet, and no one would get it.
So given this scenario:
Player 1 and Player 2 are tied for 20th place on the ladder and each play separate rounds.
Player 1's round
3 players in the round: Player 1, Player 3 (rank 30), and Player 4 (rank 40)
Player 1 wins the round, but gets no kills
Player 2's round
8 players in the round Player 2 and Player 5-11
Player 2 loses the round, but gets 3 kills, Player 5 (rank 5), Player 6 (rank 6),
and Player 7 (rank 7)
Player 1 gains points in the ladder and Player 2 loses points. In my opinion, this is unfair.
Let's look at stats in a larger scope. 100 rounds are played.
Player 1
60 round wins
30 kills
Player 2
10 round wins
120 kills
It's obvious that these 2 players have different playing styles. Player 1 is almost definitely going to be much higher in the ladder if they started at the same position,
but I think that there is an chance that either is going to be "better" than the other. If they played 1v1, I would put my money on Player 2.
Solution?
Kills should hold some sort of weight (especially since matches are determined by kills not round wins) and in my opinion, should hold more weight than round wins.
LADDER_PERCENT_BET could be split into:
LADDER_KILL_PERCENT_BET
LADDER_ROUND_PERCENT_BET
Each kill would add to your ladder score based on the level of the player that you dumped.
I'm not sure about suicide, maybe a tax based on your current ladder score ... oh, I guess you would just lose your bet, and no one would get it.
The specifics of lag charity are a bit blurry to my, but can't I change my laf charity to an obscenely high level and still not have it affect my playing? If that's the case, it seems that this could be easily abused.ishAdmin wrote:LADDER_GAIN_EXTRA x the winner gets his ping+ping charity (in seconds) times this value extra
>>I personally think this is a good idea, but it can be easily abused. For instance, it's possible to jack your ping just when this calculation is going to happen. You would get extra ladder points by cheating. Making it zero would stop that, but then higher ping players aren't compensated for their unfortunate circumstances. Low ping players have an obvious advantage over everyone else, should this used to even the ladder out a little?
What does everyone think?
I think the bet should be constant and transparent to the the player. It seems to me like this is just one way that a player can manipulate the ladder. If someone wants to goof off for a few rounds, then they can easily just change there name, like eggcozy at play.ishAdmin wrote: LADDER_PERCENT_BET x percentage of your score to be put in the pot
LADDER_MIN_BET x minimum credits to be put in the pot
Hmm. Reset every month? A months seems pretty quick for someone to drop out of the ladder completely. I think ideally each "bet" that you win should decay separate of other bets (seems like it would be very hard to implement). So if I scored 10 points for dumping n54, each day afterwards, a small tax is taken out, so after a couple months, that 10 might be down to 5. The server restart tax seems like an easy compromise.ishAdmin wrote: LADDER_LOSE_PERCENT_ON_LOAD x you loose this percentage of your score every time the server is restarted
LADDER_LOSE_MIN_ON_LOAD x but minimum this value
>>This is sort of the thing 2 was asking for. If I understand it correctly, restarting the server daily would lower everyone's score daily. Correct me if I'm wrong. Would this be a good thing? If the stats are reset every month, there is less of a need for this in my opinion. If the stats are continuous (forever), then this could be used to stop stat whores from filling up the top spots and then dropping that name.
Hum.... I maybe will say some stupidyties.... but :
WHy not tobe more simple?
There are 2x2 possibillities in this game :
-I win round
----i win but don't kill
----i win and kill
-I died
----i died but kill
----i died and dont kill
then they are 3 action for a player :
Win round
Lose
Kill another player
ok?
Why not give points like this :
-Round victory : XXpoints
-Kill another plyer : Xpoints
-Lose : -Y points
(i'm feeling really stupid.... (does someone told this earlier, and i didn't red it?...argh)
Of course, he ladder system should be abandonned in this case.
In extension, it could be wonderfull, if, a noob, could win extra points, when he kill a better player... or if ladder stayes, just take ladder from this player.
u see?
(sorry, i go out.... i feel dumb...hummm (better, for graphics...))
WHy not tobe more simple?
There are 2x2 possibillities in this game :
-I win round
----i win but don't kill
----i win and kill
-I died
----i died but kill
----i died and dont kill
then they are 3 action for a player :
Win round
Lose
Kill another player
ok?
Why not give points like this :
-Round victory : XXpoints
-Kill another plyer : Xpoints
-Lose : -Y points
(i'm feeling really stupid.... (does someone told this earlier, and i didn't red it?...argh)
Of course, he ladder system should be abandonned in this case.
In extension, it could be wonderfull, if, a noob, could win extra points, when he kill a better player... or if ladder stayes, just take ladder from this player.
u see?
(sorry, i go out.... i feel dumb...hummm (better, for graphics...))
The lag charity setting could be set extremely high. I'm not sure how that affects the ladder, as the game server sets lag charity each round. So in play, I don't believe the lag charity really means anything above 100 or so. But when it calculates the ladder, does it use your lag charity setting? Which could be 30,000? I think I should test this.
what i would think for simple and accurate would be something like:
(kills(x) + matches won(y) + rounds won(z) - deaths(a))/(total connections to server)(b)
where x y z a and b are modifiers that determine the weight each of the categories holds
the only pitfall is it doesn't determine if low ranked players are beating higher ranked ones or not
(kills(x) + matches won(y) + rounds won(z) - deaths(a))/(total connections to server)(b)
where x y z a and b are modifiers that determine the weight each of the categories holds
the only pitfall is it doesn't determine if low ranked players are beating higher ranked ones or not
stakes is high