New ladder system for fortress
- Lacrymosa
- Round Winner
- Posts: 286
- Joined: Wed Feb 15, 2006 6:44 pm
- Location: Heaven or Hell...?
- Contact:
New ladder system for fortress
I would like to ask if there is the possibility to change the ladder on fortress servers.
If I got it right, currently you get points if your team win. So even if you're dead or just drive in stupid and senseless ways (noob?), you will get points.
But I prefer, that the ladder just count for every single player.
So for example, you could get points if you:
- Conquer the enemies fortress
- Kill an enemy
This is just a proposal, if anyone knows how to change the ladder, say it.
If I got it right, currently you get points if your team win. So even if you're dead or just drive in stupid and senseless ways (noob?), you will get points.
But I prefer, that the ladder just count for every single player.
So for example, you could get points if you:
- Conquer the enemies fortress
- Kill an enemy
This is just a proposal, if anyone knows how to change the ladder, say it.
- Tank Program
- Forum & Project Admin, PhD
- Posts: 6712
- Joined: Thu Dec 18, 2003 7:03 pm
- Phytotron
- Formerly Oscilloscope
- Posts: 5042
- Joined: Thu Jun 09, 2005 10:06 pm
- Location: A site or situation, especially considered in regard to its surroundings.
- Contact:
Yeah, the current ladder is silly, as most(I think) people know, whatever the server or its type. (One exception may be servers with the ridiculous scoring system that gives an undue boatload of points to round survivors versus core dumpers. But even in that case its imperfect....)
I think that a ladder with any kind of statistical significance should be based on some sort of calculated efficiency rating, not raw numbers (and certainly not round survival). But whatever. My mathematical ability is too pitiful to devote much thought to it.
(Incidentally, I've actually looked at the ladder settings somewhat recently. Out of curiosity, whose idea was it to have the ladder behave like a wagering system? Interesting choice.)
I think that a ladder with any kind of statistical significance should be based on some sort of calculated efficiency rating, not raw numbers (and certainly not round survival). But whatever. My mathematical ability is too pitiful to devote much thought to it.
(Incidentally, I've actually looked at the ladder settings somewhat recently. Out of curiosity, whose idea was it to have the ladder behave like a wagering system? Interesting choice.)
- philippeqc
- Long Poster - Project Developer - Sage
- Posts: 1526
- Joined: Mon Jul 12, 2004 8:55 am
- Location: Stockholm
- Contact:
Re: New ladder system for fortress
Very personal opinion:
- Pause and separation of ideas -
Isn't not being able to play for a part of the round a sufficient penalty?
Yes, we would want new players to actually enjoy the game and be motivated to learn it's fine 'leet etic.
::End of venting::
Sorry if you are not of that new crowd mentality of "kick new players on sight" and where trying to make a different point.
I'm not saying your system is without value, just that for me you are proposing to change from one type of bad to another type.
Phytotron, I second your revival of the wagering based system.
-ph
Wouldn't such a scoring favor survivalist technique over intrepid dashes.LukeSky wrote:So even if you're dead
- Pause and separation of ideas -
Isn't not being able to play for a part of the round a sufficient penalty?
::Venting::LukeSky wrote:or just drive in stupid and senseless ways (noob?)
Yes, we would want new players to actually enjoy the game and be motivated to learn it's fine 'leet etic.
::End of venting::
Sorry if you are not of that new crowd mentality of "kick new players on sight" and where trying to make a different point.
Then you only favor agressive techniques, rather than technique than migh benefit the team as a whole.So for example, you could get points if you:
- Conquer the enemies fortress
- Kill an enemy
I'm not saying your system is without value, just that for me you are proposing to change from one type of bad to another type.
Phytotron, I second your revival of the wagering based system.
-ph
Canis meus id comedit.
Allow me to be the one that tries to take this thread off topic. 
So, on my server, there are 3 fortresses and correspondingly 3 teams. When it's 2 teams, giving a round win score is very sensible. Taking the other team's fortress or successfully defending your own fortress is worth the round win and it works very well.
When it's 3 teams, the system breaks down. The team that "wins" is the one who had a fortress standing and took the last standing fortress (or maybe it's just whoever takes the last standing fortress). I want to reward a team on what they actually accomplish, and winning the round is a much more nebulous concept, not very easily defined anymore.
So I give 5 points for a fortress, 3 for a core dump, 2-3 (I forget) for successfully defending your fortress, and 2 for the round win, figuring the round win credit is worth something. I also have the fortress wipe out half the team when it's conquered because I think that's neat. And what I see is this:
When it's 1 v 1 v 1, everyone defends. If I leave my zone to attack someone, the third guy will leave his to take my zone. If my attack is successful, then he will kill me by taking my fortress before I can actually take the fruits of my conquest. And then he'll attack a weakened zone. So he's a carrion bird, essentially.
When it's 2 v 1 and the other team is all dead, the 2 will first take the empty fortress for free points, then move to attack the other team. The other team may well have killed the third team and gotten those points and managed not to take their zone, then my team comes along and gets the free points for the fortress. When the round is over, whether we win our fortress attack or not, we still make out like bandits with a whole slew of free points.
Now, when the round starts, the tactical options are really great and we're really enjoying it (those that I've played with, anyway). It's neat that you can actually choose to concentrate force or divide it. It's a risky strategy to attack with your only two teammates and leave your zone undefended, but it can really pay off big, and because of the map shape, it's not entirely necessary to defend from inside the zone. In fact, I've seen folks use the terrain really well defensively.
So the point spreads wind up being quite dramatic. Winning by 100 points isn't uncommon, although it's usually more like 50 points. That's a lot of point difference in a 7 minute match! Clearly my point awards are unbalanced. Any suggestions?
I want to promote good team play and reward teams for their actual performance. I want to eliminate or at least mitigate the damage done to the "integrity" of the match by free points. You know, it's nice to take an empty zone and get the free points, but it's disorienting to see it as a strategy.
It's entirely possible that code changes are needed, and since I'm running off the trunk (well, 0.3.0 branch right now), it's entirely reasonable to make those changes.

So, on my server, there are 3 fortresses and correspondingly 3 teams. When it's 2 teams, giving a round win score is very sensible. Taking the other team's fortress or successfully defending your own fortress is worth the round win and it works very well.
When it's 3 teams, the system breaks down. The team that "wins" is the one who had a fortress standing and took the last standing fortress (or maybe it's just whoever takes the last standing fortress). I want to reward a team on what they actually accomplish, and winning the round is a much more nebulous concept, not very easily defined anymore.
So I give 5 points for a fortress, 3 for a core dump, 2-3 (I forget) for successfully defending your fortress, and 2 for the round win, figuring the round win credit is worth something. I also have the fortress wipe out half the team when it's conquered because I think that's neat. And what I see is this:
When it's 1 v 1 v 1, everyone defends. If I leave my zone to attack someone, the third guy will leave his to take my zone. If my attack is successful, then he will kill me by taking my fortress before I can actually take the fruits of my conquest. And then he'll attack a weakened zone. So he's a carrion bird, essentially.
When it's 2 v 1 and the other team is all dead, the 2 will first take the empty fortress for free points, then move to attack the other team. The other team may well have killed the third team and gotten those points and managed not to take their zone, then my team comes along and gets the free points for the fortress. When the round is over, whether we win our fortress attack or not, we still make out like bandits with a whole slew of free points.
Now, when the round starts, the tactical options are really great and we're really enjoying it (those that I've played with, anyway). It's neat that you can actually choose to concentrate force or divide it. It's a risky strategy to attack with your only two teammates and leave your zone undefended, but it can really pay off big, and because of the map shape, it's not entirely necessary to defend from inside the zone. In fact, I've seen folks use the terrain really well defensively.
So the point spreads wind up being quite dramatic. Winning by 100 points isn't uncommon, although it's usually more like 50 points. That's a lot of point difference in a 7 minute match! Clearly my point awards are unbalanced. Any suggestions?
I want to promote good team play and reward teams for their actual performance. I want to eliminate or at least mitigate the damage done to the "integrity" of the match by free points. You know, it's nice to take an empty zone and get the free points, but it's disorienting to see it as a strategy.
It's entirely possible that code changes are needed, and since I'm running off the trunk (well, 0.3.0 branch right now), it's entirely reasonable to make those changes.

Check out my YouTube channel: https://youtube.com/@davefancella?si=H--oCK3k_dQ1laDN
Be the devil's own, Lucifer's my name.
- Iron Maiden
Be the devil's own, Lucifer's my name.
- Iron Maiden
Lucifer: two additional rules may help here:
1. Attacking a zone has no effect unless
a) there was a defender around not too long ago or
b) an attacker of the zone owning team was near your zone not too long ago
2. If a team is dead, its zone vanishes after some time
A slightly negative FORTRESS_CONQUEST_DECAY_RATE and finite FORTRESS_CONQUEST_TIMEOUT do some of that. It'll simply make a zone vanish when it is not attacked for a while. It's not perfect, it may have the effect of everyone just waiting for their zone to time out, then do a regular last man standing match. But I guess it is worth a try. The suggested rules sound a bit too artificial for me.
On topic: There simply is no way to judge the performance of a single player in a team. Every metric you introduce will only have the effect of encouraging players to get a high metric, not to play for the team, as Philippe rightly points out. The only way to get a player metric would be this:
Do a ladder for teams based on round wins. Every team, referenced by the list of players that are in it, gets a slot on that ladder. Then, after enough data has been collected, look at team score differences. If team {some players} + LukeSky has a higher score on the ladder than team { the same players } + Z-Man, then that is a hint that LukeSky is a better player than Z-Man.
The obvious problem here: a lot of data needs to be collected before this gives a meaningful answer, and the ladder needs a lot of storage space, because there is a huge amount of possible teams that can be formed. It may work on a server with a maximum team size of three. But it can be sabotaged
, provided I don't care about my own rating. If I want to bring down LukeSky's ladder rating, I just play really bad whenever he is on my team. I play normal when he's not. The mechanism can't know higher performance of a team with me and without LukeSky, compared to a team with us both, is caused by me and not LukeSky.
1. Attacking a zone has no effect unless
a) there was a defender around not too long ago or
b) an attacker of the zone owning team was near your zone not too long ago
2. If a team is dead, its zone vanishes after some time
A slightly negative FORTRESS_CONQUEST_DECAY_RATE and finite FORTRESS_CONQUEST_TIMEOUT do some of that. It'll simply make a zone vanish when it is not attacked for a while. It's not perfect, it may have the effect of everyone just waiting for their zone to time out, then do a regular last man standing match. But I guess it is worth a try. The suggested rules sound a bit too artificial for me.
On topic: There simply is no way to judge the performance of a single player in a team. Every metric you introduce will only have the effect of encouraging players to get a high metric, not to play for the team, as Philippe rightly points out. The only way to get a player metric would be this:
Do a ladder for teams based on round wins. Every team, referenced by the list of players that are in it, gets a slot on that ladder. Then, after enough data has been collected, look at team score differences. If team {some players} + LukeSky has a higher score on the ladder than team { the same players } + Z-Man, then that is a hint that LukeSky is a better player than Z-Man.
The obvious problem here: a lot of data needs to be collected before this gives a meaningful answer, and the ladder needs a lot of storage space, because there is a huge amount of possible teams that can be formed. It may work on a server with a maximum team size of three. But it can be sabotaged

- Lacrymosa
- Round Winner
- Posts: 286
- Joined: Wed Feb 15, 2006 6:44 pm
- Location: Heaven or Hell...?
- Contact:
I made up my mind and I would like to talk about this type of ladder:
On Wild Fortress, 12 rounds are one match:
You get points if:
- Your team wins the match (for example: 60)
- Your team wins the round (for example: 5)
Also you could get some bonus points:
- Kill an enemy (1 point)
- Conquer the enemies fortress (2.5 points)
- Be the last one alive of your team (2 points)
Maybe you could lose points if:
- You die (suicide?)
On Wild Fortress, 12 rounds are one match:
You get points if:
- Your team wins the match (for example: 60)
- Your team wins the round (for example: 5)
Also you could get some bonus points:
- Kill an enemy (1 point)
- Conquer the enemies fortress (2.5 points)
- Be the last one alive of your team (2 points)
Maybe you could lose points if:
- You die (suicide?)
No 
LukeSky: your regular data can be combined from the highscore data that is already collected. Just take the number of won rounds * 5 plus the number of won matches * 60. The bonus points aren't currently tracked, but they're the kind of score bonus that gives incentives to do wrong things, like, not defend at all and all rush to the enemy base for the conquest point. And your suggestion wouldn't be a ladder, just a score counter that measures how often you play. In a ladder, performance differences have to be measured and taken into account. It shouldn't matter (after a certain threshold has been reached) how often you play, it should just matter how well you play.
newbie: In a close round, the defender of the winning team, after having valiantly defended against three opponents, often dies from exhaustion when he gets the message that the enemy fort has been taken (or shortly before that, but his team wins anyway because he kept the enemies out just long enough). Should that be punished?

LukeSky: your regular data can be combined from the highscore data that is already collected. Just take the number of won rounds * 5 plus the number of won matches * 60. The bonus points aren't currently tracked, but they're the kind of score bonus that gives incentives to do wrong things, like, not defend at all and all rush to the enemy base for the conquest point. And your suggestion wouldn't be a ladder, just a score counter that measures how often you play. In a ladder, performance differences have to be measured and taken into account. It shouldn't matter (after a certain threshold has been reached) how often you play, it should just matter how well you play.
newbie: In a close round, the defender of the winning team, after having valiantly defended against three opponents, often dies from exhaustion when he gets the message that the enemy fort has been taken (or shortly before that, but his team wins anyway because he kept the enemies out just long enough). Should that be punished?
- Phytotron
- Formerly Oscilloscope
- Posts: 5042
- Joined: Thu Jun 09, 2005 10:06 pm
- Location: A site or situation, especially considered in regard to its surroundings.
- Contact:
Re: New ladder system for fortress
Er, my very first sentence indicated that I do not favor the current, wager-based ladder system.philippeqc wrote:Phytotron, I second your revival of the wagering based system.
z-man, it should not be punished, just points for rounds are not credible because there is no difference if you will fight till the end or die in the first 10s
at least points should not be equal for all players, for example (14vs14)
(only winning team gets these points, one round)
the last 3 survived players - 12pts
4-7 - 8 pts
8-10 - 4 pts
11-14 (the first 3 dead players) - 0 pts
+ 2pts for each kill (both teams)
the ladder can be counted in experience points
exp points = points/120 (for stats on the web)
at least points should not be equal for all players, for example (14vs14)
(only winning team gets these points, one round)
the last 3 survived players - 12pts
4-7 - 8 pts
8-10 - 4 pts
11-14 (the first 3 dead players) - 0 pts
+ 2pts for each kill (both teams)
the ladder can be counted in experience points
exp points = points/120 (for stats on the web)
- philippeqc
- Long Poster - Project Developer - Sage
- Posts: 1526
- Joined: Mon Jul 12, 2004 8:55 am
- Location: Stockholm
- Contact:
suddenly, tk all my team mates sound like a very interesting thing to do. ;)newbie wrote:z-man, it should not be punished, just points for rounds are not credible because there is no difference if you will fight till the end or die in the first 10s
In other words, each time you only reward certain behaviors, you make it even more interesting to disrupt said behavior. That and totally neglect the fact that there are lots of good and valid strategies out there that are not rewarded by your system.
Sorry for the misunderstanding. I have a vague recollection of z-man proposing a new model for the ladder, where a good player winning over a bad player would get an insignificant gain on the ladder, but a bad player winning over a good one would get a great gain on the ladder.Phytotron wrote:Er, my very first sentence indicated that I do not favor the current, wager-based ladder system.philippeqc wrote:Phytotron, I second your revival of the wagering based system.
Personnally, I think that the ladder should represent your "skill" not how much you play. If you play a lot at the same level of skill, you should hold a constant level on the ladder. If you start to kick ass at a new level (or suck in a creativly new fashion), your position on the ladder should raise to reflect it (or sink ;) ). But that's just my own lil dream world. ;)
-ph
Canis meus id comedit.
why? then you will not get points because your team will losephilippeqc wrote:suddenly, tk all my team mates sound like a very interesting thing to do.
staying alive and dumping opponents is a sad(?) behavior? if you want to know, who is good, you need to reward only certain skilled behaviorsIn other words, each time you only reward certain behaviors, you make it even more interesting to disrupt said behavior.
at present people can get points by doing nothing (the same points as people who had won the round for your team)
