Upvote Upvoted 1 Downvote Downvoted
Pub class-ratio balance system
1
#1
0 Frags +

There is another thread vaguely discussing balance issues in pubs currently and I want to propose a very rough-draft idea for a pub class balance system. 99% of this was worked up by my teammate Slider. This doesn't address rebalancing teams by individual player skill but rather by the mix of classes on the field. Unlike something like a weapon balance discussion where you must assume that a highly skilled player is trying to get 100% effectiveness out of the item this instead assumes everyone in the pub is of the same, average pubber skill. Also the standard format assumed is a 12v12 though the system should work for any amount of players (but would be totally disabled when there's less than 6). All of the individual values and numbers can be tweaked and are subject to deliberation.

Everyone knows the scenario... You have 3 spies and 3 snipers on your team while the other team has a couple of soldiers, a couple of demos, a heavy a pyro and a med as well as some supports/specialists and the game is generally all but decided. Your team has no pushing power, no ability to maintain a front line and it's generally a one sided affair.

The basic idea is that every class has a point value assigned to it, either positive or negative. The classes in the table below are arranged by HP, but also you could say there's a pattern of generalists being at the top vs specialists at the bottom, or maybe even lower skill floor at the top (kind of, a bit). The sum of all of the values of the classes on team has to be at or above zero, and the game would not allow you to choose classes that would drive the value below zero at any point. Basically the team needs to invest into strong all-arounder fighting classes to get points with which to afford specialist classes. The system roughly enforces classes which are "good for the team" and helps maintain a certain balance and pace to the game.

http://i.imgur.com/p2mKGIF.jpg

The fighting classes maintain the same positive values regardless of their quantity but the support-ier classes have an increased negative cost with each aditional one. Not only does this system encourage a certain ratio between the "diferent types" of classes but also it discourages stacking too hard into one specialist class. A sniper and a spy is a better deal than 2 spies. 3 spies is terrible as they bump into people more, make the enemy team more paranoid (to their detriment), target the same players and generally step on each other's toes, not to mention the lack of contribution to the fighting at the front. Engineers are also kind of special because they are very important, so their initial penalty is small, but each next engie raises it by a higher amount. Too many engies on offense hurts in an obvious way but for defense it's often very good for your team-- too good if we're actually playing the objective and the game. It kind of makes the game crappy for everyone in the server.

So purely as example, if you have 10 soldiers and 2 spies on your team the total sum is 11 (good). 9 soldiers and 3 spies comes to 0 (still good). You cannot, however, have 8 soldiers and 4 spies (nor 3 spies, as 8 soldiers does not afford 3 spies).

Obviously this system isn't perfect. You can still argue that Valve wants zero restrictions or that it's too draconian and there are still ways to "game" the system and end up with some weird or less than ideal team setups (but some flexibility is a good thing). That's all reasonable but it still seems like a pretty decent system, or at least a place to start from. Once more all credit goes to Slider.

Thoughts?

There is another thread vaguely discussing balance issues in pubs currently and I want to propose a very rough-draft idea for a pub class balance system. 99% of this was worked up by my teammate Slider. This doesn't address rebalancing teams by individual player skill but rather by the mix of classes on the field. Unlike something like a weapon balance discussion where you must assume that a highly skilled player is trying to get 100% effectiveness out of the item this instead assumes everyone in the pub is of the same, average pubber skill. Also the standard format assumed is a 12v12 though the system should work for any amount of players (but would be totally disabled when there's less than 6). All of the individual values and numbers can be tweaked and are subject to deliberation.

Everyone knows the scenario... You have 3 spies and 3 snipers on your team while the other team has a couple of soldiers, a couple of demos, a heavy a pyro and a med as well as some supports/specialists and the game is generally all but decided. Your team has no pushing power, no ability to maintain a front line and it's generally a one sided affair.

The basic idea is that every class has a point value assigned to it, either positive or negative. The classes in the table below are arranged by HP, but also you could say there's a pattern of generalists being at the top vs specialists at the bottom, or maybe even lower skill floor at the top (kind of, a bit). The sum of all of the values of the classes on team has to be at or above zero, and the game would not allow you to choose classes that would drive the value below zero at any point. Basically the team needs to invest into strong all-arounder fighting classes to get points with which to afford specialist classes. The system roughly enforces classes which are "good for the team" and helps maintain a certain balance and pace to the game.

[img]http://i.imgur.com/p2mKGIF.jpg[/img]

The fighting classes maintain the same positive values regardless of their quantity but the support-ier classes have an increased negative cost with each aditional one. Not only does this system encourage a certain ratio between the "diferent types" of classes but also it discourages stacking too hard into one specialist class. A sniper and a spy is a better deal than 2 spies. 3 spies is terrible as they bump into people more, make the enemy team more paranoid (to their detriment), target the same players and generally step on each other's toes, not to mention the lack of contribution to the fighting at the front. Engineers are also kind of special because they are very important, so their initial penalty is small, but each next engie raises it by a higher amount. Too many engies on offense hurts in an obvious way but for defense it's often very good for your team-- too good if we're actually playing the objective and the game. It kind of makes the game crappy for everyone in the server.

So purely as example, if you have 10 soldiers and 2 spies on your team the total sum is 11 (good). 9 soldiers and 3 spies comes to 0 (still good). You cannot, however, have 8 soldiers and 4 spies (nor 3 spies, as 8 soldiers does not afford 3 spies).

Obviously this system isn't perfect. You can still argue that Valve wants zero restrictions or that it's too draconian and there are still ways to "game" the system and end up with some weird or less than ideal team setups (but some flexibility is a good thing). That's all reasonable but it still seems like a pretty decent system, or at least a place to start from. Once more all credit goes to Slider.

Thoughts?
2
#2
9 Frags +

Seems like it still doesn't heavily weight the fact that the team without a medic is at a significant disadvantage against the team with a medic.

Seems like it still doesn't heavily weight the fact that the team without a medic is at a significant disadvantage against the team with a medic.
3
#3
Momentum Mod
9 Frags +

that's a lot of words

that's a lot of words
4
#4
RGL.gg
2 Frags +

it resembles a brick wall

it resembles a brick wall
5
#5
3 Frags +

Your system would never be implemented with it's current class values on any server purely for the fact that a highlander team would be breaking the points limits.

Highlander = -3
6v6 with roamer offclass (a la harbleu -> spy; mackey -> sniper) = -2

Your system would never be implemented with it's current class values on any server purely for the fact that a highlander team would be breaking the points limits.

Highlander = -3
6v6 with roamer offclass (a la harbleu -> spy; mackey -> sniper) = -2
6
#6
3 Frags +

Interesting algorithm but at the end of the day, a hidden function that determines who can play what is detrimental to the make-up of casual TF2. While on a more serious or try-hardy pub where you're stuck on the team with nothing but support classes and you guys can't push, the best mechanic is an updated team-scramble.

Biggest example of an improvement would be to take the person that made the most points and the medic that made the second most points be on opposite teams, while lending some kind of mix of this equation towards balancing the classes as they were at the end of the round before the scramble.

This would spread out the snipers and spies accordingly, and add positive classes to the losing team.

There are however some makeups of maps where these support classes are much more powerful, or have a stronger time on defense vs offense, so if valve retweaked the autobalance mechanics, they would probably favor switching defensive classes (demoman, heavy, engineer) towards the defending team on Payload and A/D maps, rather than mix them evenly.

Interesting algorithm but at the end of the day, a hidden function that determines who can play what is detrimental to the make-up of casual TF2. While on a more serious or try-hardy pub where you're stuck on the team with nothing but support classes and you guys can't push, the best mechanic is an updated team-scramble.

Biggest example of an improvement would be to take the person that made the most points and the medic that made the second most points be on opposite teams, while lending some kind of mix of this equation towards balancing the classes as they were at the end of the round before the scramble.

This would spread out the snipers and spies accordingly, and add positive classes to the losing team.


There are however some makeups of maps where these support classes are much more powerful, or have a stronger time on defense vs offense, so if valve retweaked the autobalance mechanics, they would probably favor switching defensive classes (demoman, heavy, engineer) towards the defending team on Payload and A/D maps, rather than mix them evenly.
7
#7
9 Frags +

Why is multiple Scouts a bad thing? And for that matter, why is multiple Heavies a GOOD thing?

Demoman should be considered the same value as Soldier. Except at that point...Hell, even now, it doesn't take into consideration the fact that a player might run Demoknight, and thus mitigate the pushing potential Demoman normally has. So you have four Demoknights running around being about as useless as four Spies. I'm not even gonna begin with Pyro, and just say that a Scout is more valuable for pushing than a Pyro is.

It's all a moot point. All this system does is favor some classes over others based on an arbitrary sorting method (Heavy is apparently a generalist now?) and comes with the same problems regular class limits have (potentially keeping lower-skilled players on important pick classes and keeing higher-skilled players out.) The method isn't SUPPOSED to be based on skill level, but I don't see any advantage this has over just putting regular class limits. At 4 Spies, there's no combination on a 12 man team that will bring the point value to zero, and running 3 spies or snipers basically forces you to run only Soldiers and Heavies. And no Medics. Just putting a class limit on 2 Spies or 2 Snipers will accomplish the bare minimum this setup attempts to solve, though it doesn't address Engineers, but that's a whole different can of worms.

It's unnecessary work to implement and doesn't really solve anything standard class limits don't already address. Nor does it really rectify the short-comings of class limits.

Why is multiple Scouts a bad thing? And for that matter, why is multiple Heavies a GOOD thing?

Demoman should be considered the same value as Soldier. Except at that point...Hell, even now, it doesn't take into consideration the fact that a player might run Demoknight, and thus mitigate the pushing potential Demoman normally has. So you have four Demoknights running around being about as useless as four Spies. I'm not even gonna begin with Pyro, and just say that a Scout is more valuable for pushing than a Pyro is.

It's all a moot point. All this system does is favor some classes over others based on an arbitrary sorting method (Heavy is apparently a generalist now?) and comes with the same problems regular class limits have (potentially keeping lower-skilled players on important pick classes and keeing higher-skilled players out.) The method isn't SUPPOSED to be based on skill level, but I don't see any advantage this has over just putting regular class limits. At 4 Spies, there's no combination on a 12 man team that will bring the point value to zero, and running 3 spies or snipers basically forces you to run only Soldiers and Heavies. And no Medics. Just putting a class limit on 2 Spies or 2 Snipers will accomplish the bare minimum this setup attempts to solve, though it doesn't address Engineers, but that's a whole different can of worms.

It's unnecessary work to implement and doesn't really solve anything standard class limits don't already address. Nor does it really rectify the short-comings of class limits.
8
#8
0 Frags +

I will also say that this policy doesn't take into account the fact that the first of a class is great, while later ones are detrimental.

For example, in a 9v9 setting, the first sniper is a good thing, the second sniper is usually neutral to bad, and beyond that it's awful. However, the first demo is great, but further demos get progressively more powerful because it becomes harder to push through the crosscutting spam.

The system also doesn't take into account the changing value with team sizes. 3 medics on a 4 person team is bad. 3 medics on a 9-12 person team is great.

EDIT: You also have to account for the fact that "fun" and "best for winning" are not necessarily the same. Most people don't enjoy trying to push into multiple engies with wrangler+RR with pyros helping them.

I will also say that this policy doesn't take into account the fact that the first of a class is great, while later ones are detrimental.

For example, in a 9v9 setting, the first sniper is a good thing, the second sniper is usually neutral to bad, and beyond that it's awful. However, the first demo is great, but further demos get progressively more powerful because it becomes harder to push through the crosscutting spam.

The system also doesn't take into account the changing value with team sizes. 3 medics on a 4 person team is bad. 3 medics on a 9-12 person team is great.

EDIT: You also have to account for the fact that "fun" and "best for winning" are not necessarily the same. Most people don't enjoy trying to push into multiple engies with wrangler+RR with pyros helping them.
9
#9
0 Frags +

Hey all. Thanks for looking into my little pet balancing idea. It's mainly supposed to be a middle-ground between the complete chaos of a pub, and the rigidity of hard class limits by punishing detrimental redundancy; basically that, you can play any combination of 3 support classes as long as you have the combat classes to balance it out, rather than a hard cap of 2 spies, for example.

Anyway, there's a lot of good points I'd like to respond to.

MasterKuniSeems like it still doesn't heavily weight the fact that the team without a medic is at a significant disadvantage against the team with a medic.

Yeah, the numbers should probably be adjusted to encourage medic play, at least for the first one. I just didn't want the addition of a medic to contribute to the gaining of a support class because a medic with "nothing good to heal" doesn't help much.

ThomasYour system would never be implemented with it's current class values on any server purely for the fact that a highlander team would be breaking the points limits.

True, I think a good goal would be to rework the numbers such that it allows a HL setup at minimum, while still not allowing for support classes to be the first to join.

GeosusWhy is multiple Scouts a bad thing? And for that matter, why is multiple Heavies a GOOD thing?

You have to take into account that this is a balance system for pubs, where the average numbers of hours of TF2 played by most people is 74. As such it's better to encourage newer players to play classes with a lower skill floor like heavy where they can still contribute positively.

Scout is harder to justify though I admit, as in competitive we're used to it being one of the strongest if not the strongest class, but that's only in the hands of really skilled players. Your average pub scout isn't, on the other hand, going to be as effective as a pub soldier or pub heavy, or even a W+M1 pyro maybe. Side note, if you do get a lot of good scouts on one team, this can make the other team feel pretty helpless as well, a pretty common complaint I hear on pubs.

Geosusit doesn't take into consideration the fact that a player might run Demoknight, and thus mitigate the pushing potential Demoman normally has. So you have four Demoknights running around being about as useless as four Spies.

Right so, this system isn't a fix-all. Ultimately nothing, including hard class limits, can stop players from being a useless demoknight, or doing nothing but taunt in spawn, or otherwise be useless.

Geosus I don't see any advantage this has over just putting regular class limits...Just putting a class limit on 2 Spies or 2 Snipers will accomplish the bare minimum this setup attempts to solve

The advantage is it allows for different combinations of support classes provided your have the combat classes to balance it, while also keeping their total number in check, e.g., you can run 3 spies with this system, but you can't if you just do a 2 spy class limit. Furthermore, a simple class limit of 2 spies and 2 snipers still allows for the server to have 2 spies and 2 snipers constituting a makeup of an entire 1/3 of a 12 player team, which is more than enough to sink them. Worse, with class limits only, servers can start filling with snipers and spies, so you could get a situation where, you join first on the server as soldier, the first enemy joins as soldier, but your next 4 teammates are snipers and spies, while the enemy's next 4 teammates are a heavy, demo, pyro, and medic. There is really no fun to be had on that team IMHO.

Geosusit doesn't address Engineers, but that's a whole different can of worms.

I think I would adjust it so there's no penalty to the first engineer, as it's pretty important to have 1, but the reason I'd keep him going sharply into the negatives is the same reason they're limited to 1 in 6v6. It becomes next to impossible to push last if you have 3 engies with level 3s. Have you ever tried to push upward or snakewater last on a pub with 3 engies sitting on it? It takes an enormously disproportionately large amount of coordination and skill to break that compared to what it takes to build and whack a sentry, especially if they're using the magic wand that is the short circuit, with wranglers, with pyros helping. Multiple engies get a penalty to alleviate that.

Frost_BiteThe system also doesn't take into account the changing value with team sizes. 3 medics on a 4 person team is bad. 3 medics on a 9-12 person team is great.

Good point, you're right. The good news is that the numbers can be adjusted to fix that

Frost_BiteYou also have to account for the fact that "fun" and "best for winning" are not necessarily the same.

So yeah, there's always an argument that can be made that anything that takes away from any amount of freedom inherently takes away from the fun of the individual. The reason I thought this up was because I didn't think it was fair that a minority of the team can completely sink a majority of the team via their class choice alone. The 4th sniper's fun may be preserved because he got to play what he wanted when he wanted, but in doing so he made it decidedly unfun for the entire rest of the team to play, as they're outnumbered and outclassed on the front lines while the sniper sits cozily behind. Net fun for the team is in the negatives, and I would personally strongly prefer a system that was conducive to team play in team fortress two, even if it meant I couldn't play my favorite class all of the time. But that's just IMHO.
________________________________

Ultimately I think there is an ideal set of numbers that could be put together to achieve the best balance, and these responses made me realize I was farther than I thought from that. Thanks for indulging my nerd essay.

Hey all. Thanks for looking into my little pet balancing idea. It's mainly supposed to be a middle-ground between the complete chaos of a pub, and the rigidity of hard class limits by punishing detrimental redundancy; basically that, you can play any combination of 3 support classes as long as you have the combat classes to balance it out, rather than a hard cap of 2 spies, for example.

Anyway, there's a lot of good points I'd like to respond to.

[quote=MasterKuni]Seems like it still doesn't heavily weight the fact that the team without a medic is at a significant disadvantage against the team with a medic.[/quote]

Yeah, the numbers should probably be adjusted to encourage medic play, at least for the first one. I just didn't want the addition of a medic to contribute to the gaining of a support class because a medic with "nothing good to heal" doesn't help much.


[quote=Thomas]Your system would never be implemented with it's current class values on any server purely for the fact that a highlander team would be breaking the points limits.[/quote]

True, I think a good goal would be to rework the numbers such that it allows a HL setup at minimum, while still not allowing for support classes to be the first to join.

[quote=Geosus]Why is multiple Scouts a bad thing? And for that matter, why is multiple Heavies a GOOD thing?[/quote]

You have to take into account that this is a balance system for pubs, where the average numbers of hours of TF2 played by most people is 74. As such it's better to encourage newer players to play classes with a lower skill floor like heavy where they can still contribute positively.

Scout is harder to justify though I admit, as in competitive we're used to it being one of the strongest if not the strongest class, but that's only in the hands of really skilled players. Your average pub scout isn't, on the other hand, going to be as effective as a pub soldier or pub heavy, or even a W+M1 pyro maybe. Side note, if you [i]do[/i] get a lot of good scouts on one team, this can make the other team feel pretty helpless as well, a pretty common complaint I hear on pubs.


[quote=Geosus]it doesn't take into consideration the fact that a player might run Demoknight, and thus mitigate the pushing potential Demoman normally has. So you have four Demoknights running around being about as useless as four Spies.[/quote]

Right so, this system isn't a fix-all. Ultimately nothing, including hard class limits, can stop players from being a useless demoknight, or doing nothing but taunt in spawn, or otherwise be useless.

[quote=Geosus] I don't see any advantage this has over just putting regular class limits...Just putting a class limit on 2 Spies or 2 Snipers will accomplish the bare minimum this setup attempts to solve[/quote]

The advantage is it allows for different combinations of support classes provided your have the combat classes to balance it, while also keeping their total number in check, e.g., you can run 3 spies with this system, but you can't if you just do a 2 spy class limit. Furthermore, a simple class limit of 2 spies and 2 snipers still allows for the server to have 2 spies and 2 snipers constituting a makeup of an entire 1/3 of a 12 player team, which is more than enough to sink them. Worse, with class limits only, servers can [i]start[/i] filling with snipers and spies, so you could get a situation where, you join first on the server as soldier, the first enemy joins as soldier, but your next 4 teammates are snipers and spies, while the enemy's next 4 teammates are a heavy, demo, pyro, and medic. There is really no fun to be had on that team IMHO.

[quote=Geosus]it doesn't address Engineers, but that's a whole different can of worms.[/quote]

I think I would adjust it so there's no penalty to the first engineer, as it's pretty important to have 1, but the reason I'd keep him going sharply into the negatives is the same reason they're limited to 1 in 6v6. It becomes next to impossible to push last if you have 3 engies with level 3s. Have you ever tried to push upward or snakewater last on a pub with 3 engies sitting on it? It takes an enormously disproportionately large amount of coordination and skill to break that compared to what it takes to build and whack a sentry, [i]especially[/i] if they're using the magic wand that is the short circuit, with wranglers, with pyros helping. Multiple engies get a penalty to alleviate that.

[quote=Frost_Bite]The system also doesn't take into account the changing value with team sizes. 3 medics on a 4 person team is bad. 3 medics on a 9-12 person team is great.[/quote]

Good point, you're right. The good news is that the numbers can be adjusted to fix that

[quote=Frost_Bite]You also have to account for the fact that "fun" and "best for winning" are not necessarily the same.[/quote]

So yeah, there's always an argument that can be made that anything that takes away from any amount of freedom inherently takes away from the fun of the individual. The reason I thought this up was because I didn't think it was fair that a minority of the team can completely sink a majority of the team via their class choice alone. The 4th sniper's fun may be preserved because he got to play what he wanted when he wanted, but in doing so he made it decidedly unfun [i]for the entire rest of the team[/i] to play, as they're outnumbered and outclassed on the front lines while the sniper sits cozily behind. Net fun for the team is in the negatives, and I would personally strongly prefer a system that was conducive to team play in team fortress two, even if it meant I couldn't play my favorite class all of the time. But that's just IMHO.
________________________________

Ultimately I think there [i]is[/i] an ideal set of numbers that could be put together to achieve the best balance, and these responses made me realize I was farther than I thought from that. Thanks for indulging my nerd essay.
10
#10
0 Frags +

Are you saying having scouts in pubs are bad?

I'm not sure if this is what you're saying but how is having a scout bad and a pyro good.

Are you saying having scouts in pubs are bad?

I'm not sure if this is what you're saying but how is having a scout bad and a pyro good.
11
#11
0 Frags +

why would u want to class balance a pub, who cares?

why would u want to class balance a pub, who cares?
12
#12
1 Frags +

For many of us, pubs are the go-to option for training. It's really no fun to spend 15 minutes fighting a team of Engineers, especially when 4 players on your team seriously think that Spy is the ultimate counter to the Engineer. It's also not fun to keep hopping servers to find that one mythical server where Medic is a class that exists, players are conscious about their class selection or at least skilled enough to contribute in spite of the team make-up, and there aren't 9 Heavies on the enemy team. Alternatively, there's tf2center! But if I wanted to not play Team Fortress 2 for 30-60 minutes, I'd... uh. Just not play Team Fortress 2.

His head's in the right place, but I don't agree with his ideas in their current form. And there probably really is no point in trying to balance pubs, yeah.

For many of us, pubs are the go-to option for training. It's really no fun to spend 15 minutes fighting a team of Engineers, especially when 4 players on your team seriously think that Spy is the ultimate counter to the Engineer. It's also not fun to keep hopping servers to find that one mythical server where Medic is a class that exists, players are conscious about their class selection or at least skilled enough to contribute in spite of the team make-up, and there aren't 9 Heavies on the enemy team. Alternatively, there's tf2center! But if I wanted to not play Team Fortress 2 for 30-60 minutes, I'd... uh. Just not play Team Fortress 2.

His head's in the right place, but I don't agree with his ideas in their current form. And there probably really is no point in trying to balance pubs, yeah.
13
#13
5 Frags +

im 100% positive you have a better use of ur time than this shit

im 100% positive you have a better use of ur time than this shit
14
#14
0 Frags +

lmao

lmao
15
#15
6 Frags +

http://i.imgur.com/DnRfiNt.gif

[img]http://i.imgur.com/DnRfiNt.gif[/img]
Please sign in through STEAM to post a comment.