Welcome

Please register for Total War Access to use the forums. If you're an existing user, your forum details will be merged with Total War Access if you register with the same email or username. For more information please read our FAQ’s here.

Categories

Using Free Market principles to balance costs in MP?

WojmirVonCarsteinWojmirVonCarstein Posts: 416Registered Users
So I just started playing MP recently. I am doing OK, probably 50-50 record so far after about 50 games.

I just had a thought about using supply/demand mechanics for costing units in MP (Quick Battles).

IF YOU DON'T FEEL LIKE READING THE LONG WALL OF TEXT THAT FOLLOWS, THE GIST OF THE IDEA IS TO INCREASE / DECREASE COST OF ALL CHOICES IN MP, BASED ON HOW OFTEN IT GETS SELECTED AND HOW WELL IT DOES.

CONTINUE READING FOR DETAILS IF INTERESTED:


Could CA not implement an algorithm that looks at army compositions for each faction and periodically (daily, weekly?) adjusts the cost of each unit, spell, item and ability depending on how often it gets (not) included in an army.

This would mean that over time, just like in the free market, all costs should balance out. Auto-include lords, heroes, units, items, abilities etc. would slowly increase in cost to a point where players take it less and less. Never used things would slowly decrease in cost until people begin figuring it out that these previously useless things are now pretty good.

In terms of implementing this, it would literally take one person from CA's side. As a programmer/mathematician myself, I can already envision how I would do it. I could create a model that takes into account many different factors such as how often the unit is used, how often it losses/draws/wins battle, the rating of the player taking it, how many are included (if as a single entity or in pairs etc.). Then I would have an algorithm that takes into account all these factors in some way and then reduces / increases costs periodically.

So, the above proposition would have some valid counterpoints. Here they are and my answers to them:

1) This would be too cumbersome/expensive for CA to implement:

I can say with confidence that based on my own experience/education, I can do this as a full time job. I am sure CA has more talented people than myself so it might be even easier for them than me

2) Not everyone composes their army based on utility, but also based on what they like. So people who play casually would cause units to be over costed, just because they keep including them in their army. This would make the unit a bad pick for competitive players, thus making the most fun/popular units not cost efficient. The reverse is also true, if the large mass of casual players don't take a cost efficient unit because they don't like it, it's cost will keep going down, thus making them more and more OP:

The above problem would not actually exist because the algorithm I am considering would take into account the win/loss stats of each unit/ability/magic item as well as the rank of the player using it.
So if the top 100 players (who are often winning, that is why they are top 100 duh), keep using Karl Franz on Death Claw, his cost will keep steadily increasing, even if casuals don't really use him because the top 100 players winning units will have a larger effect on the increase/decrease path of said unit. Of course getting the "weighting" right will take a lot of planning before hand as well as adjustments based on what is happening and community feedback, but it can be done. So it's not like the top 100 would effect everything, just that they would have a larger effect than a random hundred of players.

3) If costs keep fluctuating, then saved army compositions will have to be re-adjusted every time a "cost update" is made. This will be both annoying and a waste of time:

This is true to a certain degree. However, if you take a typical saved army with say 15 units, it is not likely that ALL 15 units go up in price or down in price. Some will go up, some will stay the same, some will go down. The most likely result will be that the cost of your army might be under/over by say -25 to +25 gold after each update. So it's not a terribly tough fix to make. Of course the longer this algorithm is implemented, the smaller the fluctuations will be.

4) What if the algorithm is bugged and starts doing wonky things:

Not a problem! the state of the costs would be "manually" reviewed at some time intervals to make sure this does not happen. Also, community feedback will be taken into account in these "manual" review periods, so any "bugs" can be quickly fixed.

5) Wouldn't this make popular factions more and more costly and factions that are not taken cheaper and cheaper?:

True to a certain extent, but the effect would be very, very small (or non-existent) because of the way the algorithm would be implemented.

6) Finally, isn't this a huge, unnecessary complication for little benefit?:

I think this one is pure opinion. It is my opinion that this would not cost CA a lot (one full-time salary at most). I don't think this has ever been tried, so could be a big success for CA. Worst case scenario this could be scrapped if players don't like it.

If you got to this point you read everything (assuming you did not just skip to the last sentence to get to the conclusion), so you probably have an opinion. Please share :)

So what does everyone think?

Comments

  • Thorien_KellThorien_Kell Senior Member Posts: 1,529Registered Users
    Thing is, devs base the prices by unit stats and possibly multiplayer performance.

    I'm not sure how detailed or complex their algorithms are, but I assume they do track basic unit performance.

    Problem with your solution is that you assume that people know what units are best - popular unit (or lord) often reflects not it's direct power but his utility /ease of usage or current meta fit.

    All that you would achieve is to ruin favorite picks and erratic meta shifts. Furthermore having armies that weekly vary in prices would annoy the hell out of anyone.

    And the game wouldn't necessarily be more balanced then it is now. I even believe that most of the units are priced pretty fairly, all they need to do is careful removal of overused-abused abilities and spells and we all know which one they are.


  • WojmirVonCarsteinWojmirVonCarstein Posts: 416Registered Users
    @Thorien_Kell, I agree that the having the armies vary in price every week might be a bit annoying, but remember that these shifts would not be large. Perhaps a popular build might go up by 100-200 gold. Nothing game breaking.

    I actually think that it does not matter if people know which units are best. Just like in the market, prices are not set based on some objective quality but by PERCEIVED quality and demand. (that's how brands work for example)

    "popular unit (or lord) often reflects not it's direct power but his utility /ease of usage or current meta fit."

    That's fine, if people take a lord because he is easy to use and don't take another lord that is hard to use, then it only makes sense that the lord which is easy to use goes up in price. Perhaps at some point, better players will realize that the premium they pay for "ease of use" is not worth it.

    The other cool side-effect of this would be that it would actually be worth it to go "against meta". Because meta builds would be slightly more expensive than non-meta builds. Units that no one uses would be significantly cheaper and thus a cunning player might start using them. So you would likely see more varied lists.
  • ValkaarValkaar Junior Member Posts: 1,650Registered Users
    Nah.

    Lots of races and builds are specifically balanced around only a handful of units doing the bulk of the damage. Units designed to tank, defend, harass, hold, shut down artillery, buff, heal, etc. would all ‘underperform’ under this system and become super cheap.

    While ANY units designed to do damage would be perceived as ‘OP’, even though they were just doing their job, and they’d become ludicrously expensive.

    End result would just be armies of runesmiths fighting legions of revivification bastilidons basically. It would be super dumb.

    And I don’t care if it would only be super dumb for a week before it rebalanced itself. That doesn’t justify it. The system would constantly be trying to balance itself in favor of utility units instead of dps even after it ‘fixed’ itself.

    I’d rather have a game that’s slightly (or heavily if you’re a Skaven player) imbalanced, but in a predictable and sensical way that players can predict......rather than a ****, wildly changing balance that never settles on anything, and constantly skews in favor of nonsense...no thank you.

    It would eventually kill any strategy in picking your army in that trying to rememorize all the new gold costs for every race, so you could guess your opponents build and build your own accordingly would become impossible. You’d have just as much luck intelligently counter picking your opponent by queuing up with a randomly generated army.

  • Green0Green0 Posts: 4,119Registered Users
    you have too high expectations of CA devs, they currently barely manage to stay on schedule with DLCs and give us a bandaid patch every 4 months or so to gamebreaking bugs or stupidly overpowered units/spells.
  • WojmirVonCarsteinWojmirVonCarstein Posts: 416Registered Users
    @Green0, I know that the suggestion seems like a tall order and like a lot of work. But the thing, it's not as hard as it seems. I don't want to come off as a know-it-all or whatever, but I really do know what I am talking about when it comes to the theoretical/technical implementation of this

    @Valkaar, either I wasn't clear enough or you did not read my entire post. The value of a unit does not go up or down based on "DPS". Let's look at your statements:

    - "Lots of races and builds are specifically balanced around only a handful of units doing the bulk of the damage."
    I agree with this. What you are worried about is that these units would go up in cost (because they are taken often) and therefore, the faction would become unplayable and/or under-powered.

    But fear not! Because the cost of a unit would not only be adjusted based on how often the units are taken, but also (and maybe even more importantly) on how well an army does and which players pick them (higher ranked vs lower ranked ones). So, for example, if an army always includes units A & B (since they are the only DPS units), but the army is average-tier and loses as much as it wins, the cost of these units will probably stay very similar (maybe have a tiny increase of say 10 gold). Given that A & B are used very often, it means that there are other units in this army which are under-used. Because they are under-used, their cost will go down.

    What about if an army is top-tier and uses mainly units A & B. In this case, it would seem that unit A & B are in fact a bit too strong, because, the army is winning a lot AND those 2 units are always included.

    What if an army is low-tier? In that case, units A & B might stay the same price.


    - "End result would just be armies of runesmiths fighting legions of revivification bastilidons basically. It would be super dumb."

    This is extreme exaggeration and this would never happen. Remember, that the whole eco-system would be balanced. I really don't see this situation ever happening. In fact, if it did, the cost of those runesmiths would go up since they were always being used!

    - "And I don’t care if it would only be super dumb for a week before it rebalanced itself."
    this is not how it works. The process is gradual. It would not jump around like that

    - " wildly changing balance that never settles on anything, and constantly skews in favor of nonsense."
    There would be no wild changes. nothing would skew in favor of nonsense. costs would adjust slowly, up or down. Remember also, that this process would be monitored manually, so if it turn out that some unit was dropping way too low in cost, it would be a good indication that either no one is taking it, or it consistently is included in losing builds. Which suggests it needs a re-work

    - "It would eventually kill any strategy in picking your army in that trying to rememorize all the new gold costs for every race"
    You don't need to memorize anything, the cost of units is visible in army selection screen. Why would you need to memorize for every race? Remember that the costs would not be drastically different. Consider an example of vanilla goblins. Right now they are 300. Let's say that no one includes them now. After next update, maybe they go down to 285. Maybe only slightly few more people take it, next iteration, the cost will go down to 275. Now maybe they are cost effective and people start taking them. Remember that since the process is gradual, it's not like cost will drop from 300 to 200 and then everyone takes them all the time and then again it jumps to 280 to compensate etc.





  • ExarchExarch Posts: 575Registered Users
    Not a good idea for me. Not only would a weekly fluctuating balance cause problems, this entirely ignores stats based balancing and intended unit roles in its approach, instead of the statistics informed human balancing currently used.

    For example, if transformation of kadon is being chosen a lot successfully atm, because the Manticore summon is OP, the better response is to nerf the summon, and not increase the cost.

    Similarly, some units might have a low pick rate /success rate because they synergize badly with the rest of the roster, or because they are overshadowed by a better unit in their role. Making them cheap enough to be attractive, might imbalance other matchups, as they become more cost efficient in that role than a similarly statted unit in another faction which relies upon them. Or a unit can be a one of a kind in its roster.

    The final problem is that the number of variables (unit combos, faction matchups, player skill gaps, maps) is very high, and the sample size is not especially high given the playerbase, so the weekly statistics are unlikely to be robust. Not only that, the rate of information spread about the new balance will not be instantaneous, which will limit the rate at which adjustments can be made.
  • ValkaarValkaar Junior Member Posts: 1,650Registered Users
    edited March 23

    @Green0, I know that the suggestion seems like a tall order and like a lot of work. But the thing, it's not as hard as it seems. I don't want to come off as a know-it-all or whatever, but I really do know what I am talking about when it comes to the theoretical/technical implementation of this

    @Valkaar, either I wasn't clear enough or you did not read my entire post. The value of a unit does not go up or down based on "DPS". Let's look at your statements:

    - "Lots of races and builds are specifically balanced around only a handful of units doing the bulk of the damage."
    I agree with this. What you are worried about is that these units would go up in cost (because they are taken often) and therefore, the faction would become unplayable and/or under-powered.

    But fear not! Because the cost of a unit would not only be adjusted based on how often the units are taken, but also (and maybe even more importantly) on how well an army does and which players pick them (higher ranked vs lower ranked ones). So, for example, if an army always includes units A & B (since they are the only DPS units), but the army is average-tier and loses as much as it wins, the cost of these units will probably stay very similar (maybe have a tiny increase of say 10 gold). Given that A & B are used very often, it means that there are other units in this army which are under-used. Because they are under-used, their cost will go down.

    What about if an army is top-tier and uses mainly units A & B. In this case, it would seem that unit A & B are in fact a bit too strong, because, the army is winning a lot AND those 2 units are always included.

    What if an army is low-tier? In that case, units A & B might stay the same price.


    - "End result would just be armies of runesmiths fighting legions of revivification bastilidons basically. It would be super dumb."

    This is extreme exaggeration and this would never happen. Remember, that the whole eco-system would be balanced. I really don't see this situation ever happening. In fact, if it did, the cost of those runesmiths would go up since they were always being used!

    - "And I don’t care if it would only be super dumb for a week before it rebalanced itself."
    this is not how it works. The process is gradual. It would not jump around like that

    - " wildly changing balance that never settles on anything, and constantly skews in favor of nonsense."
    There would be no wild changes. nothing would skew in favor of nonsense. costs would adjust slowly, up or down. Remember also, that this process would be monitored manually, so if it turn out that some unit was dropping way too low in cost, it would be a good indication that either no one is taking it, or it consistently is included in losing builds. Which suggests it needs a re-work

    - "It would eventually kill any strategy in picking your army in that trying to rememorize all the new gold costs for every race"
    You don't need to memorize anything, the cost of units is visible in army selection screen. Why would you need to memorize for every race? Remember that the costs would not be drastically different. Consider an example of vanilla goblins. Right now they are 300. Let's say that no one includes them now. After next update, maybe they go down to 285. Maybe only slightly few more people take it, next iteration, the cost will go down to 275. Now maybe they are cost effective and people start taking them. Remember that since the process is gradual, it's not like cost will drop from 300 to 200 and then everyone takes them all the time and then again it jumps to 280 to compensate etc.

    It’s all well and good to SAY that cost changes would be gradual and would never dramatically change....but in order to limit that, you would need exponentially more manual oversight than what CA currently utilizes to patch multiplayer. This couldn’t be an autonomous algorithm.

    True ‘supply and demand’ pricing would allow for wonky and dramatic price changes.

    Artificially capped supply and demand changes....say ‘no unit’s price can change more than 15% per week’ or whatever gradual limit you set would fail to achieve any meaningful balance. Because each week a new unit would need more sharp tuning than that.

    No automated formula would be able to accurately make those calls....is 15% enough here...or too much...or did it flag the wrong unit anyway cuz it was a drop hacker win...is this unit picked more often cuz this unit more common in X match up, which is a more common match up cuz a DLC just launched?...etc etc....the algorithm won’t know....

    And if you’re going to argue that a developer is going to REVIEW all of these weekly changes, in tandem with community feedback....to make sure it’s all appropriately handled.......if CA actually had the time/resources to do that....they could also just....balance the damn game, and there would be no need for an ineffectual supply and demand algorithm in the first place. The ‘review’ guy could just make the appropriate balance changes sans algorithm entirely.

    And yes, you do need to know approximately what opposing race’s units cost to know what kind of builds to be prepared against and what units to focus fire, etc. etc. if you have no idea what enemy units cost, you’re just flipping a coin as to what to pick and what to shoot to cause your enemy the cost-efficient damage.

    As a side note....basing this whole thing on the top 100....when the top 100 isn’t power matched to begin with....is...well....horrible. The top 100 is a huge number of draw kiters, drop hackers, single-entity spammers, all missile cav cheesers, etc. etc. And yes....also actually a few good players. Still, since they aren’t matched against each other, most of the top 100 have figured out nonsense builds to beat up on noobs with, and just repeat it. Most of them don’t even get randomly defeated by the good players, because the good players have likely already privately banned them so the whole fight/potential loss for the cheeser gets dodged by one party or the other.

    If you ACTUALLY want a balanced ladder, there is no easy gimmicky formula to do it. If there was, professional esports developers would already be doing it.

    Balance and a meaningful ladder will come from matchmaking....balance patches more frequent than once every 6 months, not using roster cuts to feed pay to win DLC....and an overall objective desire for CA to WANT to balance the game...which likely won’t happen until more people play multiplayer...which won’t happen til CA puts more effort into multiplayer...in a weird chicken vs the egg situation.

    Balance will not come from any automated formula. The templates for how to asymmetrically balance an game already exist within the numerous esports based RPG/RTS/FPS games that already do it for a living.

    Most likely, CA won’t change anything they’re doing. But if they do change their approach to multiplayer, I would EXPONENTIALLY prefer that they look to somewhere like Blizzard or CCP for inspiration...not to wherever you supposedly got the ‘technical education’ you claim to have received that taught you this was a good multiplayer balance programming idea.
  • ParmigianoParmigiano Posts: 481Registered Users
    This is possibly the worst idea ever. Each change should be intended for that to be the final change for that unit, the fact that they do balance updates is just that certain egregious issues remain. That is how it will achieve the state of being a quality final product.

    They need to have an opinion on how the unit is performing in battles to match the gameplay that they want it to be. So when people are bringing up balancing issues, they're seeing the scenarios in the battles and realizing that it's not going to be good for the game for it to remain like that. When people are seeing that they have their bias, CA is the final call on that.

    So for example the revivication crystal was 1000, they switched it to 1025 and now back to 1000. So that whole time it was 1025 was wrong. If they finalized it at 1025 then that's fine. They need to have an opinion on how the units should be performing. If I was buying from someone on a weekly basis and they kept switching the cost according to their whim I would be insulted.
  • cool_ladcool_lad Senior Member IndiaPosts: 2,233Registered Users
    There's one rather problematic presumption at the base of this elegant theory; that cost precedes utility. The OP is presuming that the utility of a unit is decided in a large part by its cost when the relationship is in fact the inverse of that.

    The cost of a unit is a reflection of it's utility. The cost is ultimately decided by a combination of just how much that unit actually offers to the army, combined with the overall performance of the army as a whole and just how appropriate the "feels" vis a vis the role and vision the devs had in mind for that unit. All of this is then approximated into a value, which represents just how much utility the unit brings to that army.

    A unit is made, then a cost is assigned to it; not the other way around. True, the MP balancing is caught in what can only be described as a vicious spiral of counterpicking and specialised builds, but that not going to be solved by addressing what is essentially the product of the process instead of the flawed process itself.

    And that's before we come to the problem that is is basing such a model on the idea of a player base composed of the elusive "Homo Economicus" even at the topmost stage, or for that matter presuming that a free market can or should exist within the context of asymmetrical balancing (a simple example is the value of frontline troops, which would spiral out of control due to their extremely high pick rate, or the deflation of values of elites which would not be used as much).

    Add to all this the excessive number of caveats that the OP is proposing for the whole system, such as manual rebalancing of individual costs and community feedback, and what you get is a system that seems little to no better than the present one; one which may very well present the risk of hiding issues with units by constantly shifting costs instead of addressing the actual problem with the unit in question.

    Balancing extends far beyond mere cost and role, and requires far more than mere price adjustments (which is ultimately the most superficial form of balancing) in order to actually be effective.
  • WojmirVonCarsteinWojmirVonCarstein Posts: 416Registered Users
    @cool_lad

    Thanks for a detailed and well thought out reply.

    I do see your points and I definitely agree that what I proposed is not a perfect solution.

    I agree strongly with your example of front line troops costs going up because everyone chooses them so there would have to be some sort of restriction placed on these. Of course this would add complexity and then it would not longer be "free market" thing.

    I guess I have not thought everything through. This is why discussions are so important.

    I think CA needs to do a much better job in balancing Multiplayer by listening to the community.

  • AerocrasticAerocrastic Posts: 317Registered Users

    @cool_lad

    Thanks for a detailed and well thought out reply.

    I do see your points and I definitely agree that what I proposed is not a perfect solution.

    I agree strongly with your example of front line troops costs going up because everyone chooses them so there would have to be some sort of restriction placed on these. Of course this would add complexity and then it would not longer be "free market" thing.

    I guess I have not thought everything through. This is why discussions are so important.

    I think CA needs to do a much better job in balancing Multiplayer by listening to the community.

    They do listen to the community, but the community is often divided on how certain things should be "fixed" or implemented and CA has their own vision of the game to adhere to as well as a relationship with the owner of the Warhammer IP Games Workshop to consider.

Leave a Comment

BoldItalicStrikethroughOrdered listUnordered list
Emoji
Image
Align leftAlign centerAlign rightToggle HTML viewToggle full pageToggle lights
Drop image/file