Countering Influence Inflation

[Sidrath]

As has been pointed out by many others on these forums, Influence currently has many sources but too few sinks. You can earn it from battles, from joining armies, from forming armies with your own clan members, from towns (via forums) and from policies (typically tied to land ownership, e.g. Castle Charters or Council of the Commons). You only really spend it for forming armies and for bidding on fiefs you desire. You can of course support other Lords' bids and earn Relation with them (or outright spend Influence on Relation via the Kingdom interface), but I've barely ever seen them reciprocate outside of the early-game edge case of being fiefless (at least in my playthroughs, I haven't seen my Relation with other Lords have any impact on their propensity to support the player's bids).

It is apparent, from another thread on Food Shortage (https://forums.taleworlds.com/index.php?threads/food-shortage.422990/), that Influence inflation is a major cause for concern. Two issues come to mind, both of which negatively impact the long-term health of a Bannerlord playthrough:

1. Ruler Greed: A ruler's Influence, a few years in, will sometimes have snowballed so much that the Influence cost of them overriding their vassals, and awarding themselves everything, is a trifle. The result is 'Ruler Greed': Kings owning many towns, earning ever-more Influence from their forums, with which to further overrule their vassals.
   
   
2. Army Formation: In successful kingdoms, a great multitude of vassals are capable of marshalling the entire kingdom. It doesn't matter that you smashed their first 2k doomstack, another Lord has enough influence to immediately summon the next doomstack. Without Influence inflation, they couldn't do that.

Perhaps Ruler Greed should be addressed via a more coherent model of the Relations impact on vassals, rather than via a change to Influence. I personally believe so, but also suspect that an appropriate behaviour model for Lords to act in accordance with Relations, and their personality traits, is a LONG way off for Bannerlord. Realistically, we need short-term, functional fixes while longer-term modelling can be worked on.

Independently of issue #1, Influence Inflation is a genuine aspect of the game currently: as a currency, its value steadily debases the longer your playthrough runs. In lategame, influence becomes somewhat meaningless: as a player, you have enough to keep your kingdom permanently marshalled, and enough to max-bid everything you want. I have rarely seen kingdoms manage to field back-to-back armies in Years 1 & 2 of a playthrough, partly because few Lords have the spare Influence to raise armies after the first defeat or two. The early game feels realistic and Warband-like: army vs army battles should be rare, epic moments (which in early/midgame, pre-inflation, they are!), and winning one should feel like a kingdom-wide triumph.

To that end, I'd propose a simple means of preventing any clan from hoarding too much Influence. Much as Warband had decaying Renown (sit on your hands for too long, and people slowly forget about your deeds), Bannerlord could benefit from decaying Influence. My suggestion: make Influence decay by a percentage every day, so it becomes unfeasible for Influence to inflate beyond a certain level. The decay rate could be zero up to some basic amount, and scale up to penalise more aggressively those hoarding at the top of the pyramid.

Formally, tying this to my biggest concern for Bannerlord - game-ending snowballs - you could make this decay rate scale as a function of:

1. A Lord's current influence pool: 0% decay rate till 100 Influence, scaling up to say 2% per day decay rate at 1000 Influence. Losing 20 Inf per day probably won't happen: the Lord's Inf will probably stabilise (gains from town forums, policies, etc offsetting decay) somewhere between 500 and 1000. High enough to ensure major Lords can still run decent armies and make the occasional big bet on a fief, low enough to preserve the merry-go-round of fief allocation and prevent the formation of those 2k doomstacks - no one could hoard enough Inf to marshall the entire snowballing kingdom at once (incidentally, armies of more modest size don't starve anywhere near as much as the slow, oversized doomstacks - so this would address that glaring AI defect, albeit in a roundabout way).
   
   
2. The power of the kingdom: The more fiefs the kingdom owns as whole, the more challenging it should become for any one Lord to exert much influence. Model that disunity / logistical overhead as an additive factor on the Influence decay rate, penalising larger kingdoms and giving smaller, more streamlined kingdoms more scope to bounce back.

From an implementation standpoint, this suggestion should be pretty straightforward. Bannerlord already implements scaling Influence decay, after all! As a Merc clan, you 'decay' more daily Influence into cash, the more you have. The logic would be the same but on a different sliding scale for vassals, with a view to preserving the integrity of the Influence system and its consequences on fief allocation, army management and the relative power of small vs large kingdoms.

 

[1115141208114]

They need to just overhaul influence. I honestly hate these mana-type resources in games. It doesn't work in Paradox games and it doesn't work here. Influence shouldn't be a spendable resource, it should be similar to your renown. I.e, how famous and well know you are should determine your ability to influence people.

 

[TheShermanator]

As noted, we're talking about this on this other thread (with a different title) https://forums.taleworlds.com/index.php?threads/food-shortage.422990/ but in brief: I am a +1 for a system that scales influence costs, directly or indirectly to the size/power of the ruler (e.g. some cost multiplier that would be tied to the total prosperity of all settlements held by any given clan). Because the size of the faction, by # of fiefs + prosperity, allows for scaling influence farming, but influence costs remain flat, so runaway influence bloat feeding runaway snowballing - more influence gets you more fiefs which gets you more gold and influence which help you get more fiefs, etc. etc. - is pretty much guaranteed for powerful clans (and thus the player) over time.

Don't know if you folks have played Endless Space 2, but that game presents a fairly simple model for scaling influence costs: There are various policies that apply a bonus to all faction holdings (star systems containing planets), but the influence cost of implementing them is linearly higher relative to the total population count on all planets. Of course, the more holdings a faction has, the more opportunities it has to farm influence - both by building improvements and just by population #s. So the influence costs of policies scale proportionate to the scaling of influence farming.

 

[Badcritter]

I have a number of thoughts about the Influence system which I will type up if I have time. Just for now I thought it worth listing out the sources and sinks for Influence currently, as many of the sinks are often forgotten (and not used).

Sources:
Forums (passive, once built)
Certain laws (passive, if in effect)
Tournament wins
Battle wins
Being in an army (having your companions in your own army is effectively an exploit)
Donating troops to garrisons
Donating prisoners

Sinks:
Votes for fief allocation
Raising armies
"Support" (buy relations with clans in the same kingdom)
Proposing a law, or proposing the revocation of a law
Voting on law proposals
Certain laws (passive, if in effect)
Proposing the re-allocation of a fief
Proposing to kick a clan

Would there still be Influence inflation problems if more of these sinks were worth doing? The answer is yes, because there are fundamental conceptual problems with the Influence economy that make it impossible to balance given its current ruleset. But in considering how to make the Influence system work more effectively, it is still worth factoring in how things will be if and when the dis-used sinks become useful in practice.

 

[somethinglightthat]

Don't know if you folks have played Endless Space 2

I'm going to get around to recording a play on that sometime. I like that game, sometimes, when it gives me cookies.

I like the idea that influence always decays into denars at some scaled rate, that way it doesn't just vanish for things you've done to earn it, but it isn't good for being influential forever. Given that voting can use 300 at a time, decay might not need to start until much later than that.

Maybe the bigger problem is that influence cost to raise an army doesn't scale with army size. Anything more than the player can come up with between companions and max steward skill should get progressively more expensive, and hence rare, to form. This has to be balanced with the size of defensive armies found in cities and castles to make siege viable.

IMO

 

[Badcritter]

Or maybe it costs denars to prevent your influence decaying. It's expensive holding onto power. Not to mention that denars require inflationary control almost as much as influence does.

I find it strange that the cost of getting a party into an army is a one time fixed cost (albeit for a limited period of time before coherence decays) but the benefit received by a party staying in an army is progressive over time. I mean, I understand why it has been done this way - but how can you ever know how effectively balanced the two will be when one side of the coin is an apple and the other side is an orange (pardon my mixed metaphore)?

It would make more sense to make a payment of influence from the leader to the member party each day. Maybe the first payment is bigger, or there is variance for while the party is in transit compared to when they arrive. But the important part is that the base price paid increases slightly from one day to the next so that it eventually becomes impractical to keep the army together. The inflated cost is tracked between the leader and the member and decays back to the minimum rate over time, so that the leader can't simply disband and reform to get around it.

The leader of the army would command higher influence reward from the battles won than his followers get, and being the leader of the army that takes a fief would need to hold more weight than it currently does. Basically, it would make leading an army an important decision - not just something you do because the AI is dumb and you can do it infintely better. Hopefully the AI will get less dumb over time too.

 

[somethinglightthat]

but how can you ever know how effectively balanced the two will be when one side of the coin is an apple and the other side is an orange

When it works, I guess. I tend to value results over what makes sense, it's nice to have both. I think if you tie influence to an ongoing cost then you'll have armies disbanding from shortsighted AI decisions, as target decision making seems to be done after the army is made, so there wouldn't be much or any consideration of this when the army is formed, leading to bad results.

It's not an apple and orange but you get the idea.

 

[Badcritter]

Love the image

When it works, I guess.

My point is that it never will. Or I suppose, like a broken clock it might be right twice a day but it will be wrong the rest of the time. The goal when balancing gameplay systems is mostly about creating the right amount of tension, and if the system is too unpredictable you can't tune towards creating tension with the right frequency or timing. Meaningful moments will still pop out of uncontrolled systems, but it only happens by luck.

This is a challenge for sandbox games in general. But good sandbox games put just enough controls in place to ensure that meaningful things happen more frequently and with better pacing than being purely random.

As for AI decision making, since it is already struggling with tying its own shoelaces it obviously won't handle additional complexity. But it will get better, because it has to. So even though it pays to be mindful of the practical limits on how much better the AI will get, we shouldn't limit our thinking to what the existing AI is capable of now.

 

[somethinglightthat]

So even though it pays to be mindful of the practical limits on how much better the AI will get, we shouldn't limit our thinking to what the existing AI is capable of now.

On the surface I agree to think without limitation, in practice I think these things must be developed interdependently.

 

[Badcritter]

I didn't say to think without limitation, I said to be mindful of the practical limits of the future AI. Basically work on the assumption of limitations that are reasonable to expect at the future time when the system will come to fruition. You hit the nail right on the head that the development of both are interdependent (as are most game systems).

 

[somethinglightthat]

I'm not sure that the AI will improve in it's ability to plan objectives at the time of army formation. The army forms and cohesion is an easy metric to work with since it gives a predictable amount of time to live, it decays at a known rate. Influence would be a moving goalpost that would change based on other events, requiring a lot more changes to the way the AI evaluates things in the way I currently understand it. If my guess (and that's all it is) is correct if there's any post formation evaluation of a safe place to disband it is minimal if it exists at all. What's more likely is that there's some evaluation during objective selection (after formation, that is evaluated conditionally and every so often) based on cohesion, a known time, and if you change that to an unknown time, the AI would have to select this (objectives) based on some other, much more complex factors, that can and will lead to odd choices without a lot of testing time.

It's likely that such development is outside the scope of what they have in mind at this stage, I could be proven wrong. It's just my opinion, and its based on some assumptions of how things are done based on what I've played and read on this forum.

 

[Sidrath]

A few interesting thoughts already emerging from the discussion - will paraphrase some of them and add thoughts.

1. Make the Influence cost of actions scale with the size of the faction and/or ruler clan: I see scaling up Influence costs as a functional alternative to my suggestion of scaling down net Influence gains (via decay). However I would be cautious: increased Influence costs will be something only the largest clans of a snowballing kingdom can afford. All the lesser clans with low Influence income (including, early on, the player), may find themselves unable to vote on policies / fief allocations, breaking the merry-go-round. Arguably trickier to tune than decay rates.
   
   
2. Make army management cost Influence per day, rather than a fixed Influence cost at the start: Conceptually, I like this. Insofar as I like balanced, non-inflationary systems, there's a definite elegance to having army raising become a zero-sum Influence transaction: the army leader pays daily Influence to the attending Lords. I'd still want army raising to be encouraged behaviour rather than a burden to the leader, so I'd couple these (over time, larger) payments of Influence with a few perks: 4x Renown and Influence gain from battles for the army leader, 2x Renown and Influence gain from battles for army followers. As it stands, most players probably get a high proportion of their Renown from farming Looters/bandits with low tier units, vs what should be the true source of renown: leading and participating in massive battles. The game currently favours "combat odds" far, far too much in the calculation of Renown/Influence gain, as opposed to sheer scale of confrontation - but that's another topic.
   
   
3. Quick band-aids vs better AI: I am very keen on the deployment of simple, quick band-aids. This is not out of mistrust of a more elaborate AI - on the contrary. Based on the evolution of the Early Access phase, I'm just not convinced that robust AI systems will be developed in a timely fashion, if ever. Moreover, I have yet to carry a single character to a very high level of combat skill proficiency (weapon level speed feels out of sync with progression in e.g. Charm or Stewardship), or level (my highest was maybe Level... 23?), or even to Clan Tier 6. Most playthroughs are 'settled' before Clan Tier 6 and before your first child is of age, so the lineage mechanics, wedding between clans, etc. are all systems we can't even begin to test. Getting simulations (i.e. player passively sitting in a town, watching 40 years unfold) to evolve in realistic, balanced ways should be a priority - and that means locking down inflationary spirals, which both Influence and wealth currently suffer from.

 

[TheShermanator]

A few interesting thoughts already emerging from the discussion - will paraphrase some of them and add thoughts.

1. Make the Influence cost of actions scale with the size of the faction and/or ruler clan: I see scaling up Influence costs as a functional alternative to my suggestion of scaling down Influence gains (via decay). However I would be cautious: increased Influence costs will be something only the largest clans of a snowballing kingdom can afford. All the lesser clans with low Influence income (including, early on, the player), may find themselves unable to vote on policies / fief allocations, breaking the merry-go-round. Arguably trickier to tune than decay rates.
   
   
2. Make army management cost Influence per day, rather than a fixed Influence cost at the start: Conceptually, I like this. Insofar as I like balanced, non-inflationary systems, there's a definite elegance to having army raising become a zero-sum Influence transaction: the army leader pays daily Influence to the attending Lords. I'd still want army raising to be encouraged behaviour rather than a burden to the leader, so I'd couple these (over time, larger) payments of Influence with a few perks: 4x Renown and Influence gain from battles for the army leader, 2x Renown and Influence gain from battles for army followers. As it stands, most players probably get a high proportion of their Renown from farming Looters/bandits with low tier units, vs what should be the true source of renown: leading and participating in massive battles. The game currently favours "combat odds" far, far too much in the calculation of Renown/Influence gain, as opposed to sheer scale of confrontation - but that's arguably another topic.
   
   
3. Quick band-aids vs better AI: I am very keen on the deployment of simple, quick band-aids. This is not out of mistrust of a more elaborate AI - on the contrary. Based on the evolution of the Early Access phase, I'm just not convinced that robust AI systems will be developed in a timely fashion, if ever. Moreover, I have yet to carry a single character to a very high level of combat skill proficiency (weapon level speed feels out of sync with progression in e.g. Charm or Stewardship), or level (my highest was maybe Level... 23?), or even to Clan Tier 6. Most playthroughs are 'settled' before Clan Tier 6 and before your first child is of age, so the lineage mechanics, wedding between clans, etc. are all mechanics we can't even begin to test. Getting simulations (i.e. player passively sitting in a town, watching 40 years unfold) to evolve in realistic, balanced ways should be a priority - and that means locking down inflationary systems, which both Influence and wealth currently suffer from.

With respect, on (1), it doesn’t look like you’ve quite considered what scaling means here. There should be no need to worry about small players lacking the influence to participate because their costs wouldn’t be higher - precisely because they’re small. If costs scale with size, those with more fiefs (more ways to farm influence) would have to pay more to do stuff, not players with less. Right now, there’s no scaling - every action has a flat influence cost. (If anything, the issue is the opposite: how to make it worthwhile for the bigger voting clan if it costs more.)

Now, this would still be hard to balance; not suggesting it’s easy. E.g. It’s clear to me why policies should cost more for the bigger player, as the policies go further for them. But perhaps other expenditures- e.g army enlistment -should not scale? Tough questions like that.

 

[Sidrath]

Ahh, my apologies! I misunderstood the scaling: you meant for it to scale the agent's Influence costs in function of the agent's own power base, not the ruler's or the kingdom's. I could definitely see that working - slightly more variables to tune (which expenditures to scale and which to keep, as you point out), but it should work. Bit mindful that the justification is harder to grasp: unsure why (gameplay considerations aside, from a pure logic perspective) the influential should pay more to propose a policy than, say, a nobody clan. The overarching notion that 'Influence decays if you sit on your hands' rests a tad more easily with me.

From a gameplay perspective alone however? Totally on board with your suggestion, I suspect either approach will yield an improvement in game balance.

 

[TheShermanator]

Ahh, my apologies! I misunderstood the scaling: you meant for it to scale the agent's Influence costs in function of the agent's own power base, not the ruler's or the kingdom's. I could definitely see that working - slightly more variables to tune (which expenditures to scale and which to keep, as you point out), but it should work. Bit mindful that the justification is harder to grasp: unsure why (gameplay considerations aside, from a pure logic perspective) the influential should pay more to propose a policy than, say, a nobody clan. The overarching notion that 'Influence decays if you sit on your hands' rests a tad more easily with me.

From a gameplay perspective alone however? Totally on board with your suggestion, I suspect either approach will yield an improvement in game balance.

Agreed - I’d be on board w a both and (scaling costs + decay) approach.

The scaling costs thing would make realistic sense ONLY if we could establish that the clan (AI or player) was getting more for the influence spent relative to the clan with the lower costs. E.g. If a policy ties a resource (eg influence) gain rates to # of fiefs, it makes sense that the clan would have to spend more influence to get the policy passed. But maybe not so much sense for other things...

 

[Badcritter]

Agreed - I’d be on board w a both and (scaling costs + decay) approach.

They are both potentially useful tools for making it possible to create a sustainable influence system. There are other possibilities too, like capped accumulation with more dynamic spending profiles, zero sum designs.

Without going all the way up the chain to dismantling the question of what really is the purpose of the influence system, so that the best way to achieve that can be determined (obviously that's the designer's job)... the most important thing on my mind at this moment is seeing that the system is nudged in the direction of at least having a conceptual chance of working sustainably.

 

[TheShermanator]

Without going all the way up the chain to dismantling the question of what really is the purpose of the influence system, so that the best way to achieve that can be determined (obviously that's the designer's job)...

Right. And I think a big part of that has to be incoming features too - something we haven't considered. E.g. My favorite rant, the need for real rebellion and secession mechanics, with negative relations pressure that scaled with size / # of vassals in kingdom, and thus the need for the ruler to spend more and more influence (in new sub-features or the existing "Support" button) to preserve relations and prevent secession, rebellions, civil wars, etc. It's always possible that there's a reason that we have too much influence - maybe we're supposed to spend it on something that isn't really in the game yet!

 

[TheShermanator]

The leader of the army would command higher influence reward from the battles won than his followers get, and being the leader of the army that takes a fief would need to hold more weight than it currently does. Basically, it would make leading an army an important decision - not just something you do because the AI is dumb and you can do it infintely better. Hopefully the AI will get less dumb over time too.

+1. Being an army leader should be more influence intensive on the cost side while also promising much higher potential rewards for success. E.g. Giving rewards for conquering fiefs even if you don't get to keep the fief. This would make it harder to become an army leader, but the player would want to do it a lot. (I find that it's not that hard to get past the mercenary hanger-on stage and on to the lower tier vassal-calling-armies stage. I think it should be harder for lower tier vassals with less influence income to be those army leaders.)

 

[SturgiaStrong]

Introducing influence is one of the worst decisions in Bannerlord. Mana resource is always bad gamedesign decision

 

[TheShermanator]

Introducing influence is one of the worst decisions in Bannerlord. Mana resource is always bad gamedesign decision

For politics, engaging and using other friendly clans, etc. what would you do instead? Just like a renown system - where higher renown clans have a higher dice roll chance of getting policies passed or enlisting parties to their armies?

I also remember Attila Total War (yes, I skipped Rome 2), which had an internal politics system of sorts, and it used a -X to +X dial system for Dominion, Control, and Political Power - a system in which those 3 measurements all interacted with each other. So no bankable influence currency. It was hard in the sense that it didn't get easier the bigger you got; if anything, the bigger you got, the more likely you were to suffer civil strife. But it was kind of unwieldy and generally unpopular (I didn't hate it FWIW).