well, there'll always be some idiot, at least we've found ours
-
If you are reporting a bug, please head over to our Technical Support section for Bannerlord.
Sturigans is more weak after update?
- Thread starter Qingl
- Start date
Users who are viewing this thread
Maximum997
Squire
He acts like this in every single thread. So yes, we've found ours
Or may be Sturgia in your game got stomped by unlucky rolls, lost a big battle with Vlandia and so on. So your screenshot is as valid as quoting... wait, unlike me you newer shown any screenshot. It's just you talking empty.
Non of you ever presented any evidence that Sturgia getting wiped is a norm. It's just you 4 dudes complaining that in your random games Sturgia got wiped for some random reason. Well, ques what? There is no game rule that says that Sturgia can't get wiped out, just like any other faction. But that does not make it a norm and Sturgia weak.
Ad hominems are the last resort of losers in every lost debate.
Oh my, it's as if Mathematics itself is conspiring against Sturgia. Let's dive into that mathematical conspiracy and vanquish the evil-doers together, shall we?
Credit for the autocalc function explanation as well as all screenshots attached goes fully to nexus user keehuuu . I'm only trying to toy with numbers based on my understanding of it.
When autocalc combat is initiated, it consists of "rounds". An attacking and defending side is chosen each round.. The more soldiers you have, the higher the chance that you'll be the attacking side is.
Then two random soldiers are picked, one from the attacking side, one from the defending side.
Autocalc is based on individual units' power levels, calculated mainly from their tier, or, in case of heroes, it's hero level divided by five.
The PL is then multiplied by 1.3 for any mounted unit - cavalry, horse archer, mounted companions, mounted NPC lords.
Then we have this
Damage is calculated as follows:
The "damage" variable refers to the "damage" calculated for the attacking troop divided by the PL of the defending troop. I won't go into detail here, but essentially, if a low tier unit is going against a high tier unit, then its damage will be lowered somewhat (not by much). Let's say the enemy Looter is the attacking troop, and it has a PL of 0.6. This value gets multiplied by 50, and you get 33 damage then its divided by the PL of the defender, and it becomes 12.
And Max HP is pretty self explanatory.
Therefore, cavalry only comes into play:
- when that cavalry trooper is being attacked (he does 30% more damage)
- when that cavalry trooper is defending (he is 30% more resistant to attacks)
In the absolute best case scenario, where not a single soldier of your army is incapacitated as a result of the defensive roll, the cavalry unit and his power level bonus only comes into play when he is chosen to attack. The chance of that is (Ncav/N), where Ncav is the amount of cavalry you have in your party, and N is the amount of soldiers in your army total.
So now let's deduce the net increase the following armies gain from the cav bonus:
- An army of 100 same tier soldiers, none of which are cavalry, get 0% bonus
- An army of 100 same tier soldiers, 10 of which are cav, get 30%(cav autocalc bonus) times 0.1 (the chance that your cavalry applies that bonus on each roll), = 3% average attack bonus
- An army of 100 same tier soldiers, 20 of which are cav, get 30%(cav autocalc bonus) times 0.2 (the chance that your cavalry applies that bonus on each roll), = 6% average attack bonus
The difference in attack performance between army two and army three is 1.03/1.06, so army two is 97.17% as strong as army two attack-wise.
Now, how much more bulky is army three? Same logic applies. A cav unit gets 1.3 times the PL, so he lowers the damage he's about to take to 1/1.3 = 76.9% of the otherwise . It is only an advantage in one case: when this bonus saves the unit from an otherwise incapacitating autocalc roll.
Remember, the survival of a defending unit is decided by luck - a random roll between his HP and 1 must be greated than calculated damage value. So if we were to take keehuu's example literally, a 0.6 PL looter attacking 2.75PL cavalry does 12 damage, where a soldier of the same tier would have roughly about 2.115 PL and therefore take 15.6 damage.
What's the chance of a random value from 1 to 100 to land in the range of 12 to 15.6? It's 3.6%, right. That's a rough estimate of how much a singular cav unit (that must be first picked as a defender for it to have ANY effect, and the probability of that is Ncav/N again) directly benefits defensively from the flat cav bonus. A 10% chance to get a 3.6% advantage sounds huge, I know.
So an army with twice the amount of cavalry is 1-1.06/1.03 = 2.9% more effective attack wise and marginally better defense wise. Oh boy.
(I do not claim to be a mathematician in any capacity, and numbers listed above are my interpretation of keehuu's explanation. You're welcome to correct me if I'm wrong)
Last edited:
yeah, auto-calc need way more complex codes to work properly, it should be simulating a battle, not playing dice-roll with flat bonuses to specific units on everything. Cavalry should be good at attack, yes, but not defense.Oh my, it's as if Mathematics itself is conspiring against Sturgia. Let's dive into that mathematical conspiracy and vanquish the evil-doers together, shall we?
Credit for that data as well as all screenshots attached goes to nexus user keehuuu
When autocalc combat is initiated, it consists of "rounds". An attacking and defending side is chosen each round.. The more soldiers you have, the higher the chance that you'll be the attacking side is.
Then two random soldiers are picked, one from the attacking side, one from the defending side.
Autocalc is based on individual units' power levels, calculated mainly from their tier, or, in case of heroes, it's hero level divided by five.
The PL is then multiplied by 1.3 for any mounted unit - cavalry, horse archer, mounted companions, mounted NPC lords.
Then we have this
Damage is calculated as follows:
The "damage" variable refers to the "damage" calculated for the attacking troop divided by the PL of the defending troop. I won't go into detail here, but essentially, if a low tier unit is going against a high tier unit, then its damage will be lowered somewhat (not by much). Let's say the enemy Looter is the attacking troop, and it has a PL of 0.6. This value gets multiplied by 50, and you get 33 damage then its divided by the PL of the defender, and it becomes 12.
And Max HP is pretty self explanatory.
Therefore, cavalry only comes into play:
- when that cavalry trooper is being attacked (he does 30% more damage)
- when that cavalry trooper is defending (he is 30% more resistant to attacks)
In the absolute best case scenario, where not a single soldier of your army is incapacitated as a result of the defensive roll, the cavalry unit and his power level bonus only comes into play when he is chosen to attack. The chance of that is (Ncav/N), where Ncav is the amount of cavalry you have in your party, and N is the amount of soldiers in your army total.
So now let's deduce the net increase the following armies gain from the cav bonus:
- An army of 100 same tier soldiers, none of which are cavalry, get 0% bonus
- An army of 100 same tier soldiers, 10 of which are cav, get 30%(cav autocalc bonus) times 0.1 (the chance that your cavalry applies that bonus on each roll), = 3% average attack bonus
- An army of 100 same tier soldiers, 20 of which are cav, get 30%(cav autocalc bonus) times 0.2 (the chance that your cavalry applies that bonus on each roll), = 6% average attack bonus
The difference in attack performance between army two and army three is 1.03/1.06, so army two is 97.17% as strong as army two attack-wise.
Now, how much more bulky is army three? Same logic applies. A cav unit gets 1.3 times the PL, so he lowers the damage he's about to take to 1/1.3 = 76.9% of the otherwise . It is only an advantage in one case: when this bonus saves the unit from an otherwise incapacitating autocalc roll.
Remember, the survival of a defending unit is decided by luck - a random roll between his HP and 1 must be greated than calculated damage value. So if we were to take keehuu's example literally, a 0.6 PL looter attacking 2.75PL cavalry does 12 damage, where a soldier of the same tier would have roughly about 2.115 PL and therefore take 15.6 damage.
What's the chance of a random value from 1 to 100 to land in the range of 12 to 15.6? It's 3.6%, right. That's a rough estimate of how much a singular cav unit (that must be first picked as a defender for it to have ANY effect, and the probability of that is Ncav/N again) directly benefits defensively from the flat cav bonus. A 10% chance to get a 3.6% advantage sounds huge, I know.
So an army with twice the amount of cavalry is 1-1.06/1.03 = 2.9% more effective attack wise and marginally better defense wise. Oh boy.
(I do not claim to be a mathematician in any capacity, and numbers listed above are my interpretation of keehuu's explanation. You're welcome to correct me if I'm wrong)
True. I'd be fine with attack rolls if it was differentiating unit types and adhering to at least a rock-paper-scissors kinda scenario. Units with pikes deal more damage to any cav, units with maces are more effective against armor, these kinds of conditions.
In the current implementation, however, it's rock (unmounted units) versus bigger rock (mounted units), it's all the depth there is
but pikes can't properly attack cavalry, so they'd counter-act the cavalry attack bonus, even surpass it? It needs to be complex, not like crazy complex, but battle-logic should be applied. Pikemen can't attack a cavalry, it's just impossible, they can only defend...
But I agree, weapons should be considered.
That can be dangerous rabid hole, because you can't really take in to account and properly all the factors. Is unit with pikes on the hill or in the flat valley when cavalry attacks? Can cavalry use it's mobility to flank infantry with pikes or did allied cavalry prevented that? Did pike infantry formed a shiltron in time, or was it engaged frontally by enemy infantry before cavalry attacked?
I saw games that tried to do very complex calculations for abstract combat and it newer came close to real world results. On top of that calculations like that are nightmare for overall balancing: are cavalry heavy factions too strong? If you are dealing with just one simple number, you can adjust 1.3 to 1.2 and check if that did the job. If it's multitude of factors, have fun spending your days testing and adjusting every one of them -especially if they are all supposedly based on on something realistic, in which case you will end up adding more factors to make your calculation "more" realistic. Or start redesigning faction rosters in order that faction don't get steamrolled. Which might actually damage game lore wise, because you need to make these rosters to fit complex autocalc rather then some lore.
Autocalc will newer simulate results of battle that is actually fought in game, therefore it's easier to accept that and keep things simple. Results might actually be less frustrating then if you try to make these calculations too realistic, even if intent is good.
That's very cringeworthy and pompous, especially considering it was followed up by mathematical proof of something we're all already aware of, causing a problem we're all already aware of, a problem that has now been taken into consideration by TW and partially alleviated with the latest hotfix
Despite cries from certain individuals that a flat 30% bonus apparently not causing snowballing.. when it in fact does cause snowballing as reported by the overwhelming majority of people providing feedback on these forums who are interested in making the game better, as opposed to just using certain game topics for confrontation with other forum users because they get a kick out of doing so.
Maximum997
Squire
No. It is the best way to kill any balance at all.
It is. Not more so than blaming faulty maths, but you have a point. Better not try writing such stuff again when I'm half asleep and half boiling with disgust from "faulty maths", so I apologize for the cringe you've suffered through xD
As the dev mentioned in this thread already, Sturgia's suffering from its economy. Unless we're able to prove other factors have a significant effect, it's gonna remain in conjecture department.
And before that argument is inevitably waved at me as well, a thousand people saying something doesn't make it universally true. If my calculations are correct (which they probably aren't, haha), a very scary 30% flat bonus degrades to less than 5% benefit in a case where one army has twice as much cav as the other one. And there is no worldly fairness laws that dictate the stronger army always wins, since under the hood it means that Vlandia is 2% better at doing coin flips, which in the end may not account for their assumed dominance afterall.
And since most fights happen without the player, and autocalc doesn't care for equipment quality, this can't explain snowballs either. Yada, yada, you've read these points in that thread and dozens of similar ones.
I've previously made a point in that thread or other similar one - unless we can prove that the cav bonus wrecks Sturgia, it isn't a fact, it's a belief. It might even be correct in the end, but we don't get to assume that it is. Folks believed that our planet was flat, and based on the data they had, it very well could be flat. I was interested in the cav bonus myself, and from that wall of barely coherent pompous nonsense I've written above, I can't seem to fathom how such a difference could account for Sturgia being absolutely wrecked because of such a bonus.
These 9 pages of arguments did get us a fact though - economics are identified by devs as a factor.
Maximum997
Squire
It should be as simple as possible.
You need autocalc battles for 2 reasons:
1) AI vs AI.
2) Player can clear some trash and save time(becouse i dont want to load into boring 100 v 10 battle).
1. AI vs AI.
It should be balanced, or some factions will steamroll others. More complicated system = harder to balance. Any mod = balance goes to hell.
Even if it genious and very complicated no one fkn cares, becouse it is AI vs AI, player can see only results.
2. Plyer vs trash.
It should be reasonable. For example right now you should never ever use it, becouse casualties is too high. What is point to autocalc some battles, if after that you will lvl up two tier 1 troops and loose five t6 cav? Just leave this guy and dont spend time.
It should be predictable. For example if you have 80 to 20 balance power you can ezpz autoresolve and only get few units wounded. You should know this "magic number" when you can safely autoresolve.
Or you will never ever use autocalc. For example you can take Total war warhammer 2. Sometimes it forces you to play obvious boring battles, becouse you can loose some units.And you dont need to do anything to win this battles with no casualties.Single player game shouldn't force you to do boring things.
And for that reasons autoresolve should be as simple as possible. Right now it is useless and unbalanced becouse it is not simple enough.
Last edited:
nope, else it kills any strategy and the balance becomes binary, which will show later on in the game. They should simulate the battles and that's it, add a tweak to RNG, but that's it, the factions are already balanced within battle, so you're talking ****, and any half-brain can actually come up with the maths and the code to make that balanced. In fact, they're probably already doing something like that...
Maximum997
Squire
What strategy can be in ai vs ai battle? - _-
And wtat strategy can be in looter aouoresolving?
Maximum997
Squire
So do you want to say that devs have no even half of brain, thats why tactic skill and autoresolve for player are completly useless, right?
It is indeed so. The best and most accurate auto-calc would essentially boil down to running the exact same stuff we see in battle, except without graphics. Whatever else we try and do, auto-calc results would differ from a battle where the player's afk, and the notion of godly player presence is therefore inescapable.
My thought however is that unit performance shouldn't stray too much from their actual battle performance. Not every T5 cavalry is born the same and performs the same. Individually adjusting unit powers per unit type would be a nightmare, so here I am just daydreaming again.
Regardless, the fact that auto-calculated battles and battles where my character is AWOL ducking behind his own horse in the corner of the map (or dead from the first random pila throw) can get vastly different outcomes is still concerning. Being neither a game designer nor a programmer, I've got no real solution in mind that would fix it.
CZII
Recruit
It's still 30% advantage applied to the dying chance (which is kind of backwards since you would expect survival but whatever). Infantry is 30% more likely to die than cav. How large that is depends on how high the damage is in the first place.
Let's say it's tier 4 units:
Looter PL = (2 + 1) * (10 + 1) * 0.02 * 1 = 0.660
Infantry PL = (2 + 4) * (10 + 4) * 0.02 * 1 = 1.680
Cavalry PL = 1.680 * 1.3 = 2.184
Looter Damage = 0.660 * 50 = 33.000
Infantry Mitigated Damage Taken From Looter = 33.000 / 1.680 = 19.64
Cavalry Mitigated Damage Taken From Looter = 33.000 / 2.184 = 15.11
For tier 3 it's:
Infantry PL = (2 + 3) * (10 + 3) * 0.02 * 1 = 1.300
Cavalry PL = 1.300 * 1.3 = 1.690
Infantry Mitigated Damage Taken From Looter = 33.00 / 1.300 = 25.38
Cavalry Mitigated Damage Taken From Looter = 33 / 1.690 = 19.53
For tier 2:
Infantry Mitigated Damage Taken From Looter = 33 / 1.120 = 29.46
Cavalry Mitigated Damage Taken From Looter = 33 / 1.456 = 22.66
What if a tier 4 Highwayman attacks our tier 2 units instead of a Looter?
Infantry Mitigated Damage Taken From Highwayman = 109.200 / 1.120 = 97.50
Cavalry Mitigated Damage Taken From Highwayman = 109.200 / 1.456 = 75.00
Looter PL = (2 + 1) * (10 + 1) * 0.02 * 1 = 0.660
Infantry PL = (2 + 4) * (10 + 4) * 0.02 * 1 = 1.680
Cavalry PL = 1.680 * 1.3 = 2.184
Looter Damage = 0.660 * 50 = 33.000
Infantry Mitigated Damage Taken From Looter = 33.000 / 1.680 = 19.64
Cavalry Mitigated Damage Taken From Looter = 33.000 / 2.184 = 15.11
For tier 3 it's:
Infantry PL = (2 + 3) * (10 + 3) * 0.02 * 1 = 1.300
Cavalry PL = 1.300 * 1.3 = 1.690
Infantry Mitigated Damage Taken From Looter = 33.00 / 1.300 = 25.38
Cavalry Mitigated Damage Taken From Looter = 33 / 1.690 = 19.53
For tier 2:
Infantry Mitigated Damage Taken From Looter = 33 / 1.120 = 29.46
Cavalry Mitigated Damage Taken From Looter = 33 / 1.456 = 22.66
What if a tier 4 Highwayman attacks our tier 2 units instead of a Looter?
Infantry Mitigated Damage Taken From Highwayman = 109.200 / 1.120 = 97.50
Cavalry Mitigated Damage Taken From Highwayman = 109.200 / 1.456 = 75.00
What we can already notice is that a cavalry unit is slightly better than an infantry unit a tier higher in autoresolve. But that's only as far as the battles go. It has even bigger implications for faction balance on the world map. Let's say we have two factions. One turns half of its recruits into tier 2 cavalry. The other only gets cavalry at tier 3 and only in one of the four branches. Let's dumb it down and say they get 1,000 recruits who get attacked by looters and upgraded if they survive once per iteration until they get to tier 5.
Recruits
Cav heavy faction: 1,000 recruits
Other faction: 1,000 recruits
Tier 2
Cav heavy faction: 250 t2 cav, 250 t2 inf
Other faction: 500 t2 inf
Tier 3
Cav heavy faction: 176 t3 cav, 193 t3 inf
Other faction: 88 t3 cav, 265 t3 inf
Tier 4
Cav heavy faction: 142 t4 cav, 144 t4 inf
Other faction: 71 t4 cav, 197 t4 inf
Tier 5
Cav heavy faction: 121 t5 cav, 116 t5 inf
Other faction: 60 t5 cav, 158 t5 inf
Cav heavy faction: 1,000 recruits
Other faction: 1,000 recruits
Tier 2
Cav heavy faction: 250 t2 cav, 250 t2 inf
Other faction: 500 t2 inf
Tier 3
Cav heavy faction: 176 t3 cav, 193 t3 inf
Other faction: 88 t3 cav, 265 t3 inf
Tier 4
Cav heavy faction: 142 t4 cav, 144 t4 inf
Other faction: 71 t4 cav, 197 t4 inf
Tier 5
Cav heavy faction: 121 t5 cav, 116 t5 inf
Other faction: 60 t5 cav, 158 t5 inf
It's pretty obvious who would win in a battle and the advantage that the cav heavy faction gets would be that much higher if instead of looters we had higher tier bandits. If it was forest bandits or sea raider chiefs infantry would just die out by tier 3. Obviously the situation is not that exaggerated since units don't necessarily need to be attacked to upgrade but they also need to kill more than a single unit. Chances are they will be attacked at some point before they get their turn to attack someone in autoresolve. In every single battle they win over the years of a single player campaign the cav faction will have less units dying and more units getting promoted than the rest. It adds up and it shows. There could also be disproportionate effects from marginal advantages because winning by a few men is significantly better than losing by a few men but I won't go into that.
Last edited:
Consider me beaten. There's also NPC lords having a passive XP cheat to consider (they don't have to fight at all to get some XP), but that's irrelevant to the bandit slapping example.
As for cavalry disadvantage Stugia supposedly gets out of all, it's still a moot point. Sturgia gets mounted troops at T4, being on par with Vlandia, and overshadowing the Empire with the T5 Bucellari as the only non-elite tree horsemen. Though Vlandian ET2 Squire and Imperial ET3 Equite are more accessible than ET4 Druzhinnik where the elite Sturgians get mounted (duh).
I'm not sure about the exact distribution between the elite/common types, so a direct comparison with the Empire is impossible (how many Equites it would take to compensate for the lack of Hardened Brigands vs how many T3 nobles are being recruited/trained versus T4 commons).
Actually i think You should calculate the chance that a random value between 1-100 land in range 1-15,6 and it's 15,6%.
From my calc looter has 9% chance to kill T6 cav and judging by how often it happens i think i'm not wrong in it.
Here You have my calculation on % chance that one unit will kill another based on current autcalc method.
| Tier/Tier | T1 | T2 | T3 | T4 | T5 | T6 | T1M | T2M | T3M | T4M | T5M | T6M |
| T1 | 50% | 34% | 25% | 19% | 15% | 12% | 38% | 26% | 19% | 15% | 12% | 9% |
| T2 | 72% | 50% | 36% | 28% | 22% | 18% | 55% | 38% | 28% | 21% | 17% | 14% |
| T3 | 98% | 67% | 50% | 38% | 30% | 25% | 75% | 52% | 38% | 29% | 23% | 19% |
| T4 | 100% | 87% | 64% | 50% | 40% | 32% | 97% | 67% | 49% | 38% | 30% | 25% |
| T5 | 100% | 100% | 80% | 62% | 50% | 41% | 100% | 84% | 62% | 48% | 38% | 31% |
| T6 | 100% | 100% | 98% | 76% | 60% | 50% | 100% | 100% | 75% | 58% | 46% | 38% |
| T1M | 65% | 44% | 33% | 25% | 20% | 16% | 50% | 34% | 25% | 19% | 15% | 12% |
| T2M | 94% | 65% | 48% | 37% | 29% | 24% | 72% | 50% | 36% | 28% | 22% | 18% |
| T3M | 100% | 88% | 65% | 50% | 40% | 33% | 98% | 67% | 50% | 38% | 30% | 25% |
| T4M | 100% | 100% | 84% | 65% | 52% | 42% | 100% | 87% | 64% | 50% | 39% | 32% |
| T5M | 100% | 100% | 100% | 81% | 65% | 53% | 100% | 100% | 80% | 62% | 50% | 41% |
| T6M | 100% | 100% | 100% | 99% | 79% | 65% | 100% | 100% | 98% | 76% | 60% | 50% |
| Tier/Tier | T1 | T2 | T3 | T4 | T5 | T6 | T1M | T2M | T3M | T4M | T5M | T6M |
| T1 | 50% | 34% | 25% | 19% | 15% | 12% | 41% | 28% | 21% | 16% | 13% | 10% |
| T2 | 72% | 50% | 36% | 28% | 22% | 18% | 60% | 41% | 30% | 23% | 19% | 15% |
| T3 | 98% | 67% | 50% | 38% | 30% | 25% | 82% | 56% | 41% | 32% | 25% | 21% |
| T4 | 100% | 87% | 64% | 50% | 40% | 32% | 100% | 72% | 53% | 41% | 33% | 27% |
| T5 | 100% | 100% | 80% | 62% | 50% | 41% | 100% | 91% | 67% | 52% | 41% | 34% |
| T6 | 100% | 100% | 98% | 76% | 60% | 50% | 100% | 100% | 82% | 63% | 50% | 41% |
| T1M | 60% | 41% | 30% | 23% | 18% | 15% | 50% | 34% | 25% | 19% | 15% | 12% |
| T2M | 87% | 59% | 44% | 34% | 27% | 22% | 72% | 50% | 36% | 28% | 22% | 18% |
| T3M | 100% | 81% | 60% | 46% | 37% | 30% | 98% | 67% | 50% | 38% | 30% | 25% |
| T4M | 100% | 100% | 77% | 60% | 48% | 39% | 100% | 87% | 64% | 50% | 40% | 32% |
| T5M | 100% | 100% | 96% | 75% | 60% | 49% | 100% | 100% | 80% | 62% | 50% | 41% |
| T6M | 100% | 100% | 100% | 91% | 73% | 60% | 100% | 100% | 98% | 76% | 60% | 50% |
Also take into account that party with more troops has higher chance to become attacker and land first hits and all HA counts as cavalry in autocalc giving advantage to armies relying on cav+ha like Khuzaits and sometimes empire.
Last edited:
