Automatic battles and other things...

[drus45]

Last time (don't bother to look up) I wrote about inconsistency in unit costs. It's not only that.

I have searched and seen very little talk about automatic battles. The thing is that they are not balanced at all. I am one of those who don't want to command every battle personally, mainly because there is no challenge in the campaign if I do so. And it gets so boring, I want to get ahead with the campaign and do most stuff in the strategy map.

I have made myself some rules like; I can only command battle personally with faction leader (you are one man, not god), and to make the battles more challenging put troop limits. And try to get the AI to mass their troops for one big battle rather than many small ones.
-Seleucid troop limit for example: 1 general + 6 phalanx units + 4 whatever infantry + 2 scirmishers/archers + 2 cavalry + 3 siege engines
-Roman Troop limit: 1 general + 7 whatever infantry + 2 scirmishers/archers + 2 cavalry + 3 siege engines

Still I win too easily. I play on hard/hard, to give the AI 5000 gold every turn (keeps the poor factions in the game) and +4 attack +4 defence? +4 morale on the battle map. I will not use "very hard" because in strategy map very hard is totally mad, you guys must know. And in battle they get +7 morale on very hard, so just forget very hard.

Anyway the automatic battles are a disaster. For what I have learned is that automatic battles take into consideration when deciding a winner:
-troop numbers
-(melee?) attack
-hit points

There is also a bit of luck. Generals stars don't really have that much effect, only when u have some 7+ stars it has somewhat average effect.

At the moment if I put my Seleucid Silver Shields against some much cheaper 240 unit phalanx they will lose. Also hypaspists, sacred band, legionaries, gallic chosen swordsmen etc will lose against cheap numbers. So there is really no point in making those elite units. Also in battle map the Silver Shield for example is only slightly better than normal phalanx and it costs two times more!!!

BUT there is one more way to mod the units automatic battle ability. There is a second hp in export_descr_unit, it is the mounts hp, with cavalry it don't matter, but with infantry it has big effect on the automatic battle outcome. For example:

-when you have a seleucid phalanx with 14 attack, 1hp and 5 second hp and 240 men, you could calculate it's automatic battle power by 14(attack) x 240(men) = 3360 (this is the automatic battle power)
-then you have ptolemy legionary with 10 (melee)attack, 1 hp and 5 second hp and 160 men: 10(attack) x 160(men) = 1600 (battle power)

If you would like that the phalanx and legionary are about as strong then you could double the legionarys second hp to get this:
10 x 160 x 2(the doubled second hp; phalanx has 5 and legionary has 10) = 3200 battle power

Something like that. I hope someone will understand what I mean and I will propably edit my message later and comment more to make more sense.

Also one funny note, I remember seeing someone talk about... It was that the mercenaries should not be so good in maintaining public order. I have heard that if you put something like peasant_unit or whatever in their "attributes"? in export_descr_unit it will reduce their public order effect by 50%. I really need to check this one out and tell more later. If someone knows something about this then good.

And these issues are not only Extended Realism issues, but have always been with the game.

[Quinn Inuit]

Hello again. I think the tag is is_peasant or some such.

I'm pretty sure the secondary HP works on all units. See here: http://forums.totalwar.org/vb/showthread.php?t=88859

I think one of the reasons you're seeing better units do poorly in autoresolve is that the AI gets a bonus on autoresolve based on the campaign map difficulty.

[Scutarii]

Well, autoresolve is heavily biased against the player on H and VH campaign difficulties, so you can't really base much on it.

The tag is in fact 'is_peasant', it reduces a unit's effect on Public Order by half. Its used by peasants in BI.

Secondary HP definitely works with cavalry. Check out the Scythians in XGM, they're units are given a secondary HP bonus and suddenly HA's wreck other comparable units in autoresolve.

And general stars have lots of effect on autoresolve, more than the actual battles themselves anyhow... Granted, you won't see your general's pulling off amazing victories if they aren't 8-10 stars. Just steadily better performance.

[Quinn Inuit]

We give our HA a secondary HP bonus, too, but I'm not sure it's enough. In the most recent version I think I upped it by one, though.

[drus45]

What you don't seem to understand is that in Extended Realism Mod the automatic battles are totally awful. So this is VERY IMPORTANT stuff that must be taken into account immeadedly.

Let me tell you something about my testings...

"merc cretan archer" is about as strong as "aor gaul roman spearmen" for example. I tested with two full stacks attacking each other and most of the time Cretan archers won slightly!!! That is outrageous! Archers are far inferior to any spearmen. Cretan archers have melee attack: 7, armor: 0, defnce skill: 7. gaul romans have attack: 9, armor: 7, defence skill: 19 + also a shield.

This happens because both have about the same "auto battle power". Cretans have 1960, while the Gaul roman spearmen have 1800. It gets more outrageous since the hypaspists lose to both the cretan archers and gaul roman spearmen. The Hypaspists have only auto battle power of 1400! Hypaspists lose to archers and cost like hell. That is crazy. Silver shields also have lower than gold shields. Triarii has auto battle power of 1100!

What is really important is that the ONLY things (+command stars etc) the automatic battle calculations care is:
1. Troop numbers
2. Primary melee attack
3. HP
4. Second hp.

And the second hp thing DO NOT WORK with cavalry, except elephants. You can test it yourself. The way to calculate this when the unit has 1 hp is to (forget what I said in the first post): troop numbers x primary melee attack x second hp. With cretan archers this is 40(this is in the export_descr_unit) x 7 x 7 = 1960.

So I ask you to check this thing because in this mod and every other mod too the automatic battles are horrible.Also... the AI thinks cretan archers are inferior because it backs up from a fight if I take aor gaul roman spearmen and attack. But in reality they are not. So when you start modding with this new knowledge it will chance the way the ai factions are going to succeed. It can change the whole balance upside down.

And I don't want to be rude or anything. This is the best mod with best graphics, historically accurate and all. But there are some really weird going on with the auto battles. Also like I have said there is something wrong with the unit cost also.

I will write more later.

[Quinn Inuit]

Two things:
1) Are you still testing this on Hard campaign difficulty? If so, your results are tainted.
2) Where are you getting your numbers for how autoresolve works? I think you're leaving out a few things. I'm in favor of systematizing, but one must have a solid system first.

[drus45]

Still I don't want to be rude, but why don't you test it out yourself? I mean, people have done all these great graphics, auxilia buildings, got rid of the senate, put some extra factions etc but have never payed attention to the auto-battles? Have you not noticed it yourself when you play? Hypaspists suck, they are the ultimate losers in this game when in reality they were the opposite. Even in battle map mode they are almost worthless with pathetic 80 men. If archers or peltasts target your "elite" hypaspist then you can say goodbye to your over expensive pussies.

And I have tested all these things with medium/medium difficulty, first with the other side and then the other like 10+ times each. I should not have said anything about me using hard/hard, because thats only when I play for fun and not testing.

But one interesting thing happened; I tested full stack of falxmen (5000 auto battle power) vs full stack of gold shields (4200 auto battle power) and the goldshields won slightly every time. So there seems to be some other factors after all, but with other troops my system seems to be working. The most powerful non-cavalry unit in the game with autoresolve seems to be the macedonian/seleucid/ptolemy phalanx.

But even if my system is not 100% accurate the elite units are still underpowered in auto-battles and in some cases in the battle map also.

Also there is problem with hoplites whose formation is:
formation 0.7, 1, 2, 2,...
They are too close together and even when highly_traines their formation is disorderly.

Also the german berserkers have javelin instead of sword in their hand in battle map mode. This is because their javelin weapon have been deleted from export_descr_unit, but their skeleton still has it. I don't know how you can take the javelin away without them using it in battle map mode.

There are all kinds of problems but I encourage you to test these things yourself. And when you have done that you can tell me if the results were the same and if you believe me or not. Assuming the results are the same, this means that some major work needs to be done with this mod (and other mods too).

And the way to calculate the "auto-battle power" is to use multipiclation or whatever it is called:
troop number x (primary) melee attack (with phalanx units this is the spear) x second hit points (hit_point_extra)
with macedonian phalanx this is 60 x 14 x 5 = 4200

[Caligula Caesar]

Still I don't want to be rude, but why don't you test it out yourself? I mean, people have done all these great graphics, auxilia buildings, got rid of the senate, put some extra factions etc but have never payed attention to the auto-battles? Have you not noticed it yourself when you play? Hypaspists suck, they are the ultimate losers in this game when in reality they were the opposite. Even in battle map mode they are almost worthless with pathetic 80 men. If archers or peltasts target your "elite" hypaspist then you can say goodbye to your over expensive pussies.

I do have to agree that a lot of the elites are underpowered. However, I would say that you don't need to get annoyed about our replies. Quinn isn't being unreasonable.

And I have tested all these things with medium/medium difficulty, first with the other side and then the other like 10+ times each. I should not have said anything about me using hard/hard, because thats only when I play for fun and not testing.

You can say that without sounding very irritable. If you say "I have tested repeatedly on m/m, and the results don't vary greatly", we'll believe you. By the way, I have noticed that on autoresolve, I seem to get one result the first time, then a different (in terms of numbers lost) result if I reload, then the same result again and again if I keep reloading. There is also a ctd if I reload and win with naval battles, btw.

But one interesting thing happened; I tested full stack of falxmen (5000 auto battle power) vs full stack of gold shields (4200 auto battle power) and the goldshields won slightly every time. So there seems to be some other factors after all, but with other troops my system seems to be working. The most powerful non-cavalry unit in the game with autoresolve seems to be the macedonian/seleucid/ptolemy phalanx.

Try testing non full stacks with various units and see if your rule still holds, if you want to make yourself useful.

There are all kinds of problems but I encourage you to test these things yourself. And when you have done that you can tell me if the results were the same and if you believe me or not. Assuming the results are the same, this means that some major work needs to be done with this mod (and other mods too).

And the way to calculate the "auto-battle power" is to use multipiclation or whatever it is called:
troop number x (primary) melee attack (with phalanx units this is the spear) x second hit points (hit_point_extra)
with macedonian phalanx this is 60 x 14 x 5 = 4200

So to make the Macedonian phalanx not so strong in autoresolve, all we need to do is add secondary hp? That should be easy enough. If you want to make yourself useful, try it and see if it works.

[Delvecchio1975]

does anyone know if we can put decimals in hitpoints? aqd put all HP on 3 and then the second HP 3 for infantry, 4 for peltasts, 5 for archers, and so on, to stabilize the autoresolve. it works really well (tried it in ExRM and RTRPE), but it sorta screws up your self-fought battles who last around three times as long now. decimals in the hitpoints could easily solve this.

i should test this myself, shouldn't i?

[drus45]

Ok first I must say that I have done more testing and there are other factors as well. Very important factor is the armor and other important seems to be ranged weapon.

There are many units that beat makedonian phalanx, some with CRUSHING victory. These include: roman velite, aor numidia spearmen, numidia javelinmen, aor asia roman archer.

The numidian javelinmen and Berber Javelinmen seem to be some of the most powerful units in auto-battles. They beat macedonian phalanx with crushing victory, and hypaspists get their ass kicked as well by them (hypaspists lose to basically everyone). Funny thing is that these units cost very little (maintaining 1 turn is 168 and 102 for these two. Hypaspist costs 900 and macedonian phalanx costs 350 per turn).

The "aor asia roman archers" have 8 armor and that makes them very powerful combined with their ranged weapon. They beat the macedonian phalanx with crushing victory etc... might even be THE most powerful unit. Again they don't cost much.

Anyway some things what I suggest is:
-hypaspist and triarii strength should be doubled to 160 men
-sacred band at least to 160 men
-silver shields and other agema to need boosting too
overpowered auto-battle units hit_point_extra must be greatly reduced:
-velite,bactrian peltast and italy skirmisher to 4
-egyptian peltast to 5
-macedonian peltast to 5
-aor numidian spearmen to 1 (any other way they are extremely powerful)
-numidian javelinmen to 2
-aor numidia skirmisher to 5
-aor asia roman archer to 3
.... basically every unit needs to be tested and hp_extra replaced

I have been testing much, but I can't figure out a way to calculate the armor and ranged weapon auto-battle power. They have (especially armor) REALLY big effect.

One thing what I have done when playing the game myself (this don't need to be changed but I think they made the game better):
-all infantry units maintaining cost -1/3

I can't come up with anyway to change the auto-battle power of the cavalry. They are extremely underpowered.

And if I sound irritable then I propably am. But don't care about that. Just listen what I say, because this is very important stuff.