Printable View
EDIT: SCRATCH THAT. I've just retested the phalanx charge bonus, and it is not doing anything. Wierd that it did before, arrrgggg. Believe it or not, I tested it more than once and was seeing an impact--must have been unlucky with the random parts of combat. Time to correct my other post on this.
You know I didn't even consider the fact that certain units would be worse at defending their shieded side than their swordside. Units that comes to mind are Urbans, Arcani, Bastarnae, various cavalries (Light Lancers in particular), all pikemen (yes even Levies), Spartans and Armoured Hoplites. That can't be true??? ~:confused:
I refuse to believe that unshielded units are totally screwed from the left and that super elites are in fact easier to attack on their shielded side. That doesn't make sense.
And what about elephants? They have a very high defensive ability and not so high armour, does this mean we should take them on on the left and stay clear of the right?
Yes, the info Jerome posted was very useful, and quite a bit different than I expected in some regards. I hope he has a chance to give us some clarification on the defensive skill/shield thing. It would give us a much better idea of how to properly rate/evaluate units.
I suspect there is more going on, however. One reason is that shield values cap at 31 in unit stats (0-31 range), while defensive skill can reach 63 (0-63.) That might lend itself toward something like the unshielded side getting 1/2 the defensive skill stat added to it on attacks to that side (just a guess.) Direct frontal might get both or the higher of either stat or something else...gives me some idea for testing.
A good way to test it take a unit of Companions (good charge) and mod a unit with no shield to have 63 defensive ability (and preferably a low mass). Then present the Companions with the left side in a very deep formation so that they all hit the side. If the Companions penetrate deep or even just kills a fairly largen number of men, then Jerome is right and we need to revaluate every single unit, if it only kills a few men and is killed itself then we know that defensive ability is applied to both sides (perhaps not equally but at least the left is not screwed over).
Ok I tested it myself...
Modded the Cilicians to have 63 in defensive ability. They are most likely the best since they have low mass, no armour and no shield.
I put them into a deep square formation about two times deeper than wide and presented teh side to the Companions. The charge went very badly for the Companions, they killed only 3 men, but then again I had noticed that the entire line of men on the left flank had turned to face them. So when they pulled back to retry the charge I made my Cilicians (who had lost 8 men to the 27 Companions) march perpendicular to the Companions to force them to present the left flank. This time the Companions killed 56 men in the charge, and my unit routed instantly there after.
Now I have heard that men marching won't recieve a charge as well as those standing still, so that might have impacted the result.
I tried it again and the results were the same, if the Cilicians faced the Companions they were pushed back and the Companions penetrated the unit but killed less than a hand of men. Marching made my Cilicians die like flies.
Obviously the defensive ability doesn't add much to the left. Now I will try marching with the right side presented.
[EDIT]
Hmmm... The test was inconclusive. The Companions could indeed penetrate the right side as well, but were far from as good at it as the left. But if my men were marching they got killed very badly, in one case I lost every single man and the Companions lost 1 single rider. Standing still the Cilicians were a veritable wall, but against the men on the right flank turned to face the enemy. So the results are impossible to understand, only that you should never march with your side to enemy cavalry, your men can't stand up to a charge at all.
I didn't have much luck testing it either. As you say, the men turn to face the threat, and that wrecks the test. I tried 2vs1 unit tests with phalanx and other types, but I would need to seriously reduce the kill rate and increase morale for it to work as a test. The unit routs too easily, or turns to face the threat too easily.
From which we can conclude, presumably, that if there is a difference, it's not important in the scheme of things.
-Simetrical
That is my hope. The problem is that things can be having an impact, but not be easily verified until some definitive test is found (pri/sec bug being an excellent example.)Quote:
Originally Posted by Simetrical
Three questions:
1.Why there's no difference between a Companion with 1 charge bonus and a Companion with 60 charge bonus? They did exactly the same damage on average when charging after many testings.
2.If the spear attritube really did use the cavalry's charge bonus against themself, then why a group of Round Shield Cavalry with 52 charge bonus charged head on a hoplite phalanx suffered the same casualities as Round Shield Cavalry with 2 charge bonus charged head on?
3.Does the Charge Bonus exist and APPLY IN COMBAT or not?
AI:
1. Is it possible to mod the game so the AI keeps their troops more together, and attack as a group?
2. Can i change some of the priorities in descr_formations_ai.txt in order to achieve this? If so, how?
3. What does 'linked' mean?
4. descr_formations_ai.txt and descr_strat.txt allow us to mod the AI to some extent. Are there other *.txt files (perhaps more hidden) that allow similar moddability?
1. How much attack penalty do spears have against foot units?
2. Do swords have attack penalty against mounted units?
1. Explain how the battle difficulty is bugged.
They fight at +4 against mounted, and probably -4 against infantry. A penalty anyway, see export_descr_units.txtQuote:
1. How much attack penalty do spears have against foot units?
No penalty is mentioned in the txt file. Probably none.Quote:
2. Do swords have attack penalty against mounted units?
AFAIK, the +4 bonus to attack for the AI on Hard, the +7 bonus for the AI on Very Hard, and the +4 bonus for the player on Easy, are given to the other party as well.Quote:
1. Explain how the battle difficulty is bugged.
It effectively means that battles last longest on Normal, with the combination of lowest atack/highest morale (for both parties, though that would of course work as intened: The rules for morale are always the same).
A +4 bonus (146% combat effectiveness) against mounted units in RTW is almost insignificant. It's equivalent to +2 points in STW or MTW. STW had an anti-cav differential of 8 points (400% combat effectiveness) for spears vs cav. That would be 16 points in the RTW system. On top of that, cavalry in STW cost at least twice what a the spear unit which could defeat it cost, but that's not the case in RTW. Really STW is the only game in the entire Total War series where CA got the RPS right. It was weakened in MTW and the spears were too expensive.
Weak RPS in a game means that overpowered unbalanced armies have no counter army other than the same army.
Ah, it's optimistic Puzz3D again.. havent seen a 'i love RomeTW' post from you yet :)
Now since we dont know the engine formulas, how can you say that the effect in STW/MTW is more pronounced than in RTW? Also, they DID try to balance RomeTW, and it is pretty balanced stat wise, but game mechanics like cavalry charging proved to have a too large effect on things.
In my opinion, swords and spears are better balanced than in say, MTW. Also the cavalry in RTW behaves a lot more realistically than cavalry in STW, which was basically just treated like a somewhat faster moving infantry unit (ie, units locking together in hand to hand quite boringly, not happening very much after the charge except a drudgematch). Anyone playing STW and MTW can verify this. And you could not really withdraw your men from melee, but in Rome TW, you can.
We do know the basic formula. We know exactly what the formula is in STW and MTW, and I did enough tests to show it's basically the same formula in RTW.Quote:
Originally Posted by sunsmountain
Could someone from CA clarify on loyalty and emerging factions system on BI?
There is a LM thread open at https://forums.totalwar.org/vb/showthread.php?t=66032
Thanks.
Puzz3D (or anyone would would be willing to help), I have been trying to reconstruct parts of the RTW formula.... from scratch. Would you please tell me the MTW formula or point me to it and/or you best guess as to the RTW formula?Quote:
Originally Posted by Puzz3D
I am mostly interested in the exact mathematical effects of sword/shield/experience/terrain/orientation/exhaustion on chance to kill.
Any help you can give would be much appreciated.
Thank you!
EDIT: Acckkk!! Only one post? That means the forum "machine" lost the intro post, that I thought I had posted 2 weeks or so ago, in the welcome thread. Waahhh. Must have been the machine or the moderator getting even for all the whining I did in it about not being able to post!
Anyway; Hi, and mucho admiration to you all for the wonderful array of research material. I've been reading the very informative posts in this forum for months, and have been trying to post for most of that time. Nice to see one finally stuck to the wall instead of fluttering down to the gutter.
In MTW the formula is:
Each man has an attack value and defend value used in calculating the chance to kill his opponent on each "strike" in a combat cycle. Each man who is fighting gets one "strike" within a combat cycle, but can parry many times.
chance to kill = 1.9% * 1.2 ^ df
difference factor df = attack - defend + bonus
attack is the attack value of the striker
defend is the defend value of the opponent
bonus is any combat modifier that applies
charge is added in as a combat modifier
You can see all the combat modifiers in this thread Defensive Bonuses. Most of this comes from the Official MTW Strategy Guide, and some came directly from LongJohn who wrote the chapters on the game mechanics.
In RTW, the formula has become more complicated. We don't know the exact formula, and because it's more complicated, it's hard to determine empirically. Some tests I did indicate that the 1.2 constant has been changed to 1.1. That means the size of the steps in chance to kill has been reduced from 20% to 10%. The cutoff on df is + or - 20 in MTW. This has apparently been increased to 64 in RTW. In addition, there is a lethality factor for the melee weapon in RTW. I think this comes in as a multiplier on the base probability. Also, the shield now directly contributes to the defend value. but is directional as is the parry ability. From post 20 in this thread by Jerome: "Armour is applied as a modifier to attacks from any facing, while defense is a parry-style bonus on the front and right flank, and the shield bonus applies front and left flank. Missile weapons ignore defense." Missle weapons have a separate formula, and it is more straightforward. Tests in STW indicate that doubling armor cuts the kill rate in half from projectiles such as arrows. Height increases the range and accuracy of ranged weapons.
Thank you, Puzz3D !
Just knowing the form of the formula (pardon the play on words) helps a lot. You delivered exactly what I was hoping for.
So the ranged weapon equation (discounting terrain and weather effects) is similar but simpler?
Thanks, again.
preliminary removed
duplicate removed
In STW, accuracy determined if you hit the target, and power determined if you penetrated the armor. My measurments indicated pentration was a simple inverse proportion. Penetration = C * power/armor where C is just a constant to scale the value to a particular range of probability. If the projectile penetrated the armor a hit point was removed from the target. All men had one hitpoint except for general which had 6 in multiplayer and custom battle, but could have more than that in SP campaign.
In MTW, the system was changed. Accuracy still determined if you got a hit, but projectiles had lethality and power and an armor modifier. Penetration = C * lethality/(armor modifier * armor). The power determined how many hitpoints were removed from the target if the projectile penetrated the armor or hit a wall. This was instituted to handle artillery, but other weapons got higher power as well such as guns which got power = 4. This meant a gun could kill a general in MP or custom with 2 hits.
In RTW, I don't really know how it works. I see missles have an attack rating in the unit file and this changes when you upgrade the experience or weapon, and the projectiles file has damage and radius parameters for each weapon type. I don't know where the accuracy parameter is, but there must be one. I would expect it to work something like MTW, but there is no armor modifier that I can see. I suspect attack is equivalent to lethality in MTW, and damage is equivalent to power.
Just wondering what the default video settings for multiplayer are? They do not stay the same as what I have set in options. This is frustrating as it makes it difficult to know what I can comfortably host.
Why are some skeletons better in combat than other skeletons? (all other things being equal)
Probably a related question:
How is a unit's attack speed determined?
I believe it is some combination of animation speed and the min_delay parameter (second to last on the stat_pri) line, but which animation?
fs_dagger has 6 attack animations that are 36,38,28,28,36, and 36 frames long respectively. average=33.6667
fs_2handed has 6 attack animations also that are 24,22,20,20,22, and 18 frames long. average=21
Repeated test battles have shown that I need to set the fs_2handed min_delay to 13 to get the two skeletons approximately even. This seems to support the idea that min_delay is the time between the end of one animation and the beginning of the next in FRAMES, not as the EDU comments say...tenths of seconds.
However, when I do the same test between fs_spearman(average animation time =25 frames) and fs_2handed. I find that the fs_spearman skeleton is actually *faster* than the fs_2handed skeleton and I have to give it a small min_delay to get the two even. I expected to need to give the fs_2handed a 4 frame delay, but it turns out I need to give fs_spearman a 2 frame delay. This destroys the idea supported by the previous test and leaves no theory that I can think of remaining.
So, my point is that I can't figure out how to tell the affect of a given scale 1 skeleton on combat short of doing 100s of test battles with different min_delay values for *every* skeleton in the game. There is no formula that I can find that makes sense.
Please help...