Combat Explained....

[Bushface]

As I said, both elephants had Combat 2. And if the "extra combat" figures, which differed between the two fights, were applied to the attacker with their signs as I showed, then the second elephant (with a nett -10% extra combat) should have been weaker. Why in any case should the program have produced two separate "extras", one positive and one negative, if both are to be applied to the same combatant ?
Since the above occurrence, I have been accumulating data on relative combat strengths, winning odds, and amount of damage, and can already see that the values in-game with patch 2.08 do not totally match those in the estimable Arathorn's tables. In particular, a hit from a relatively weak defender always does more damage than his tables indicate - that is if the defender gets lucky and does actually make a hit. Note that here I am assuming that damage done by a weak defender against a stronger attacker is the same as would be inflicted if the attacker were weak and the defender strong.
I have also had instances when the combat log showed "extra combat" but the strengths of the combatants (shown on the main screen and in the combat log) took no account of such an extra. They followed the figures for basic strength, remaining hit points, and bonuses that appeared on the main screen, repeated in the combat log.

 

[DaviddesJ]

As I said, both elephants had Combat 2.

I know you think that both units had Combat 2. But I think you're mistaken.

And if the "extra combat" figures, which differed between the two fights, were applied to the attacker with their signs as I showed, then the second elephant (with a nett -10% extra combat) should have been weaker.

As the previous poster explained, the attacker's Combat modifiers show up as "-X%" in the logs, but they actually count +X% for the attacking units.

 

[malekithe]

...I have been accumulating data on relative combat strengths, winning odds, and amount of damage, and can already see that the values in-game with patch 2.08 do not totally match those in the estimable Arathorn's tables. In particular, a hit from a relatively weak defender always does more damage than his tables indicate...

Note that damage dealt during combat is now based upon the combatants' power with maximum health, not their current health. So, a wounded longbow defending against a full-health axeman is going to do the same amount of damage per "hit" as a full-health longbow would.

 

[Bushface]

O.K. The preceding post means that Arathorn's tables are now out of date, which is a pity but does add meaning to the eventual interpretation of the data I am collecting. However, his tables refer to relative strengths at the time of combat, not to whether these were affected by previous damage. Damage now being calculated from full strength removes a possible variable from the calculations, though it remains obviously easier to kill an opponent if he begins with less than 100 hit points - and easier for him to do the killing if your unit is the one with reduced hp.

 

[Roland Johansen]

As far as I know, there is only 1 tiny change in the game compared to the information stored in the article from Arathorn.

The chance to hit is still calculated from the strength values that are visible when you ALT-mouseover an enemy unit. The damage per hit is now however calculated by averaging the maximum strength value of a unit and the damaged strength value. 1 tiny change (which can have significant consequences on combat) doesn't devalue the article. It just needs this one line addendum, that's all.

 

[malekithe]

The damage per hit is now however calculated by averaging the maximum strength value of a unit and the damaged strength value.

That is correct. The information in my previous post was out of date (patch 1.52).

 

[Bushface]

This does not agree with the values I have been getting. For example, three fights in which my Infantry have attacked Catapults :-
1). Full strength Inf. - mouseover + ALT showed modified strengths of 22 and 5.5, at odds of >99.9% in my favour (all such values have agreed with those in the Combat Log except that the latter shows 100% not >99.9%) - so R = 4: damage per hit was 37 by the inf. and 10 by the catapult.
2). Inf. 52HP, str.11.44 vs 5 for cat., R = 2.288, damage 34 and 11 respectively.
3). Inf. 40HP, str. 8.8 vs 1.2 for cat., R = 7.33, damage 39 and 10.
Now these figures for damage do generally agree with Arathorn's table (though damage of 33 he said needed R = 3, not 2.88), but do not appear to show any relation between reduced strength and less damage caused. Indeed, the most-damaged infantry did more damage per hit than did the full-strength one, though to a weaker opponent.
In these three examples, the mysterious "extra combat" did not appear. I still fail to see when these bonuses are applied, and why they appear only after combat has been initiated. A 20% change to the strength of either combatant could greatly affect the odds of winning, even reversing the likelihood of victory if the originally-displayed strengths are very similar; for example, you see 10 vs 9 on mouseover, with about 62% chance of winning, attack, then (and only then) your opponent picks up an extra 20% and the odds drop to only about a 35% chance. Of course, the extra could be in your favour - theoretically. I have yet to collect enough data, from similar combats with and without extra combat occurring, to offer any conclusions, but no doubt someone familiar with Python and XML coding can enlighten me. Or is the information in the core program ?

 

[malekithe]

1). Full strength Inf. - mouseover + ALT showed modified strengths of 22 and 5.5, at odds of >99.9% in my favour (all such values have agreed with those in the Combat Log except that the latter shows 100% not >99.9%) - so R = 4: damage per hit was 37 by the inf. and 10 by the catapult.
2). Inf. 52HP, str.11.44 vs 5 for cat., R = 2.288, damage 34 and 11 respectively.
3). Inf. 40HP, str. 8.8 vs 1.2 for cat., R = 7.33, damage 39 and 10.
Now these figures for damage do generally agree with Arathorn's table (though damage of 33 he said needed R = 3, not 2.88), but do not appear to show any relation between reduced strength and less damage caused. Indeed, the most-damaged infantry did more damage per hit than did the full-strength one, though to a weaker opponent.

I fail to see the problem. When I run the numbers for the scenarios you give, I get the exact numbers you say the game gave you. Maybe you're not computing the damage properly. You can't use the same R-value for damage, you must recalculate it using the average of maximum health and current health for both the defender and the attacker. So, in your scenarios above, I get the following values for the "damage" R:

1) 4
2) 3.344
3) 4.968

Also, I have always found that "extra combat" in the logs corresponds directly to combat promotions. It will be positive for combat promotions to the defender, negative for combat promotions to the attacker.

 

[Bushface]

Point taken (about working out R-value). I didn't "compute" the damage, just took the values from the combat log. Perhaps the statement about relative damage by a full-strength attacker compared with that by a weakened attacker should be amended to say "against defenders of equal strength". Anyway, I now know the truth.
Incidentally, in the second example the Infantry, despite its 97% win chance, lost to a series of 5 hits from the catapult. The chance of such a series I reckon to be 0.0000000243% (0.03^5). Yet more curses on the streaky RNG !
But still. the exra combat is a value unknown to the player until the combat log is seen.

 

[Bushface]

More thoughts. My apologies, but there are still things my poor old brain doesn't quite understand.
In my second example, the average of full to damaged strengths I get as 15,72 and 5, the ratio of which is 3.144 not 3.344. Whether this is sufficient to get past a damage step point remains to be seen. And for example 3 I get average values of 14.4 and 3.1, ratio 4.65 not 4.968. A further point: if either combatant shows a strength greater than its basic one (e.g. full strength + Combat 1), does it fight at the higher strength or is it again taken as the average of basic and modified values ? Averaging would then reduce the benefit of such promotions.

 

[malekithe]

But still. the exra combat is a value unknown to the player until the combat log is seen.

Not true. For every combat promotion had by the attacker there will be a negative 10% extra combat modifier. For every combat promotion had by the defender there will be a positive 10% extra combat modifier. It's actually not a very difficult concept to understand. Maybe you're confused because these modifiers have already been applied to the numbers you see in the log?

In my second example, the average of full to damaged strengths I get as 15,72 and 5, the ratio of which is 3.144 not 3.344.

You're averaging the wrong numbers, apparently. avg(22, 11.44) = 16.72 (not 15.72)

And for example 3 I get average values of 14.4 and 3.1, ratio 4.65 not 4.968.

Again, avg(22, 8.8) = 15.4 (not 14.4)

You don't seem to be taking the infantry's combat promotion into account in your calculations. This would explain your last question...

A further point: if either combatant shows a strength greater than its basic one (e.g. full strength + Combat 1), does it fight at the higher strength or is it again taken as the average of basic and modified values ? Averaging would then reduce the benefit of such promotions.

A combatant's combat value is calculated while including all applicable combat modifiers. So, an infantry with a single combat promotion would have, when at maximum health, a combat value of 22. When at lower health, it's combat value for damage purposes would be the average of 22 and the health-modified combat value (HP/100 * 22).

 

[Bushface]

More thanks, Malekithe. However . . .
Had a combat where both sides had Combat 2. According to what you said, this should bring extras of +20% and -20%. But only +20% appeared, in the log, which also showed both sides as having strength 18, which is correct for undamaged cavalry with combat 2 each. Hence the extra +20% had not been applied to the logged strength of the defender. This could perhaps have been the reason for the defender AI to win with 5 consecutive hits to none, the chance of that occurring at the quoted 50% odds being 3.125% whilst with another 20% strength addition there would be different odds, favouring the defender.
Are these extra combat % applied to the defender's modified strength or to the averaged value ? And when ? Definitely not visible before the fight.
I take your further point about what values should be averaged. Silly me. And again thanks for your patience.

Edit: Got it at last ! There is no magical "extra combat, just the strange way the combat log reports the additions from promotions. End of discussion, I think.

 

[_alphaBeta_]

I've been asking some questions about aircraft interception over on the quick answers thread, but wanted to confirm a few things here. The big question I'm left with is:

If a plane attacks, and there are multiple units that have the ability to intercept, do they each get a chance to intercept it? I'm just curious if having tons of interception units around (each with a percentage chance to intercept) pretty much guarantees an interception.

Let's take SAMs who have a 40% chance to intercept. As an example - if I have three SAMs, and an enemy fighter comes into range, what exactly happens? I'm guessing SAM #1 will get a 40% chance to intercept it. If it fails, SAM #2 will get a 40% chance to intercept. If that fails, finally SAM #3 will get a 40% chance to intercept. If that fails, the attack is successful. This would apply to all interception units like other planes, destroyers etc. If this is true, then having a boat load of interception units would be extremely effective. However, if the first eligible intercepter fails, and no other units get a try, that changes things quite drastically.

I'm being told over there that only the best intercepter gets a chance and that's it. Can anyone over here confirm?

 

[VoiceOfUnreason]

I'm being told over there that only the best intercepter gets a chance and that's it. Can anyone over here confirm?

Depends on what you think that means.

On an Air Bomb/Strike, the intercept test first checks to see if the plane evaded interception. If not, the game figures out which is the bestInterceptor, and that unit gets a chance to damage the plane.

Avallable interceptors are the units defending that tile which have not previously intercepted an aircraft. bestInterceptor is the available interceptor with the highest intercept probability.

See CvUnit::bestInterceptor

 

[_alphaBeta_]

Depends on what you think that means.

On an Air Bomb/Strike, the intercept test first checks to see if the plane evaded interception. If not, the game figures out which is the bestInterceptor, and that unit gets a chance to damage the plane.

Avallable interceptors are the units defending that tile which have not previously intercepted an aircraft. bestInterceptor is the available interceptor with the highest intercept probability.

See CvUnit::bestInterceptor

I'm not exactly following you. You're saying that if I'm attacking a plot that only has one possible defender, once that defender intercepts once, any subsequent attack by additional aircraft will get through 100% of the time for that turn?

If that be the case, then it would be beneficial to have many defending units around since bad things can happen if you run out of available interceptors.

 

[VoiceOfUnreason]

I'm not exactly following you. You're saying that if I'm attacking a plot that only has one possible defender, once that defender intercepts once, any subsequent attack by additional aircraft will get through 100% of the time for that turn?

If that be the case, then it would be beneficial to have many defending units around since bad things can happen if you run out of available interceptors.

You are following me exactly, and have drawn the appropriate conclusion.

 

[Roland Johansen]

There is still some confusion in the 'Quick answers and newbie thread' about interception. Some people have different experiences than others.

Is the following scenario possible?

Plane A attacks a city and is intercepted by plane B, but plane B is again intercepted by another plane C. Can such a double interception occur?

I would prefer an answer by someone who really knows (by looking in the SDK or from some ingame testing).

Thanks in advance.

 

[sixs_monkey]

Where would the "fighter escort" (plane C) have come from? Civ air units don't move in groups, armies, or stacks currently (to the best of my knowledge, though it is rusty as the Fall from Heaven mod isn't yet implementing Civ air combat units), so plane A (a bomber, or at least on a bomb run) would be hoping for a friendly fighter to be launched from his side to keep him safe.

At this time, I'm gonna say its a pretty safe bet that Firaxis didn't choose to have all nearby flying units check for 'dogfighting', though that is something that I would like to see someday, as long as it was kept local to each bomb run and each plane had a finite number (fuzzy perhaps) of sorties it could participate in per turn.


Actually, I was posting here in hopes that the founders of the thread might be revisiting the subject once more, in case details (minor or major) changed with 1.74 and BtS. In any case, thanks for the illuminating work!

 

[Benford's Law]

Merry X-Mas all

I don't mean to be a thread necro but I was doing a little research on combat and ran into this great (if a little dated) strategy article.

I think I understand how the defender's strength is calculated, but is this passage from the article correct?

Example: A sword (str:6) with City Raider I, II, and III is attacking a pike (str:6) defending a city. The city has a 30% defensive bonus. The attacker’s strength is just 6, but the defender’s strength is 6/(1+1.1-.3) = 3.33

First of all, just to be clear...do the CR promotions stack? Even if they did wouldn't the swordsman only have a 55% bonus (ie. 10% for swordsman's natural city attack bonus + 20% for CR1 + 25% for CR2 + 30% for CR3 - 30% city defense)? In this case, the defender's strength would actually be 6/(1+0.55) = 3.87, wouldn't it?
If CR promotions *don't* stack, then the swordsman would only have a pitiful 10% bonus, giving the pike a strength of 6/(1+0.10) = 5.45

Can anyone help clarify this problem for me as a nice X-Mas gift? Thanks!

~Benford's Law

 

[oyzar]

CR promos stack, seems like there was an extra shock added in for the attacker(don't ask me how he got it without any combat promos).

That said i am pretty sure at least some of the info in this article is dated since combat now(in bts 3.17 at least) uses firepower instead of unit strength when calculating damage/chance to win. Firepower = (strength at max hp + current strength(which is strength at max hp multipled by hp) )/2 .

Ask niklas / DanF5771 if you want to know exactly how combat works now though...