Defensive Order handled by poor AI, Option needed for player control

[Neomega]

AI didn't really pick a choice other than running a max function and picking the hardest hitting unit it has on its stack. And this type of combat has already existed in Civ 3. Basically the combat system picks the strongest defender for the attacker.
There is no good alternative to make a tactical choice on who's defending aganist what unit.
A warrior probably couldn't even land a blow on a knight. the knight has a 10:1 odds of killling the warrior, and I would be surprised if the warrior would even take down one strength of that knight. Not allowing you to choose the defender would definitely speed up a MP game where both plaers would try very hard to outsmart each other in choosing which units to attack. I haven't seen how useful is the 2 first strikes in combat to comment whether keeping the samurai would be a good decision.

There is a combat breakdown around here somewhere.. I can't find it right now, but basically... it says first strike is VERY powerful.... especially since it saps the enemy's strength before they start fighting. 2 landed samauri first strikes could make a mace man useless, and leave the samuari almost unscratched, without a too generous RNG roll.

 

[calyth]

I understand the combat system, I am just mad the AI decided to ignore the fact my Samauri were needed for the melee units, and chose to pit them against knights.

Suppose you're given the advantage of choosing which units are up to defend your city. How would it be fair to the AI city when you attack them, and it has no reasonable way to tactically choose, as well as you could've, which units should attack which?
The programmer's choice to pick the strongest defender is a heuristic to minimize loss for the defender. Clearly it has done pretty well because it has been pissing a lot of the warmongers off already. In this case, it didn't work out because the AI happen to bring a unit that nullify your UU's advantage. This is partly because there isn't a good way to factor in the 2 1st strike advantage into choosing which is the best defender against the attacking units. Frankly I haven't seen what would happen to a pikeman if you use a samurai with 2 1st strikes so I don't have a clue to fix this problem (if it is a problem in the first place).
Would a human do the same thing if they're planning to attack the Japanese and knows that they have samurais? Probably, so either the AI is smart, or more likely, just lucky. Alternatively, would there be a gripe about it if it was a Monguls sending Keshiks? or how about war elephants, which are prone to 1st strikes?
The other alternative that I could think of is to run minimax games with the attacking and defending stacks, come up with the probabilities of which units should face which, and then just pick them in the battle. I don't think this would yield better result, and result in more computation time which this game (being so CPU intensive already), doesn't particularly need.

EDIT: going to crack out the world builder to see how the Samurai does its 2 1st strikes.

 

[Neomega]

The other alternative that I could think of is to run minimax games with the attacking and defending stacks, come up with the probabilities of which units should face which, and then just pick them in the battle. I don't think this would yield better result, and result in more computation time which this game (being so CPU intensive already), doesn't particularly need.

The CPU intensity has little to do with the AI choosing how to fight... this is pretty straightforward at the moment. The Ai probably would have a tough time trying to pick it's collateral damage, if the occasion were to arrive, and could possibly be exploited by a player knowing the AI would accept collateral damage. (killing the collateral, then withdraw laughing)

However, an AI could easily flag units as defenders, (as it did in civ III), and of done randomly, a human would not be able to figure it out.

It's not chess, with various movements, at least not for a city attack, allthough open field combat could get quite hairy, trying to decide which units will descend from which direction.

For example, a stack of 3 types of units attacking another stack of 3 types of units in a city would only need to perform 9 calculations, then pick the right strategy.

 

[Psyringe]

Let me ask you a personal questioon... do you use the city governors? Do you automate workers? I sure don't. I like micromanagement. People complain about it like smokers who hate the way cigarettes smell, and always talk about how bad they make them feel.

I use the governors, and I automate some workers. Usually, in the beginning of the game, I micromanage. But later on, this just becomes tedious. When i feel that the micro-managing detracts from the fun, I automate my workers and use governors.

I want control. This is a TBS... emaning you have all the time you want to make the tough decisions... do you sacrifice your longbowmen, to allow your samauri the first strike hand to hand combat they are itching to fight?

I want fun ... control isn't always fun for me. But I can see where you're coming from.

 

[Neomega]

EDIT: going to crack out the world builder to see how the Samurai does its 2 1st strikes.

Don't forget, samauri always have a 10% combat bonus.

 

[oldStatesman]

Very good points Neomega.

Perhaps the advantage vs the aI criticism could be answered by making this an option .... not available in easier levels where it is not realy needed, but available by player choice in the higher ones...for those that want to use it, and will not be bothered if in fact it does increase mm.

 

[calyth]

Curious - I seem to find a bug.
It seems like the programmers took a short cut. I tried putting a screenshot into the gallery, and it's pending.
So I took the challenge. Started a duel size map, with Noble. i'm the Japanese, and I'm fighting the Indians. I add enough culture to the city so that it's 40% city defense, and added a unit and wait until it has 5% fortify.
I didn't notice any problem - besides the fact that the samurai didn't survive and I don't have a good way of resetting the PRNG's seed to test alternative battles of the same situation.
Then I read OP's post about how Samurai starts with 10% strength upgrade. Then I looked at the numbers. It says combat odds are 8.8 vs. 11.6.
I thought this couldn't be a difficulty setting problem because I started it as Noble. Then I figured that there might be a short cut made. Both units have 50% melee. And both 8.8 and 11.6 are figures in which the 50% melee bonus is not factored in. The correct values should be 12.8 and 15.6 respectively.
Then I took my calculator, and figure the ratio. apparently 8.8/(8.8+11.6) != 12.8/(12.8+15.6). The correct odds should give the attacker 2% advantage. It is a small advantage, but I think I've seen similar miscalculation before.
Anyone can tell me if I could change the seed through world builder? So far, I can't really see any evidence that the Samurai can save himself by using the 2 1st strikes, but because there's no seed change, there's no way to see what would happen if the PRNG rolled differently.

Stacks of 3 vs 3 is very little on the CPU, but mind you that the tableau needs to be recalculated for every battle, because odds changes, and units die. It means very little for a normal cpu, but I don't think that a) it would solve the problem of nullifying the samurai's 2 first strikes and b) the game is bad enough as is with eating up the CPU, and I'm not sure if I want them to do this until they fix up make the game more efficient first. There needs to be a way to define when is it advantageous to use a weaker unit in place of a stronger unit with first strikes if it is facing a knight. I don't favour the player control idea simply because we can't make sure that the AI would perform as well tactically.