Attack force and attack points

[bnj.s]

Hello all,

Are the attack/defense points the only criteria used to decide who wins a battle? I have the feeling that my Archers (attack 2, defense 1) get much better results than my Legionaries (attack 3, defense 3) when I am attacking a Pikeman. It seems very strange, though. Is it possible?

Thank you.

 

[MAS]

The defense of the defender is modified by factors such as the terrain type, being behind a river, in a fortress or a town with walls, being in a city or metropolis.

If you swordsman is attacking a pike fortified on a hill ofer a river inside a city. Then he needs to fight a defense of.
Pikeman: 3 (100%)
Fortified: +25%
Hill: +50%
City: +50%
Behind river: 25%
Defense modifier: 1+0.25+0.5+0.5+0.25=2.5
Modified defense; 3*2.5=7.5
If your archer is attacking a pike that is not fortified on flat land, then he is attacking a defense of 3.3 (flat land still gives a 10% def bonus)

Also, are your swordsman and archers veteran units? (4hp) or regulars? (3hp) or maybe even elites? (5hp) and the defending pikes? If you use a regular sword to attack a veteran pike, or an veteran archer to attack a regular pike, it makes a big deference. More HP means the unit can soak up more damage before it finally dies, meaning it will have an overall higher chance of winning the whole battle.

If these defense factors for the enemy pike are the same, then there is still the random factor.

A lot of people misunderstand how randomness works.
For example, if you have a 50% chance of something being true, then a lot of people would expect a series of events like this:

1011 0101 0010 1010 1010 1010 0110

As opposed to this:

1111 1011 1011 0000 0011 1100 1010

However, if you think about it, you'll realize the first series is much more predictable than the second, and thus less random.

Ever heard of the saying "bad luck never comes alone" Well, its true! Random events happen in strokes!

 

[bnj.s]

Hey MAS,

Thank you for your answer. It does fully answer my question.

About randomness, here is a quote from, I believe, John von Neumann:

The generation of random numbers is far too important to leave to chance.

Best regards

 

[Tasslehoff]

A lot of people misunderstand how randomness works.
For example, if you have a 50% chance of something being true, then a lot of people would expect a series of events like this:

1011 0101 0010 1010 1010 1010 0110

As opposed to this:

1111 1011 1011 0000 0011 1100 1010

However, if you think about it, you'll realize the first series is much more predictable than the second, and thus less random.

Ever heard of the saying "bad luck never comes alone" Well, its true! Random events happen in strokes!

This is not quite true. Randomness means that both of the strings you have shown have an EXACTLY equal chance of occurring. Specifically, (1/2)^28.

I think the point you were trying to make is that it is far less likely that

1010 1010 1010 1010... is less likely than a string having several 1s or 0s in a row.

 

[MAS]

@Tasslehoff

I'm sorry for not making myself more clear.
I was not talking about a set string of 28 50% chance events, and comparing it to an other set string of 28 50% chance events.
I was talking about the predictability of the next event being 1 or 0. Imagine both series of events going on for ever, and the first one continues to have about 2 1's and 2 0's in every group of 4.

If a string of 14 1's where to occur, a lot of people would be poised to think there is something wrong with the random number generator (RNG). While actually, its the RNG that is generating the first series is the one being less random.

 

[pickle]

in a random series the next number has an equal chance of being a one or a zero regardless of what has happened before, so any length sequence has the same chance of occurring as any other sequence of the same length. There is no such thing as more or less random, either the sequence is random or it is not. randomness having nothing to do with the numbers in the sequence but how they are generated. Which brings me to the fact that the RNG is not random. It is a pseudo random sequence, a sequence in which the pattern is very difficult to decipher. These are done by using a combination of shift registers and logic gates

 

[civverguy]

If you played civ enough, you won't need to know this. You can just know that X usually beats Y and X can't usually beat Z but X with artillery and beat Z.

 

[MAS]

@pickle

I realize I didn't explain myself very well again. (sorry for that) Of course, the chance of the next one being a 1 is 50% always. But look at the patterns I was trying to show with my 2 example strings. The first one has a more predictable pattern, but many people unconsciously expect it more than the second.

Also:
The pseudo RNG used by civ produces results that match a true RNG close enough that we can consider it the result of a true RNG for any of our discussions about randomness. (With the additional advantage that the same results can be re created as long as we save the seed and other factors)

 

[anaxagoras]

Just to clarify for the original poster, the game calculates combat based on attack factor versus defense factor as modified by various terrain and status factors and resolved using pseudo-random number generation as referred to above. But even though some units have special capabilities (such as the ability to build forts or to attack from boats) no unit type has any special modifiers versus any other unit type. So even though it seems that TOW infantry should be stronger defending against tanks and weaker against, say, regular infantry, that just isn't the case.

The combat system is very much simplified to keep the learning curve reasonable.

 

[Tribute]

If only there were a weapon triangle system, Civ combat would be really interesting. But I guess that would be CRAZY, though it would really emphasize unit diversity and tactics. No more fighting with only artilery under defenders. No more spamming cavalry. And there would be a point to building medieval infantry when longbowmen are around!

 

[Quintillus]

Civ4 kind of has that. Macemen (strength 8) have a 50% bonus against melee units. Crossbowmen (strength 6, 14% cheaper than Macemen) have a 50% bonus against Melee as well. Pikeman (strength 6) has a 100% bonus against mounted units. Knights (strength 10) have no bonus but are the strongest. War Elephants (no longer a UU) have strength 8 and a 50% bonus against mounted units. Longbows (strength 6) have a 25% city defense bonus. I forget all the unit costs, but that plays a role, too. What it comes down to is Macemen are usually the best choice, but a mix of crossbows and knights can really nix that strategy.

 

[PaperBeetle]

Which must have sounded really exciting and dynamic on paper, but in my experience the Civ4 system has two basic problems...
1. Too much detail. There are simply too many different unit classes and situations, against which various units have bonuses and penalties, with promotions providing further elaboration. I only played it for a couple of months, but I couldn't imagine ever being able to form some kind of mental map of which units to use against which.
2. The attacker has to select which unit to use for each combat, and the defender gets to pick the best response. So if both sides are using combined arms stacks of similar size and tech, the defender should expect to win every single combat!
So I think point (2) is the more serious problem for implementing rock/paper/scissor-style weapon triangles in Civ. Suicide artillery isn't a very meaningful solution to this problem, but I don't know what is. Thinking about RTS games which use weapon triangles, they don't have the same problem because both sides can simultaneously try to manouvre their units to get the match-up that they desire, so at the lowest tactical level, there is no such thing as "attacker" and "defender".
[/rambling]

 

[pickle]

@MAS
I get what you are saying, that they could both as likely be random sequences and that strings of ones or zeros often occur, i was just picking holes in your argument. But essentially i agree with you.

@PaperBeetle
I think that in Civ4 they were trying to make there be a point to having a varied military, but because the defender got to choose the best unit to take on the attacker it overpowered the defence a bit. the collateral damage from the artillery was a way of balancing things a bit, but not all that good a one. perhaps if some units had an ability like stealth attack in conquests then that would simulate being able to outflank the opposition.