Fast Unit Retreat

[Quasimodo]

What in tarnation determines if a fast unit retreats or dies when it gets down to one hit point??

I was attacking a city, that was defended with a regular spearmen, with half a dozen horsemen. They all had full movement and hit points. It seemed every other one died in the attack while the others backed off when they got down to one hit point.

It was on grasslands with no hills.

Just curious,

Thanks.

 

[Oystein]

Only fast units fighting slow may retreat. Also, units defending a city wouldnt retreat. The unit must obviously have a place to retreat to. The other unit has to have more than 1 hp. Migth be some other restriction I am forgetting.

When retreat is possible, the chance is 100A/(D+50), where A and D are the retreat bouns for the attacker and defender.

Retreat bouns can be found in the editor, they are
34, 50, 58, 66 for conscript, regular, veteran and elite.

If you want to read about the investigation:
http://forums.civfanatics.com/showthread.php?s=&threadid=64755&perpage=20&pagenumber=1

 

[Wakboth]

This has been discussed in another thread (the one about armies being fast units).

I'll repeat my understanding as it applies to regular units:

Rules of retreat:


1. In order to retreat, a unit must be a "fast" unit (typically units with more then 1 movement point).

2. A fast unit cannot retreat if it's fighting another fast unit.

3. A fast unit will not retreat if the unit it's fighting is at 1hp.

4. During any round of combat in which a qualifying unit begins the round at 1hp, there is a chance of that unit retreating (based on the unit level). A retreat will move that unit one square away from the combat, and avoids that round of combat entirely (and ends all remaining combat with that unit).


Your horsemen have an attack of 2. The spearmen have a defense of 2, and were presumably fortified (giving them a 3?). I have no idea of the size of the city, but considering horsemen and spearmen were involved, we'll assume no additional bonus in this case.

You don't say if you had veteran or regular horsemen. However, even considering veteran horsemen, you're going to see lots of losses.

You've got 4 hp, he's got 3. However, he'll do damage to you at a 3:2 ratio. One way to do a quick and dirty comparison is to say that for every 5 total hps between you and him, he'll dish out 3 and you'll dish out 2. What's really nasty here is that you have 4 and he's got 3. So our first nominal "round" of 5 hps results in your losing 3 and him losing 2. That results in both you and him having 1 hp left. Per the rules above, your horseman will *not* retreat.

I don't feel like calculating the odds of variation results. However, it's not uncommon at all for you to end up in that 1 to 1 hp situation during this fight, so that's why you didn't get many retreats. You should get a roughly even number of times when you'll come out ahead of that 1 to 1 or he'll come out ahead. If you are ahead, you probably won the fight, so retreating didn't matter. If he's ahead, you'll lose the fight, but each round that he's ahead there's only a chance of you retreating. Interestingly enough, your odds of losing that last hp (60%) is almost the exact same as your chance of retreating.

Just eyeballing the odds, I'm surprised you got 50% of your horsies to retreat. Should probably have been more like 30%.


EDIT: What's even more interesting is that I think you'd lose less units fighting with regular horsemen in that situation then veteran (they're more likely to hit 1hp while the spearman still has 2 or more). You'd do less damage to the spearmen per attack, but you'd have more live units after the battle. Go figure...

 

[Ranos]

@ Wakboth

I don't claim to be an expert, but I think you're wrong about how battle damage is calculated. Battle is all based off of a die roll type of thing. I believe the way it works is that the computer looks at what the attack and defense are, 2 and 3 respectively in this case, and builds a 5 sided die, in the computer sense that is. Lets say that 1 and 2 are assigned to the horseman and 3-5 are assigned to the spearman. It then "rolls" this 5 sided die and if it lands on one or two, the spearman takes damage. If it lands on 3-5, the horseman takes damage. So, the spearman has greater odds of hitting the horseman, but it is completely random, which is why a warrior can damage a modern armor, even though its highly unlikely.

The reason I say this is simple. If your way of doing it was how damage was calculated, then how could a spearman ever damage a modern armor?

As far as I know, the only benefits of having a higher experience level is the hps the unit has and its chance to retreat if its fast.

About your edit, that makes even less sense. If you claim that a 3 def would cause 3 hp damage and a 2 off would cause 2 damage, then with a regular horseman, after the firts battle, the spearman would be down 2 hp to either one or two, depending on if it was regular or veteran, and the horseman would be dead.

No offense intended by this, just wanted to put that out there. If somebody knows for a fact how this works, please let us know.

EDIT: Fast units retreating used to be gauranteed when Civ3 first came out but was modified in one of the patches or with PTW. Now it is a percentage. Default is as Oystein said: 34%, 50%, 58% and 66%.

So, if you get to the one hp left situation, you will either win or lose but you won't retreat. If the defender still has more than one, you get a die roll based of of the attackers exp. level. If you have a regular horseman, the die will be 1-50 are retreat and 51-100 are stay and fight. If they are veteran, 1-58 are retreat and 59-100 are stay and fight.