Combat Odds?

[NackBarnes]

Has there ever been a Civ game that gave you even remotely accurate odds on combat?

In BtS with new random seed on reload if the odds are 3 to 6 then figure you will win 1/3rd of the time.

Huh. Reloading 10, 15 times, you might win after that many. Shouldn't the odds display 1 to 10 then? Or 1 to 15? And this is consistent, all ages and military units. Rarely will you win on a less than 1 to 1 combat and the odds don't really tell you how UN-likely it is

I might gamble on a 1.5 to 2.0 fight. But if it is REALLY 1 to 10 but I'm being told by the program it's 1.5 to 2.0 then I'm being screwed over by the program. *laughs*

[NackBarnes]

Turn 51, Marathon, Prince level. Archer with no promotions vs. Archer with City Def 1 in +20% city, fortified.

3.00 to 6.45 odds.

Figure one out of three, one out of four, the attacker wins that, yes?

So far, 25 reloads, not one victory. Why can't we get real odds?

[rjmatsleepers]

For reasons quite unconnected with testing the probability display, I have replayed an attack sequence a large number of times. I have five axemen against three barbarian archers. Promotions have varied slightly, but the displayed odds (on the first attack) are between 60% and 70%. The most frequent result has been three victories for my axemen. A close second is one defeat. Two defeats is rarer, but none the less reasonably common. I have had three defeats a couple of times and never four defeats. I haven't calculated the statistical tests, but these results seem to be exactly in line with the displayed probability.

RJM

[NackBarnes]

I ran the example above 60 times before I just got tired of it.

60 to 0 is a lot different than 3.00 to 6.45 or whatnot. Heh.

[couerdelion]

You should be able to get a good idea if attack odds are wrong by looking at the combat stats and bonuses.

ie Strength 3 vs Defence 6 will never get you 30-40% win probabilities

[Risa]

You made wrong assumptions how combat resolves. A combat of 3.00 vs 6.45 is far worse than the probability you thought.

[NackBarnes]

3 to 6 odds is as I said earlier, 1 out of three.

If Civ IV says "3.00 to 6.45" and means something other than 3.00 to 6.45, can someone explain what it means, so I can read the combat odds accurately?

If I've made wrong assumptions how combat odds mean do you have a resource you can point me to that explains what "3.00 to 6.45" means in Civ-speak?

[ColdPhoenix]

I'm not sure how it works but you can see the odds of winning by hovering over the unit to be attacked with "go-to".

It doesn't work the way you're saying though. The number is not only the chances of each unit hitting the other but the damage they do when they hit and how much health they have. Not sure what the % would be for 3 to 6 but I reckon it would be less than 10% chance of winning. Someone who knows more will probably come along give a more accurate description.

[NackBarnes]

Hrm, guessing I was still stuck in Civ I and II. Then, 3 to 6 means out of 9 shots you'd have 3 wins and 6 loses (more or less). Never really worked OUT that way, but that's what the odds system meant.

I note in Civ IV it doesn't say 3 to 6, it says 3 vs. 6, a subtle but important distinction that I have missed in spite of 5 zillion hours of Civ IV play. *sighs*

[Kuciwalker]

Originally posted by NackBarnes
3 to 6 odds is as I said earlier, 1 out of three.

If Civ IV says "3.00 to 6.45" and means something other than 3.00 to 6.45, can someone explain what it means, so I can read the combat odds accurately?

If I've made wrong assumptions how combat odds mean do you have a resource you can point me to that explains what "3.00 to 6.45" means in Civ-speak?

Read the combat odds by looking at the ****ing percentage. The strenght numbers are only meaningful if you've got a clue how combat resolves.

[NackBarnes]

Wow, thanks for that helpful data.

[Virdrago]

Just mouseover while holding the right-click button, and it will show the odds for 3.00 v 6.45 (or thereabouts). I'm pretty certain it will be around 5%, probably (a lot) less. I'm certain someone has an algorithm for this; I just use what the computer says and save the calculating for the calculators...

If the chances are around one percent or two, you have to do approximately 200 to 300 tests to have it even out. Good luck.

[Kuciwalker]

Originally posted by NackBarnes
Wow, thanks for that helpful data.

It was, it solved your problem completely

[Blake]

It works something like this:

The base damage per combat round won is 20.

This is modified by strength ratio,
the 3str unit will deal about 15 damage/round (must win 7 rounds to deal 100 damage)
the 6str unit will deal about 25 damage/round (must win 4 rounds to deal 100 damage)

Of course the odds ratio also determines the chance of winning a round. Not only must the 3 str unit win almost twice as many rounds to be victorious, it also has only about half the chance of winning a round. This double whammy is why a little extra strength gives a significant increase in victory odds, and why units which are badly outmatched have little chance of winning.

The full details can be found around, but you don't really need to know them, the odds mouseover is reliable enough.

[NackBarnes]

Thanks, that explains pretty well, Blake.

Find it odd when a 4 vs 3 sort of fight ends up with an 8% success chance, but still, knowing that it is 4 vs 3 instead of 4 to 3 helps my understanding a lot!

So, this pretty much puts the B.S. to the "experienced but obsolete units can beat newer green units" civ iv help blurb, eh?