31-01-2001 17:23									#32
	Support Apolyton, buy Civilization: The Boardgame

My idea about 100% for a defender originated when I tested to attack dips/spies by warriors.
Now, I have found my record of older tests: only 3 vets after 20 successful defences.
It may depend on power of a unit...

	08-02-2001 19:28									#33
	Support Apolyton, buy Galactic Civilizations

As I understand, the Hitpoints and Firepower stats were added in Civ2 so that there would be a random element rather than a straight win for attacker or defender.

If you had a unit of A3,H1,F1 and attacked a unit of D3,H1,F1, its different from attacking a unit that is D1,H3,F1. Its also different from attacking a D1,H1,F3 unit.

In the first case, you might win or lose but the surviving unit will be half dead.

In the second case, if the defender survives, it will be almost dead

In the third case, if the defender survives, it will be only slightly damaged.

A second thing to take note is that HPs and FPs become important with modification that takes when air, helicopter and naval units fight land units or when city improvements like city walls are involved. Some modifications reduce the hit points while others multiply the attack factor without modifying the hit points. This give more depth in the design of interaction between units.

To understand the use of hit points and firepower, you need to take them to their extremes. In a game where all units have FP=10 and HP=1

You can have single blow deaths
The probability that a weak unit kills a strong unit becomes significant
Ships performing shore bombardment or defending cities are not disadvantaged - their firepower is reduced from one to one!

In a game where all units have FP=1 and HP=10

The weaker unit is sure to die
The stronger unit is sure to get damaged
Ships performing shore bombardment or defending in cities have fire power reduced to one (i.e. 10 times)

In the case of air units attacking ships in port, the planes have firepower doubled so the effect is not so pronounced as the ship example but there is still an effect.

	08-02-2001 20:50									#34
	Support Apolyton, buy Galactic Civilizations: Deluxe Edition

Pretty much trying to hang on to some of the discussion here by its coat-tails (and failing to do so).

But I'll throw in a comment on vet status. I seem to have noticed that a unit gets to be made up more often when the outcome of a particular combat was a close run thing. And (just maybe - memory loss makes me doubtful) it may be that this is more of a factor for the successful defender than for the attacker. If there's anything in this, I'm not altogether sure whether the state of health of the unit at the outset matters - but I think it does.

Or, put another way, a unit that emerges badly damaged but alive is likely to get made up.

So a hardy phalanx which repells a chariot on open ground, dropping into the red while doing so, would become a vet. The same phalanx, fortified on a mountain, that sees off the chariot without breaking sweat probably won't.

	09-02-2001 00:35									#35
	Support Apolyton buy from Amazon

East Street Trader,
You may be right...Good idea

kobayashi,
I don't agree. If I have understood the combat system well, then (accordingly to my post "posted January 29, 2001 19:03")

a)A3,H1,F1 attacks D3,H1,F1,
attacker wins (3-1/8)*1*1 times
defender wins (3+1/8)*1*1 times
(Note: I have doubled numbers in my table: Total number of attacks = 2m)
If we disregard eights then wins are divided in a ratio 3 to 3, i. e. 1 to 1

b)A3,H1,F1 attacks D1,H3,F1,
attacker wins (3+2-1/8)*1*1 times
defender wins (1+1/8)*3*1 times
wins are divided in a ratio 5 to 3

c)A3,H1,F1 attacks D3,H1,F3,
attacker wins (3+2-1/8)*1*1 times
defender wins (1+1/8)*1*3 times
wins are divided in a ratio 5 to 3

quote: Some modifications reduce the hit points while others multiply the attack factor without modifying the hit points.

I don't understand what do you want to say...

quote: To understand the use of hit points and firepower, you need to take them to their extremes. In a game where all units have FP=10 and HP=1... ...In a game where all units have FP=1 and HP=10...

In short, a quotient H/F lowers accidental behaviour of units. OK?

{combat system}{combat formula}
{pikemen}{vet status}
{}{SlowThinker}{end2}

------------------------------
This is a post with keywords. See a thread The Great Library: a hierarchical structure" thread.
[This message has been edited by SlowThinker (edited April 15, 2001).]

	09-02-2001 08:36									#36
	Support Apolyton

Well I guess in the simplest terms,

Your calculations are not wrong as far as the statistical average is concerned. In a two system world, either A is killed or B is killed, they odds
are as you state them.

But they simply do not take into account

1. the distribution of events, i.e. the shape of the distribution curve

2. the amount of damage the survivor sustains after the opponet is killed.

If every unit had 1000 hit points, the outcome and damage would be exactly predictable, the stronger unit would always win everytime and be damaged by the same amount.

If every unit had one hit point, by accident, the weaker unit might kill the stronger unit one time in five but in exchange the stronger unit is less likely to be damaged when he does kill the weaker unit.

	09-02-2001 09:05									#37
	Lose 30 kilos (of popups)

To further clarify:

if ONE (2A 1H 5F 1H) attacks TWO(1A 1D 5F 1H)
(remember that 1H = 10 hit points)

Every round ONE has a 2/3 chance of landing a blow and TWO has a 1/3 chance of landing a blow.

Lets say ONE wins the first two rounds. In each round TWO loses 5 hit points and after losing 2 rounds, TWO is dead and ONE is undamaged.

TWO only has a 10.9% (33%x33%) of winning the first two rounds but if it does, ONE will be killed.

BUT if ONE (2A 1H 1F 5H) attacks TWO(1A 1D 1F 5H)
(Now they each have 50 hit points)

After ONE wins two rounds, TWO is down to 48 hit points from 50, far from dead. Now you need to win fifty rounds to kill an opponent (as opposed to two above). There is no statistically significant way that TWO can win fifty rounds before ONE so TWO is as good as dead - but ONE is sure to get damaged in the process.

	09-02-2001 09:10									#38
	Inflate your Upload Space

Opps, I missed the last line of your post.

Yes, in short, a quotient H/F lowers accidental behaviour of units.

	03-03-2001 06:04									#39
	Support Apolyton buy from Amazon

My understanding is that HP/FP have two effects to change the game from Civ1.

1. combat damage is accumulated and units must "heal"
2. The FP/HP tends to make thestronger units more consistent at killing the weaker units. (i.e. removes much of the chance element that a battleship will lose to a phalanx becuase mulitple rounds of combat are simulated.

I have always thought that element 2 is a little bit of a crutch for the human and makes things too easy for the player who races ahead in technology, makes it easier to calculate exactly how many crusaders you need to capture a city, etc.

	06-03-2001 23:15									#40
	Support Apolyton, pre-order Civilization IV

Forget all the formulas and calculations. What it comes down to - explaining a concept behind the algorithm - is that more hit points mean the unit can survive more damage. More fire power means the unit can deal more damaging blows.

For a simple comparison, attack a warrior and a settler with a horseman.

warrior: 1d 1h/1f
settler: 1d 2h/1f

The horseman will have more success against the warrior because it is twice as stong (2a vs. 1d). The settler is just as weak as the warrior, but takes twice as much damage to kill. In other words, the settler is as hard to kill as 2 warriors in succession, all other factors being equal.

	29-10-2003 00:54									#41
	Increase Your PM Length

I posted a summary of math of this thread in Info: Combat (GL)

	29-10-2003 01:55									#42
	Get a bigger avatar today!

Complex algorthims are fine for SP when you have time to screw around, but Marguis is right that it's best to understand the concepts behind it when you're calcing on the fly in an MP game. I'm fascinated by the combat threads and believe I have a firm grasp on it, but in an MP game I just do rough calcs and estimate because the outcomes vary. I just want a good guess on how many vet crooks (or whatever) I'll need to throw at three pikeman on whatever terrain they're defending.

	29-10-2003 02:23									#43
	Support Apolyton, buy Galactic Civilizations: Deluxe Edition

I think the algorithm I posted (with point B. omitted) is just what you are speaking about: it doesn't tell you exact probabilities, only an approximate strength of units. And it is fast if you are familiar with it.