What people mean about firepower...

[Trifna]

UUHHHH????

Bravo Venger. Now you established perfectly well the portrait it seems.

The big question for me is: how come Firaxis didn't see it? And how come Firaxis didn't saw a lot of little other things?

[Emperor Aries]

Ok, I don't have the game (yet) but from all of these arguments, I think Venger has a VERY valid point. I was talking with one of my friends that has the game before I read this and he told me about the whole spearman-beats-musketeer thing. I see two solutions to this problem. One of them has already been posted. Give the units a multiplier equal to their "age" such as [give warrior from "age" 1 a multiplier of 1 for all combat stuff so he would have no advantage and give musketeer from "age" 2 (or whatever age he is from) a multiplier of 2 for his attack, HP, and defense. That way the unit from the greater age has a better chance of winning but it isn't (might not, I don't know) garaunteed.

The other solution I see is to go over and take a look at Call to Power 2. In that game (as far as I know, didn't look at any programming BS) the unit with the greater attack value had a better chance of winning or something like that. You really just need to play the game to find out. But whatever system they had in call to power 2, it worked. You could have 2 nuclear subs go against 2 regular subs. The nuclear ones would usually win but just barely and might even lose one. In CTP2, the balancing was perfect, higher age didn't mean better but lower didn't necesarily mean loss.

This is totaly confusing, completely disregard this post!!!!!

[WhiteElephants]

Originally posted by Rusty Nail
WhiteElephant: OK.Your point is well taken, but Soren is still wrong. Referring to my above post, consider now:
Case 4: We double A's FP to 2 and halv his attack to 3. D is unchanged at D=4, FP=1, but his hit points are now reduced effectiely to HP=1 from HP=2 before (one hit and he is dead). So:

Prob A wins a round is now 3/3+4 =0.4286

Prob A wins = 0.4286 in 1 round + (1-0.4286)x0.4286 in 2 rounds
= 0.674, which is not the same as 0.648 as in Case 1 as claimed by Soren.

In this case the two results are fairly close. I have not analyzed other cases. Soren's claim may be a reasonable first approximation, but it is not correct mathematically. Further analysis is required to determine how good an approximation it might be. If the differences are always small, then he does have a good point. Maybe somebody out there can look at some other cases based on typical civ2 FP values.

Look, to be honest I'm not clear on your math and I have a sneaking suspicion that if you had carried out your example in a more realistic model (say each unit has ten hit points as opposed to a one hit one kill) the numbers would have been closer than they are. I'm under the impression you have set up this model as a model to prove your point (yes I know enough about statistic to know you can use them to prove just about any point). You've also changed the attack and firepower for one unit and left the other the same, how about changing both?

Example #1: Unit A has 3 Att. 2 FP, Unit B has 2 Def. 2 FP odds are 60%. (3/3+2 = .60)

Example#2: Unit A has 6 Att. 1FP, Unit B has 4 Def. 1 FP odds are still 60%. (6/6+4 = .60)

Given that in both example both units had ten hit points as opposed to 2, the results would be very similar. The reason your's were so different was because you included the speculation that one unit would be dead if it was hit once meaning that that unit could rely on a lucky hit to win whereas the other unit had to work for it, which would certainly skew your results.

[Sevorak]

WhiteElephants,

Think about Civ2. Take a Howitzer (12.2.2 2/2) and a defending Mech Inf (6.6.3 3/1).

Assume they have only 10 hp each. We'll use two different attacking units to demonstrate, a 12.2.2 unit with 2 FP, and a 24.2.2 unit with 1 FP.

12.2.2 Howitzer wins undamaged 13.2% of the time ([12/12+6]^5). 24.2.2 Howitzer wins undamaged 10.7% of the time([24/24+6]^10). The removal of firepower has created a decrease in the amount of times the Howitzer wins undamaged.

Therefore, Soren was wrong when he said it was identical. I'm not going to get into whether it is sufficiently identical or not, simply that the chances are not identical.

I am not making any judgements on how good or bad combat in Civ3 is by this statement, just math.

-Sev

[WhiteElephants]

Originally posted by Sevorak
WhiteElephants,

Think about Civ2. Take a Howitzer (12.2.2 2/2) and a defending Mech Inf (6.6.3 3/1).

Assume they have only 10 hp each. We'll use two different attacking units to demonstrate, a 12.2.2 unit with 2 FP, and a 24.2.2 unit with 1 FP.

12.2.2 Howitzer wins undamaged 13.2% of the time ([12/12+6]^5). 24.2.2 Howitzer wins undamaged 10.7% of the time([24/24+6]^10). The removal of firepower has created a decrease in the amount of times the Howitzer wins undamaged.

Therefore, Soren was wrong when he said it was identical. I'm not going to get into whether it is sufficiently identical or not, simply that the chances are not identical.

I am not making any judgements on how good or bad combat in Civ3 is by this statement, just math.

-Sev

Just as Rusty Nail has done you've reduced that amount of hit points in order to skew your results. The further you reduce the hit points the further the results will differ. Lets take your test and double the hit points in both situations.

Unit A with 24 attack and 1 FP = (24/24 + 6)^20 (because we've doubled the hit points) = .8^20 = 0.0115 or 1.2% of the time the unit goes untouched.

Unit B with 12 attack and 2 FP = (12/12 + 6)^10 (again we've doubled the hit points) = .666^10 = 0.0171 or 1.7 % of the time the unit goes untouched.

So now there's only a half of a percent difference, I will agree that it's a difference nonetheless, but really? Gotta love those stats! And if we carried out the equation the other way, say 1/2 the amount of original hit points the numbers would be even farther apart rather than closer together. Solution? Eliminate FP and reduce hit points as one precludes the other.

[Sevorak]

Just as Rusty Nail has done you've reduced that amount of hit points in order to skew your results.

I don't need to "skew" my results. Your example works just as well. All I set out to point out was that Soren is technically wrong when he says that they're identical.

On a more on-topic note, I'm starting to believe that there IS a serious problem with the random number generator, and I'm more and more convinced that it tends to generate long strings of related numbers.

I just rushed to Military Tradition in a game, I was the Russians and started churning out Cossacks. I then stormed five English cities in one turn, sending Cossack after Cossack against Spearmen.

The first Cossack goes in. It wins. Without being damaged. It's a Regular Spearman, but the chance of a Cossack defeating a Regular Spearman undamaged is only 27%. Sure, I figure it's a one-in-four chance, should happen every now and then. Then it happens again. And again. My first EIGHT Cossack attacks defeat their opposing Spearmen UNDAMAGED. That's a less than one percent chance! I should not be winning so much. And the ninth guy has a pretty ordinary back and forth combat, then the tenth Cossack falls on his face and loses 4-0.

I would really like to see RNG raw output. Long strings of related results, be they winning or losing, points to a flaw in the RNG.

-Sev

[WhiteElephants]

Originally posted by Sevorak
I don't need to "skew" my results. Your example works just as well. All I set out to point out was that Soren is technically wrong when he says that they're identical.

I just found it interesting that both you and Rusty Nail chose units with low hit points to demonstrate the differences. I assumed you realized that by doing so you were accentuating those differences and thought everyone else reading this thread ought to consider that. A 2.5 percent difference is 5 times greater than a .5 percent difference.

[WhiteElephants]

What? No comments?

^BUMP^

[Akka]

Originally posted by WhiteElephants


I just found it interesting that both you and Rusty Nail chose units with low hit points to demonstrate the differences. I assumed you realized that by doing so you were accentuating those differences and thought everyone else reading this thread ought to consider that. A 2.5 percent difference is 5 times greater than a .5 percent difference.

Mmh, actually, the Civ3 fight system is based on low hit points, while the Civ2 system is based on high hitpoints. And this thread was talking about the difference between those systems. So, any comparison with the Civ3 system SHOULD be based on low hit points, to see the actual differences between both.

[WhiteElephants]

Originally posted by Akka le Vil
Mmh, actually, the Civ3 fight system is based on low hit points, while the Civ2 system is based on high hitpoints. And this thread was talking about the difference between those systems. So, any comparison with the Civ3 system SHOULD be based on low hit points, to see the actual differences between both.

Right, but the addition of fire power requires the addition of hit points, and the subtraction of fire power means there's no need for high amounts of hit points. Otherwise the effectiveness of a 2 FP unit is just as effective against a unit with 1 or 2 hit points, and a unit with 3 or 4 hit points, and a unit 5 or 6 hit points, etc.

Besides the additon of FP to a combat system with such low hit points would only change the outcomes of battle by a round or two. A 2 FP unit could kill just about any thing in 2 successful rounds. Talk about odd outcomes? There would be no need for a unit with more fire power than that would there? Meaning the significance of a modern unit with 3 firepower vs. a industrial age unit with 2 would hardly make a difference.

[Akka]

Originally posted by WhiteElephants


Right, but the addition of fire power requires the addition of hit points, and the subtraction of fire power means there's no need for high amounts of hit points. Otherwise the effectiveness of a 2 FP unit is just as effective against a unit with 1 or 2 hit points, and a unit with 3 or 4 hit points, and a unit 5 or 6 hit points, etc.

Besides the additon of FP to a combat system with such low hit points would only change the outcomes of battle by a round or two. A 2 FP unit could kill just about any thing in 2 successful rounds. Talk about odd outcomes? There would be no need for a unit with more fire power than that would there? Meaning the significance of a modern unit with 3 firepower vs. a industrial age unit with 2 would hardly make a difference.

I suppose that's why so much people ask to simply come back to the Civ 2 system, with FP and a big bunch of hit points

Personnally, I'm all for the FP/MP system, if only for the scenarios. I mean, if they don't want to change the balance of Civ3, they could at least include the FP/MP system and set it all to 1. But being able to make special units with many HP or high FP in a scenario is invaluable tool. Much better for subtle tweaking than just putting insane amount of A/D.