 |
|  |
 |
|
vulture
|
 |
Leeds, UK
Jun 2001 time: 05:32
|
|
Some more numbers from the combat system clarified by Jesse. This is using the combat system as works in civ3 (well, almost - I know the civ3 system essentially generates random numbers from 0 to 1023 (IIRC) - 10 bits of precision rather than the 26 or so my program is using), generating 4 random numbers and taking the average for each combat round. Retreat and promotion chances are handled by a single random number, rather than an average one.
For various things attacking a single veteran spearman, fortified in a town (10% terrain bonus, 25% fortified bonus, 2.70 total defense strength). All attackers are veteran. Numbers listed are % chance of spearman being killed in the fight.
attackers PtW C3C
1 horse 30.45 12.07
2 horses 69.24 38.92
3 horses 88.41 61.82
4 horses 96.02 77.82
1 sword 55.74 64.95
2 swords 91.02 94.90
So in vanilla and PTW, 2 horses were slightly better than 1 swordman, if your criteria is the chances of taking a town defended by 1 spearman. In C3C beta, we now have 3 horses are roughly equal to 1 sword - you need to bulk up the size of your horse forces by 50% or more (I'd say more if taking on towns with more than 1 defender). The chance of a single horse in the circumstances here is reduced by more than a factor of two.
Note that if you want a 60% chance of taking a town (not brilliant), you need 1 swordsman (30 shields), and lose him 0.44 units on average (44% chance of losing 1 unit). Or you can use 3 horses (90 shields, 6% better chance), and lose 0.93 on average - that's roughly 65% chance of losing at least one, with 25% chance of losing 2 or all 3.
I suspect horse warfare just became something where you need to replenish losses at a far greater rate to carry on the campaign.
Definitely time for the artillery - it just became much more powerful.
|
|
|  |
 |
|  |
 |
|  |
 |
|
vulture
|
 |
Leeds, UK
Jun 2001 time: 05:32
|
|
quote: Originally posted by alexman
Dramatically increase retreat odds for fast movers: So horsemen are again an alternative to swordsmen as an offensive unit in the ancient age.
|
I'd have thought that the flip side to this would be needing even more horses in the first place, since each horse then has a reduced chance of getting the kill.
Then I tested it out - to take an extreme case, I put retreat to 100% (and kept the usual rule that a fast attacker doesn't retreat when the defender is on 1 hp).
This really didn't change much. With the usual veteran retreat chances, 4 horsemen had a 77.8% chance of killing the spearman, suffering 0.93 average losses. With 100% retreat chances (back to old vanilla civ3), they had a 78.0% chance of killing the spearman, suffering 0.67 casualties on average. Maxing retreat hardly affects the odds of winning, but does reduce the casualties by 1/3 in this case. I think that's pretty small compared to the PTW -> C3C beta change.
But this is a beta patch. It's as much to test out how this idea works in practice as anything else. If after playtesting by everyone here (and elsewhere) we find it is horribly unbalanced, then it probably won't appear in the release version of the patch. If it works and everyone is happier with combat results, it will stay.
BTW I don't think it will slow the game up any. User combat time is dominated by the graphics, and for off-screen AI-AI combat, the calls to the RNG aren't a huge part of the time taken calculating combat results.
|
|
|  |
 |
|
Cryptor
|
|
quote: Originally posted by sabrewolf
but that's what beta-versions are for: to test features _before_ they get released officially.
|
I agree totally. Tell the faraxians that. Tell them not to test new features, such as the new GPT feature and corruption feature, in their box retail releases. You may say these are bugs, but they're bugs because they decided to experiment in a main release with something which wasn't broken before. I could forgive this mistake if they promptly released a patch which address these two very important issues, reversing the experimentation they did with the initial release, but they aren't. They're releasing a patch that breaks the game in a new and exciting way. If this experiment is important to them, it would take a good 10 minutes of work to include a menu option that makes it default but optional. Whether their customers are worth that 10 minutes work is anyones guess, but soon to be revealed.
|
|
|  |
 |
|
asleepathewheel
|
|
listening too long to one song
Mar 2002 time: 00:32
|
|
quote: Originally posted by Cryptor
I agree totally. Tell the faraxians that. Tell them not to test new features, such as the new GPT feature and corruption feature, in their box retail releases. You may say these are bugs, but they're bugs because they decided to experiment in a main release with something which wasn't broken before. I could forgive this mistake if they promptly released a patch which address these two very important issues, reversing the experimentation they did with the initial release, but they aren't. They're releasing a patch that breaks the game in a new and exciting way. If this experiment is important to them, it would take a good 10 minutes of work to include a menu option that makes it default but optional. Whether their customers are worth that 10 minutes work is anyones guess, but soon to be revealed. |
I've really missed these posts!
"I haven't even played Conquests, but its broken beyond playabilty"
"I haven't used the beta patch, but I know they royally screwed up combat"
A bit of a credibility problem. Speculation at what is going to happen is all well and good, but the above is just...ludicrous. download the patch next week, then ***** all you want if you feel its necessary. Until then, you are as in the dark about it being broken as someone who thinks the change is great. And what's up with questioning Firaxis' valuation of the customer? Is that really necessary, I think it detracts from your points.

|
|
|  |
 |
|
The Kaiser
|
|
Stoke-on-Trent, England
Dec 1999 time: 05:32
|
|
I'd prefer it if they saved this new formula for Civ 4.
Firaxis, please just fix the bugs and leave the combat calculations alone. I like losing a Battleship to a Ironclad now and then because I know that when the roles are reversed it could happen to my own advantage also.
I want both sides of a battle to have an even chance of winning, according to their Unit values regardless of the "apparent" stupidity of it.
We always have posts from players who Whine and complain about Spearmen defeating their Tanks but never hear about the times when their own Spearmen have defeated a Tank etc.
Is it because these Lamers never have the self discipline to finish a game where it looks like they won't take over the whole map for a change? Personally I enjoy the games when it becomes apparent that I don't have any chance of winning but the challenge to me is the surviving against all odds with little or no resources.
I repeat, please don't include it with this patch, use it later or in Civ4 or something.
|
|
|  |
 |
|
Gyromancer
|
 |
North Carolina, USA
Jan 1970 time: 00:32
|
|
OK, for those of you who want to try it, I worked out a way to mod the game to simulate the combat changes using the best 3 out of 4 method (I need a little more free time at work to get the averaging method done. )
I don't know if there is a way to mod the unit fortification bonus. (at work ATM...) If there is, change it from 25% to 60%, otherwise see below.
Mod attack/defense values for all units as follows:
1 becomes 1
2 becomes 5
3 becomes 15
4 becomes 30
5 becomes 54
6 becomes 86
If you can't mod unit fortification bonuses, and you are interested in something where the town assaults on standard terrain will be close to the new method, use the following
1-> 4 attack / 5 defense
2-> 24 attack / 31 defense
3-> 60 attack / 78 defense
4->120 attack / 156 defense
5-> 216 attack / 280 defense
6-> 344 attack / 447 defense
This method gives all units the extra piece of the fortification bonus regardless of whether they are fortified, so the results from open field combat will be a little off.
I could work on some terrain mod numbers, if anyone wants those.
Edit: For example, the first mod scheme changes swordsmen from 3/2 to 15/5 the second scheme changes swordmen to 60/31
Edit #2: Fixed math error in calculation of factors. I'm amazed at some of the equivalences now.
Last edited by Gyromancer on 12-12-2003 at 02:00
|
|
|  |
 |
|  |
 |
|  |
 |
|
ducki
|
|
I'm glad they decided to separate it out of this patch, but rather than them experimenting and deciding on the issue, I'd like them to experiement and decide what sort of customizable option to give the player to decide on.
Give us a "PRNG Streakiness" item in the C3C.ini file or make it part of an "Advanced Options" screen. Players that don't care, don't have to wonder what it is. Players that want to tweak it can try out the settings for minimizing streaks.
At least, that's my pie in the sky item for the wishlist. I like the streakiness of the PRNG, because without the streakiness, you'll never see a weaker unit string together enough individual HP roll wins to finish off a stronger vet unit(at least 4 rolls out of 7, right?)
I don't want to get off into probability theory, though, so suffice to say...
Thank you Firaxis!. Thanks also to the Firaxians that frequent these forums above and beyond their job description. Thank you.
|
|
|  |
All times are GMT. The time now is 05:32. Apolyton Time is 00:32. |
top of page
|
|
|
Forum Rules:
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is ON
vB code is ON
Smilies are ON
[IMG] code is ON
|
|
|
|
|
|