You know the combat system is FUBAR when...

[Pyrrhos]

...it takes 37 reloads for an elite 4-4-1 unit to defeat a non-fortified regular Impi in the open without losing a hit point whereas in that time, the Impi manages to kill the elite 4-4-1 unit no less than twice without losing a hit point itself and four times losing just one.

...when the aforementioned elite 4-4-1 units on average lose two hp killing unfortified warriors and three hp killing spearmen and swordsmen as seen from several hundred combats.

...when you, fed up with this rank cheating, create a 24-48-3 superunit with 10hp in order to test if something "funny" is going on and find that the superunits still manages to lose hitpoints to unfortified regular warriors, spearmen and swordsmen and where losing 3-4 hp in one combat is not uncommon and once even being redlined by an unfortified veteran swordsman on a hill.

Way to go, Firaxis!

PS. I must be suffering a massive attack of cognitive bias.

 

[Chamnix]

*Looks for the attached save - finding nothing attached (again) walks away shaking his head.

 

[Mursi lives]

Never saw a 4/4/1 unit defeated by an impi.

 

[Pyrrhos]

*Looks for the attached save - finding nothing attached (again) walks away shaking his head.

Since I never play with the Preserve Random seed...

...knock yourself out! Should keep you occupied the next week or two.

 

[TheOverseer714]

I came close to last night, a regular pikeman defeated a healthy tank of mine. The town was on a hill and the pike was fortified. I did red-line a tank killing a regular spear on grass the same turn.

 

[Pyrrhos]

Bad as "spear" is, my favourite FUBAR candidate is when an elite BB of mine lost two hp to a curragh. When you think about it, it's ridiculous. A curragh was a craft built using saplings and greased hide and wouldn't even scratch the paintwork of a BB. "The intrepid curraghsman paddling his craft at 30+ knots chipped off flakes of paint with his stone axe."

 

[Chamnix]

Since I never play with the Preserve Random seed...

...knock yourself out! Should keep you occupied the next week or two.

Thank you , but wouldn't it make more sense to give me the save that you reloaded 37 times instead of the 4000 BC save? If your claim is still that the RNG is biased only at certain points of game, how am I supposed to be sure I reach one of those points? I've never managed to reach one before...

 

[timerover51]

...when you, fed up with this rank cheating, create a 24-48-3 superunit with 10hp in order to test if something "funny" is going on and find that the superunits still manages to lose hitpoints to unfortified regular warriors, spearmen and swordsmen and where losing 3-4 hp in one combat is not uncommon and once even being redlined by an unfortified veteran swordsman on a hill.

Way to go, Firaxis!

PS. I must be suffering a massive attack of cognitive bias.

Based on the combat unit testing that I have been doing, the AI bias in the RNG gets worse as the tech differential and combat value differential increases between the human and AI player increases, and as the hit point differential increases. I have had a longbowman in the open take down one of my upgraded tanks, after being redlined, he took down the remaining 8 hit points on the tank. (Note, I said it was upgraded.) I have also had a trebuchet do 2 hitpoints of damage to one of my upgraded battleships from inside a city, and because I do not have the battleship set for lethal land bombardment, I really chewed up the city trying to knock it out. Finally got it with a bomber. So far, no ancient or medieval units have been able to engage aircraft, so I have been using bombers a lot. Hard on the landscape and cities though.

The problems with the AI have confirmed my belief that board and miniature war games are superior to computer games when you can find human opponents. Lacking those, I figure that I will have to keep modifying units, and periodically gnash my teeth at the AI for Civ3, and give serious thought to hacking into the game files and finding the source of the problem.

Against that, the bizarre combat results at times do keep you on your toes. I was once running a test game for a naval war game that I have designed, and had a player miss 4 times at a 90% hit probability range, for a 1 in 10,000 odds event. Then, a few turns later, the same player took a controls hit on his Italian battleship, went out of control, and RAMMED the best British battleship by accident, sinking it immediately. Things like that give game designers and testers fits, and have them talking to themselves.

 

[timerover51]

Bad as "spear" is, my favourite FUBAR candidate is when an elite BB of mine lost two hp to a curragh. When you think about it, it's ridiculous. A curragh was a craft built using saplings and greased hide and wouldn't even scratch the paintwork of a BB. "The intrepid curraghsman paddling his craft at 30+ knots chipped off flakes of paint with his stone axe."

Hmm, that beats my trebuchet verses battleship by quite a bit. A curragh, no less! And 2 hitpoints, plus that means that it survived the battleship long enough to do that.

 

[D0NIMATRIX]

I get this kind of stuff a lot, mainly because I often fight obsolete units with tanks and stuff. has happened quite a lot, but my personal favorite is losing a fortified full health elite rifleman to a spearman.

 

[Simak]

I've noticed that I lose elite units more often then veteran ones when I'm stomping the cr*p out of the AI. So, usually, I'll send in the veteran ones first......illogical but seems to result in less lost units.....go figure?

 

[D0NIMATRIX]

I notice that whenever a unit does really well, the next one will get massacred, and vice versa.

 

[Lord Emsworth]

Who'd thunk ... ?

*yawns*


*unsubscribes*

 

[Pyrrhos]

Thank you , but wouldn't it make more sense to give me the save that you reloaded 37 times instead of the 4000 BC save? If your claim is still that the RNG is biased only at certain points of game, how am I supposed to be sure I reach one of those points? I've never managed to reach one before...

The thing is that since I play with the "preserve random seed" function disabled, you would never be able to locate that exact sequence of the prng. But if you play this at least as well as I did to around 500AD (not hard to do, I assure you), build loads of legions plus three to four "slavemasters", wait for your neighbours to declare and do not accept their proposals for peace, then you should see all these things for yourself.

As a bonus, you'll experience a mod that has some interesting naval features. (And learn to trust my word, one hopes. )

 

[Pyrrhos]

I notice that whenever a unit does really well, the next one will get massacred, and vice versa.

I've seen that a lot too. It as if there are built-in factors such as "If player X's score is 50% greater than Player Y's, add +1 to the attack and defence values of player Y's units", "If Player X's score is 100% greater than Player Y's, add +3 to the attack and defence values of player Y's units" and "If Player X's unit won the last combat without losing a round (=hp), subtract 5 from the prng results for player X for the next combat". I suppose it was put there so that no player, human or AI, could run away with the game.

PS. Did Lord Emsworth go back to "Command and Conquer" or Bullfrog's classic "Assassin"?

PPS. We'll see you back soon, Lord E! Probably before the week is out if not the day.

 

[Pyrrhos]

I've noticed that I lose elite units more often then veteran ones when I'm stomping the cr*p out of the AI. So, usually, I'll send in the veteran ones first......illogical but seems to result in less lost units.....go figure?

That's the impression I get too. It sometimes seems (I know, cognitive bias) as if an inferior regular unit such as a warrior or spearman when attacked by an elite immortal or legionary, takes its own number of hp off the attacking superior unit.

 

[SeriousCaller]

Hmm, I have my suspicions about that too. If I have a mix of vet and elite cav for example and I go for a little leader fishing, I find that my vet cav will kill a defending musket every time, but my elite cav will get redlined and run away, or worse I find there was a horseman garrisoned in the town so my elite cav die and I take the town with vets, gaining no leader.

 

[Chamnix]

The thing is that since I play with the "preserve random seed" function disabled, you would never be able to locate that exact sequence of the prng. But if you play this at least as well as I did to around 500AD (not hard to do, I assure you), build loads of legions plus three to four "slavemasters", wait for your neighbours to declare and do not accept their proposals for peace, then you should see all these things for yourself.

You disappoint me, Pyrrhos - we have already had this conversation. As I mentioned last time, I have no intention of duplicating anyone's exact sequence. I simply want the save that you played 37 times, so I can play it 37 times... and another 37... and another 37... and another 37. Will I get results that (like yours) cannot realistically be explained by an unbiased RNG, or will I get results that are more in line with what you would expect from an unbiased RNG?

Which do you think will happen, and why are you so reluctant to find out?

 

[Pyrrhos]

I'm not reluctant at all and I am very confident about what will happen, I just don't have that particular save anymore! You, however, seem reluctant to go to the trouble of playing through the game in order to see for yourself. Are you afraid of what you will find or is it that you doubt your own ability at Emperor to reach such a position of superiority that these things will manifest?

Just do it, buddy!

 

[Chamnix]

I can understand my request for a save this time may have surprised you - however, for future reference, the next time you have incontrovertible proof of AI bias in the RNG, please be aware that I would love to see a copy of the save and would greatly appreciate it if you would post it before deleting it.

I am so tired of excuses that I will even play by your rules. I will play your save until 500 AD, but I need a couple things from you first:

if you play this at least as well as I did to around 500AD (not hard to do, I assure you), build loads of legions plus three to four "slavemasters", wait for your neighbours to declare and do not accept their proposals for peace, then you should see all these things for yourself.

1. I need a definition of playing “at least as well as you did”. Do I need a higher power score, more citizens, more territory, more luxuries, more techs, all of the above? I need to know approximately what you had at 500 AD in whatever categories are relevant so I know what I am shooting for (obviously, I would prefer to know exactly what you had – it’s a shame about the lack of a save).

2. I need to know what other conditions there are to reach this point of AI bias. Apparently, the AI (a neighboring AI?) must declare on me, and I must refuse offered peace at least once. Are there any other conditions I should know about?

3. After I have satisfied all of your conditions in game, I will post a save. As many people as are interested (including you) will be able to download the save and run as many combats as they like. If the results over many combats as posted by all participants are within the range of expected results based on the hypothesis that the RNG is not biased, then you will publicly disavow your theory.

Do we have a deal ?