I've done it

[Balderstrom]

New Random Seed on reload:
After reloading a save:
* The AI's actions may be different (new Random Seed).
* If you do everything the same, the results may be different.

Without "Random Seed on reload", (which can't be used in Multi-Player)
After reloading s save:
* The AI's actions will be the same.
* If you do everything the same, the results will be the same.
* If you change what you do (or the order) the results may be different.

 

[CHEESE!]

My luck is opposite. I won an unfortified, unpromoted axeman to a rifleman.

It said: "combat odds-0%"

I was like w/e and did it.

Next thing- "Your axeman has PWNZORDZ an Indian Rifleman!"
HAHAHAH!

 

[a4phantom]

I can't play Civ until I replace my cd-rom, but yesterday I dropped Chuck Norris with a half empty box of green tea packets.

 

[Everest]

"Your axeman has PWNZORDZ an Indian Rifleman!"
HAHAHAH!

Ghandi told them not to fire.

 

[BakingTheArt]

They were over confident.

 

[NHJ BV]

Just today I lost a battle at 98,8% odds - the highest I've ever lost at. It was a Knight of mine, full health, four promotions* against a damaged Numidian Cavalry or something. That sucked

*then again, my production city is turning out units with four promotions by default, every other turn

 

[Peteyboy]

Balderstorm - I see you mentioned seeding - I thought most things were fairly pre-determined, I had a test a few months back where no matter how I did the combat I would always fail to take the city

It was something like I was attacking against 1 longbowman 1 maceman and 1 ctapult.

I had a Treb, 2 cavalry & 2 riflemen

If I did Rifle or Cav first they'd lose about 70% vs LBM.
If I did treb it would hurt the LBM to 3.6/6
At which point my Cav/Rifle would win (63%)

The problem came the Maceman - Whatever unit i'd used I couldnt use again.

If I used Rifle He'd lose a 65%'er Followed by Cav (99.9%) would lose.
If I used Cav He'd lose a 72%'er Followed by Rifle (99.7%) would lose.
If I had 2 Rifle's left - 2nd would lose a 98.9%
If I had 2 Cav - 2nd would lose a 97.2%

I tried other combo's i.e using various units on LBM first - but every time it would coem down to a unit vs a Maceman at 95% plus and would lose.

Thus it was impossible in that instance for me to take the city - due to the game already seeding the fight against me. I know you can pick random seeding as a choice but that leads to such an easy temptation to cheat (reload until you get the 'correct' outcome)

Concluding - do odd's really matter? Also I'm not sure whether difficulty makes a difference, I wrote down a list of fights in a recent game (on Monarch for what its worth) and found the following

Based on 262 fights (all were pre tanks so no fights involved 2nd attack) - rounded to nearst %

When attacking:

0 - 20% - I won 32% of the time
21- 33% - I won 23% of the time
34 - 50% - I won 56% of the time
51 - 67% - I won 37% of the time
68 - 80% - I won 50% of the time
81 - 95% - I won 71% of the time.
95% + I won 99% of the time (lost a 95.6% and 98.8%)

Clearly it seems attacking is disadvantageous comapred to odds.

Defending

0 - 20% - I won 9% of the time
21 - 33% - I won 22% of the time
34 - 50% - I won 40% of the time
51 - 66% - I won 73% of the time
67 - 80% - I won 72% of the time (Curious I won these less then the above)
81 - 95% - I won 80% of the time
95%+ - I won 87% of the time (a very disturbing fact)

In these fights I ignored fights which involved withdrawing of units like cavalry - but including siege units as them withdrawing = them winning.

Overall statistics showed over all 262 fights I should have won 58% of tbe fights - and won 47%

What other factors ar ein place to determine fights anyway?

 

[Balderstrom]

@ Peteyboy, what you need to do - if you are trying for a different/new result, is go off to another part of your empire and move a unit (somewhere else) and/or do some other actions, like perform a trade.

If you keep trying the same combat - but just with different units ... then the value (seed) that is used to Roll the Dice will be the same every time. Now sometimes it is possible to change the win/loss of a combat by just changing which unit you use - BUT if you've already lost, and already used your best unit (best odds), then none of the other units (which have worse odds) have any chance whatsoever.

These days I just take my losses and continue. Though I think I am going to stop playing Huge maps and large numbers of civs - those games just take far too long.

 

[Peteyboy]

Is it really as uncontrolled as using other units elsewhere - I haven't reloaded because of lost fights for a long time, just accepted that sometimes things don't go as planned, but somehow the dice being fixed as such seems a little annoying - but to be honest, if those fights are going to be critical it means something else is going wrong somewhere!

Hehe the old saying like it or lump it comes into mind. I did manage to win a 0.4% the other day so I can't complain

 

[say1988]

I don't think it is that simple, I think you have to do something that requires the computer to make a decision, such as hitting the explore button. Anything that will use a number from the seed.

Also, for your results you only have, on average, 19 per range, for each of you 14 ranges.
IF you had 20 battles in 95+ [defending] you went ~17 (actually 85%), if you won one more you would have won 90%. That is best case, odds are you had less than 20 battles in the section. You have a very small sample size when calculating something like this. And, IIRC, you also have to take into account that odds aren't really combat odds, but strength ratios, so damaged units actually have worse odds that actually shown [as they will take fewer lost rounds in order to be killed], and I am unsure if first strikes are included in the combat odds calculator.

 

[Kiwi Tyrant]

Obviously your Riflemens gunpowder got moist when you marched past that river. By the time they realised this, the spearmen were stabbing them to death while they tried to fix their bayonets! The computer (caught off-guard)just could'nt reconfigure the numbers in time to show the combat odds were actually 100% chance of losing.....

Also, if the dopey Persians allowed their archers to rain down hell (thus making the Spartans cower under their shields) and simultaneously rushed in their numberless slave fodder, would'nt that efficently deal with their formation so they could quickly move on?
If they can ship across elephants, rhino and huge thrones; then why did'nt they bring some catapults and simply smash the Spartan formation to bits where it stood?
.....then again, would'nt make for an interesting, cool movie though!

 

[Eclipse-X010]

-The size of the Persian army was extremely exaggerated.

-What do you expect, they used archers on the offensive when they clearly had iron and bronze to build swordsmen/axemen.

-Strange how the Spartans fight with spears when there aren't very many Persians with horses.

I've lost a few battles with a 99.9% chance of victory but this might have been 100% in vanilla.

I lost 2 95% battles in a row, very disturbing. (2 German early infantry vs. regular infantry [15 vs. 5 health plus defenders had hill defense.) finally I just sent in the tanks.

EDIT: Looked at the minimap that showed you only had one city. I'm going to assume either
a) You're playing one city challenge (PLEASE tell me your playing one city challenge)
b) Sury captured all of your cities
c) You are very bad at expanding

OF course no missionary came your way... you only have ONE city.

 

[Peepers]

a) You're playing one city challenge (PLEASE tell me your playing one city challenge)

See post#4

 

[DeadShot]

Lost a 99.9% once, my tank vs some rubbish unit of theirs... But my second avenged the 1st's death.... It's just proves it's better to be born lucky!

 

[onomastikon]

I have written a combat simulation program which gives 0.00076 probability of the defender winning, as it did, in round 12. What did the damage was, obviously, the string of defender hits at 0.285 probablity each: the probability of such a string occurring is 0.000544.
Such results are to be expected occasionally.

Are you saying it's 0.00076 percent or 0.00076 chance?

76 in 10,000 isn't all that astronomical, after all. Your chances of winning a smallish prize in the lottery (in Germany, at least) are much, much worse.
7 in 1000 is actually quite a bit, considering that if you play half as much civ as half the people on these boards, you are bound to have 1000 battles in half a week's time. Half of those will be halfarsed, but 7 of them might very well show a suicide squad of knife-bearing but well-dresed lunatics who manage to hide inside houses of some town, lurking, pretending to be civilians, so that when the enemy riflemen come, secure in their feeling of superiority, they feel (or rather: fail to) the cold blade of steel enter their jugular in the wee hours after midnight when they are asleep in their tents, drunk on the spoils of what they thought was a victorious war.
Statistics can be dramatic.

 

[Cytral]

hehehe

drunk rifleman

 

[Civsassin]

In my current game as Shaka, I went against a Greek Spearman with 0.0 hit points. I didn't think that was possible either...

 

[say1988]

It is possible that it was something like 0.02 and rounded down. I believe hit points are calculated to more than one decimal place, but not sure.

 

[Kranden]

I once lost a 99.9 I was so embarassed

 

[CivOtaku]

That's amazing! The highest i have ever lost is probably somewhere around 70!