Fighting bug

[davew9999]

I'm playing a single player game on Emperor difficult right now and just had a fight, my horsemen with an attack of 2 vs their archers with a defence of 2. So you'd expect me to win this fight around 50% of the time. I've just done this fight 25 times, reloading after each fight, and I lost all 25 times!

Surely this is a bug? Unless the fate of my horsemen was predetermined before the fight? Or is it the difficulty I have it on? As far as I'm aware there isn't anything that can alter the attack and defence rating in a fight that doesn't get shown in the fight scene.

I posted a thread quite a while ago about how a computers knight army seemed to win almost all the time when attacking, even if my defence was much higher. Nobody else seemed to agree with me though and put it down to luck.

Anybody else noticed this?

[danthechan]

the variables of chance are created by the actions leading up to the fight. if you reloaded and did anything different at all the outcome could be different.

[tennispro75]

sometimes it never changes at all, so you'll have to move to a different square, preferably to a hill, which will give you the advantage.

[davew9999]

Oh ok, I guess that explains it!

But that leaves me thinking, if people found out which actions have a positive effect on their chances then they will exploit this.

And another thing, it seems unnecessary for the programmers to write code that alters outcomes of fights based on actions leading up to the fight. If we go by this logic, it is possible that because I found 2 artefacts a couple of turns before the fight this could have caused me to lose the next fight.

It would be much simpler, and make more sense in my view, if the game was programmed to use some simple maths to decide who wins based on the two numbers when the fight is started.

I've just tried a different scenario, I've reloaded the game, ended my turn and then tried the fight. Now their archers defence is on 3. I did this 10 times and lost every time. One thing to note though, the initial 25 fights not one of their archers died, but in these 10 fights exactly 2 of their archers died each time.

[tennispro75]

i had a time where i was trying to take over washington early (only warriors). at certain times, the same battle, with the same moves and such occured. if you cared, i took over the capital and knocked them out ^_^

[SVPM]

I've just tried a different scenario, I've reloaded the game, ended my turn and then tried the fight. Now their archers defence is on 3. I did this 10 times and lost every time. One thing to note though, the initial 25 fights not one of their archers died, but in these 10 fights exactly 2 of their archers died each time.

Move an other unit before attacking, do an other attack it will change the seed used to calculate the results of the battle. You don't necesserly need to end your turn.

[MadDjinn]

everything is determined by a core seed that is used for 'random' things.

when you save the game you save the seed, so if you constantly save/reload expecting different results, it won't happen.

[davew9999]

Thanks for your answers.

Does this seed bias towards the computer on harder difficulty modes? Not just from my example in this post, but generally over the past few months of playing the game it seems like they get an advantage. Haven't noticed this when playing multiplayer though.

[MadDjinn]

there's little bias in most cases. there are random factors, which can allow wins despite the odds for both attack and defense. but that happens for both AI and players. (I lost a 30 vs 18 defense against another player yesterday... and in this weeks GoTW you can win a 4.5 vs 8 attack against the AI to take their capital if you do it on a specific turn with specific actions prior to that turn )

sometimes it really does look like they get some defense advantage with their capital especially the Aztecs and English, but it's likely not there.

at harder levels the AI gets starting boni rather than 'help' with fights. In most MP (or GoTW) the AI is a useless factor that is there mainly to give one or more players another city or tech early. the only saving grace for the AI is in teams where they seem to have bumped up the agression factor so that the AI does *something*.

the seed itself is just a random factor but seems to be more pre-defined than some very basic pseudo-randomization techniques would provide.

it's been that way since the very first Civ/Colonization games, and unfortunately they never got past that useless bit of code. (by the look of the game and other parts of it, it's entirely plausible that it is the exact same code as before)

[elthrasher]

I haven't seen anything to make me think the odds of the fight are anything other than the odds displayed. 30-18 odds means that you have a 30/48 or 5/8 chance of winning each round, so you certainly can lose. I've had outrageous strings of victories and outrageous strings of losses. This will happen when chance is involved.

davew9999, see the GotW threads in the strategy section for more on how to exploit the seed. For regular SP games, it can hardly be termed an exploit, because nobody really cares how you do except you. It's not (as far as I'm aware) possible to exploit the seed in a MP game because there's no way to tell what the seed is and there's no saving/reloading.

Also, build an army and then try.

[MadDjinn]

and since MP games are simultaneous turns, the seed changes constantly.

[Elanza]

It's not (as far as I'm aware) possible to exploit the seed in a MP game because there's no way to tell what the seed is and there's no saving/reloading.

Also, build an army and then try.

In MP mode the seeds are not really existing to begin with.(or should I say practice seeds with SP games and apply them in MP games - there are hundreds of them, so get your pencil and paper ready)

On 1 MP game the Spanish Galleon attacked my defensive mode Galley and lost. He was attack points 3 and I was defense 1.5.

As you can see the chance of winning this battle was 200% for him and losing this battle was 50% to me. It is hard to figure out why such things happen but they do for a reason. It may be related to the fact that I was catching up to the Spanish in techs. So the loss of his Galleon could be related to the fact that my race has yet to meet its demise.

At the end of that MP game he had tanks on my vulnerable city with 40 attack while I had 2 Hoplites with 24 defense. The first time he attacked he won. In the next turn the tanks got wipped. I was amazed with those results and soon after he left the game but after that it froze so abruptly leaving me and Aztecs player waiting for the game to unfreeze.

So I have seen battles in which the odds of winning are 3 to 1 and my attacking units lose that fight. But the most common losses were 2 to 1 while attacking a defended city. I usually attack a city with 2 attacking armies and 1 defensive units. That will increase the chance of winning. <<<< Hint

Most MP games are based on chances and risks factors. You can either wait to build a massive attacking unit to take over a city or risk it all and attack a city wall with an undefended army. Chances are you may retreat and depending on the losses that densive unit in that city may actually retaliate and wipe your wounded army especially Arabs archers with attack of 6 when Elite.

[elthrasher]

In MP mode the seeds are not really existing to begin with.(or should I say practice seeds with SP games and apply them in MP games - there are hundreds of them, so get your pencil and paper ready)


I'm sure you mean millions if not billions.


On 1 MP game the Spanish Galleon attacked my defensive mode Galley and lost. He was attack points 3 and I was defense 1.5.

As you can see the chance of winning this battle was 200% for him and losing this battle was 50% to me. It is hard to figure out why such things happen but they do for a reason. It may be related to the fact that I was catching up to the Spanish in techs. So the loss of his Galleon could be related to the fact that my race has yet to meet its demise.

No. You won because you had a 1 in 3 chance of winning. 1 in 3 means you'll win sometimes. 1/3 of the time, in fact.

At the end of that MP game he had tanks on my vulnerable city with 40 attack while I had 2 Hoplites with 24 defense. The first time he attacked he won. In the next turn the tanks got wipped. I was amazed with those results and soon after he left the game but after that it froze so abruptly leaving me and Aztecs player waiting for the game to unfreeze.

So I have seen battles in which the odds of winning are 3 to 1 and my attacking units lose that fight. But the most common losses were 2 to 1 while attacking a defended city. I usually attack a city with 2 attacking armies and 1 defensive units. That will increase the chance of winning. <<<< Hint

Most MP games are based on chances and risks factors. You can either wait to build a massive attacking unit to take over a city or risk it all and attack a city wall with an undefended army. Chances are you may retreat and depending on the losses that densive unit in that city may actually retaliate and wipe your wounded army especially Arabs archers with attack of 6 when Elite.

Yup, you can't count on the odds always falling your way. The only way to be sure is to send an overwhelming force. Even then, you could lose to the Aztecs.

[davew9999]

It makes me wonder why they decided on using a seed as it's just asking for people to exploit it!

The only benefit I can think of is that it stops being people able to reload a single player game if a fight doesn't go there way. But that isn't much of an issue

I think if the game just decided who was going to win after the fight is started that would make more sense and be easier for the developers.

[Pedal2Metal]

I'm playing a single player game on Emperor difficult right now and just had a fight, my horsemen with an attack of 2 vs their archers with a defence of 2. So you'd expect me to win this fight around 50% of the time. I've just done this fight 25 times, reloading after each fight, and I lost all 25 times!

Surely this is a bug? Unless the fate of my horsemen was predetermined before the fight? Or is it the difficulty I have it on? As far as I'm aware there isn't anything that can alter the attack and defence rating in a fight that doesn't get shown in the fight scene.

I posted a thread quite a while ago about how a computers knight army seemed to win almost all the time when attacking, even if my defence was much higher. Nobody else seemed to agree with me though and put it down to luck.

Anybody else noticed this?

This isn't a bug. The fate of the fight IS determined before the fight occurs BUT it can be changed by a large number of events:
Other fights
Creating a city
Going to next turn
Changing city configurations
Attack order (attack with different unit first)
Building units
etc....

In your case, you simply reloaded the same exact game 25 times & observed the same result 25 times, which is as expected based on how the game works. If you wanted to change the battle outcome, you need to insert another "seed-changing" event (or perhaps more than 1) prior to making the attack. Also movement alone does NOT affect the seed. However, movement that result in another seed-changing event such as introducing you to another civilization, adding gold that causes you to build another unit, etc... can indirectly cause a seed-change.

Basically the game is such that if you play EVERY single turn from the very beginning IDENTICALLY, you will always get an IDENTICAL outcome. However, it's very difficult to play a game IDENTICALLY, although this is basically what you do to get 2900 BC domination victories in Diety GOTW.

It is impossible to exploit this game mechanic in MP as there is no saving/restoring but it is possible to exploit it in SP by saving/restoring as you discovered & simply changing 1 thing that causes the seed to be recalculated. It is also used by GOTW players (myself included) to find the most favorable outcomes to maximize GOTW scores by replaying the GOTW multiple times.

I put a bit of write up in the GOTW thread I think 2-3 weeks back (when it was the Aztecs) that describes some of the events that can impact the seed.
Also it can be challenging to determine when the seed is affected as it can take 3 or more turns before you see a difference.

thanks & best regards,
Pedal2Metal