The AI CHEATS! (Admitted to by Sid) lol

r_rolo1 said:

And to be honest, it would not give a better game experience if you are in the down side of the stick A RL example: once I had the luck of seeing a warrior of mine beating the vedic aryans ( 4 barb archers, since it was a standart map ) in BtS. The odds of that happening are astronomically low, but they are far higher than the ones of the system you are advocating ( 0,0% ), and if would be a game over situation, period.

Click to expand...

Well there may still be a chance with such a strong unit favoring system, that the defender will survive every attack, even though the strong unit doesn't die.

But it's easier to suspend disbelief of with spearmen always 100% loose to a tank, then to have 99% odds, but the lucky spear men happens to overcome a crucial offensive.

So Sid is not going to go back on his realization that player intuition matters more than mathematics. There will be no spearmen beating tanks in civ 5. At least not in a single turn.

Souron said:

But it's easier to suspend disbelief of with spearmen always 100% loose to a tank, then to have 99% odds, but the lucky spear men happens to overcome a crucial offensive.

Click to expand...

To be honest, I find the oposite far more believable. C'mon, you should be old enough to have seen a couple Apache helicopters in Iraq during the 2003 offensive being taken down by a combination of bad luck, bad weather and some rather archaic gunpowder weapons... i also remember the that East Timor guerillas in the the time of the Indonesian ocupation using effectively as mock up muskets bamboo filled with gunpowder and nails ... medieval stuff at best ( and they won ). 100 % odds of something is simply not real. 100% odds are Hollywood stuff at best ( not even them do that as much nowadays ).

And to add, like I pointed before, treating a less than 100% odds event like a certainty introduces errors in the game sequence probablilities that do not come from the combat engine in it self: the odds of winning 20 97% battles are of 54% , meaning that there is almost a 50% error if you treat those 97% battle as 100% ones. that, by any standart, is a lot . This things would cascade a lot, especially if you ( or a AI ) started in a high tech unit rampage ( it is the diference between a rambo styled action sequence with a pile of enemy corpses and a alive hero , and the unit losing after killing 20 of their weaker foes by just bad luck ).

I can live with a combat engine that makes odds of 90-99,99% extremely unlikely, that would be far more than enough to kill the ( and to be honest, it is pretty much dead in Civ IV as it is now ). I'm just against getting a 98% odds combat and treat it like it was 100% .

Another proof positive of importing worthless ideas from CivRev, like some of us have said all along. I'm very disappointed in this - not just because it's silly and childish to lie about combat odds, but because odds with vastly different outcomes still exist. Sure, maybe a 99% battle is still won, but if at 90% a player can lose and have the opponent take no damage, it's still just as bad. Since they were already committing to 1upt, not destroying units after every combat, and so on, they really should have implemented a simpler system with more straightforward combat outcomes. Just let all results fall in a reasonable range - no "winning and taking no damage" at 2% odds and no auto-wins either.

r_rolo1 said:

C'mon, you should be old enough to have seen Avatar.

Click to expand...

Fixed.

Earthling said:

Another proof positive of importing worthless ideas from CivRev, like some of us have said all along. I'm very disappointed in this - not just because it's silly and childish to lie about combat odds, but because odds with vastly different outcomes still exist. Sure, maybe a 99% battle is still won, but if at 90% a player can lose and have the opponent take no damage, it's still just as bad. Since they were already committing to 1upt, not destroying units after every combat, and so on, they really should have implemented a simpler system with more straightforward combat outcomes. Just let all results fall in a reasonable range - no "winning and taking no damage" at 2% odds and no auto-wins either.

Click to expand...

Wait, so you're disappointed with the change in handling combat odds. I got that much. Which would have to mean you like the CIV way of doing things? But you can't like that because you said it's bad if you can lose a 90% victory prediction and not deal any damage? So then you also say you want a simpler system with more straightforward outcomes, but isn't that what you just said at the beginning you didn't like (since that's what CiV does)?

I am so confused.

Fun fact:

Infinite Interactive (Puzzle Quest series) found that, for players to perceive the game to be fair, need approximately twice as many Good Things happening to them than the opponent.

Just sayin'.

For fun I played a Civ I game the other day. Civ I is more than happy to let my V battleship die to a phalanx. I lost at least two. That is a 27 to 3 ratio given that the spearmen were fortified. I should lose 1 in 9. This makes the odds sit at 91%. Pretty sure though most everyone will be furious to lose this battle.

Iberian said:

For fun I played a Civ I game the other day. Civ I is more than happy to let my V battleship die to a phalanx. I lost at least two. That is a 27 to 3 ratio given that the spearmen were fortified. I should lose 1 in 9. This makes the odds sit at 91%. Pretty sure though most everyone will be furious to lose this battle.

Click to expand...

That's not how the odds work though. Being twice as strong as a unit doesn't give you a 66% chance of winning.

In civ 4, 27 strength vs 3 would have a 99% chance of winning, at least.

Im glad they gotten rid of the percentage being shown in combat, intellectually i realise you can't win every 75% chance battle, but from a gameplay point of view i honestly don't see why i shouln't expect too, so i always reload when a high % fights turns out badly.

Schuesseled said:

Im glad they gotten rid of the percentage being shown in combat, intellectually i realise you can't win every 75% chance battle, but from a gameplay point of view i honestly don't see why i shouln't expect too, so i always reload when a high % fights turns out badly.

Click to expand...

This is a little bit ridiculous. If you reload when you lose a 75%, you must be reloading all the time, which means you haven't really ever won a true game of civ.

The reason why you shouldn't expect to win those, is because the AI doesn't always win them, it makes for a fair playing field where there's always a chance that you get really lucky or unlucky.

This is exactly why I had to turn off the "new random seed on reload" option.

Lyoncet said:

Wait, so you're disappointed with the change in handling combat odds. I got that much. Which would have to mean you like the CIV way of doing things? But you can't like that because you said it's bad if you can lose a 90% victory prediction and not deal any damage? So then you also say you want a simpler system with more straightforward outcomes, but isn't that what you just said at the beginning you didn't like (since that's what CiV does)?

I am so confused.

Click to expand...

As it stands, the civ5 system appears to just be about the same as civ4 but worse, there's the place we start. The civ4 way was all right, it is true I was not a crazy fan about it - the very worst was a lot of counterintuitive micromanagement. The difference between 6 strength and 5.9 strength (eg. collateral damage) could be like a 20% difference in combat odds, plus we had weird promotion systems and so on. Then, we have the complaint they are trying to address here is still there - people just don't like losing 99% battles.

So in short, civ5 has ZERO indications of actually being a simpler system with straightforward outcomes. All they did was lie about the odds past a certain point, which helps nothing except for people who really don't get math I guess. There's no evidence to suggest that at, say 80% odds there is less variance than before. If you can win 99% guaranteed, fine, but if at a reasonable 80% battle you could still have an invulnerable victory or crushing defeat the system is just as random (and annoying in the same way to players) as always.

A simpler system, someone else said something similar in this thread and I and others have said elsewhere, would translate combat odds into a reasonable range of outcomes. No multiple rounds of combat, nothing decided entirely by RNG rolls where the 1 in 50 chance crops up to much frustration. Units not having to die is a good thing with the rest of the combat and war system at least. And if they have clear hitpoint totals, creating the range of outcomes would have been very easy. So then a battle would just result in "unit 1 takes a-b damage, unit 2 takes x-y damage" determined by final strength analysis. The system where any unit could win with no damage or be defeated by random chance is what should go, and it doesn't appear it has.

So if they weren't going to a better system, at the least they could just leave it like civ4 and be honest about how random the game is.

In fact, hiding things from the player (real combat calculations, diplomacy, etc...) is fast looking to be one of the very worst and most frustrating decisions made in a civ game. These capabilities were almost universally praised when added in successive expansions/patches of civ4 or previous games. When combat, or an AI civilization, or barbarians or city states or so on, is not only going to act very randomly but the chance of even knowing accurate odds/chances is obscured, that's just a recipe for annoying the player.

That's not how the odds work though. Being twice as strong as a unit doesn't give you a 66% chance of winning.

Click to expand...

It did in Civ1.

Civ1's combat calculator was very simple; no damage, the attacker either killed the defender or died.
And the probability of killing the defender was [Attacker strength]/[Attacker strength + defender strength]

Souron said:

Well there may still be a chance with such a strong unit favoring system, that the defender will survive every attack, even though the strong unit doesn't die.

But it's easier to suspend disbelief of with spearmen always 100% loose to a tank, then to have 99% odds, but the lucky spear men happens to overcome a crucial offensive.

So Sid is not going to go back on his realization that player intuition matters more than mathematics. There will be no spearmen beating tanks in civ 5. At least not in a single turn.

Click to expand...

My goodness, finally. Thank you! In one post you've perfectly described my way of thinking, which took me several posts to explain and people still didn't get it!
...Probably because I'm not a native speaker and I'm unable to express myself clearly.
It seems you've seen the footage from the conference and get what's this about. Things like Schuesseled described above are exactly what Sid wants to avoid, because it ruins the experience of an Epic Journey.

Don't worry folks, Mr Meier knows what he's doing. Regardless how dumbed-down you might think it'll be, when it'll come to playing on Monarch+ you'll find that losing is still a perfectly valid option ;P

P.S. And don't get condescending on Schuesseled because of reloading. We can talk all day long about power of will, noobness and whatever but it's a game, not a life. I myself several times when lost Great General or lost Liberalism race thought to myself " f*** it, I'm reloading". Even the great Sisiutil himself worldbuilded barbarian city in one of his games, and then reloaded whole lot to basically, prevent barbs from settling his spot.

By eliminating incentive to reloading thousands of players won't be tempted to go for it, and everybody's happy

Earthling said:

As it stands, the civ5 is the civ4 but worse, there's the place we start. The civ4 way was all right, it is true I was not a crazy fan about it - the very worst was a lot of counterintuitive micromanagement. The difference between 6 strength and 5.9 strength (eg. collateral damage) could be like a 20% difference in combat odds. Then, the complaint they are trying to address here is still there - people just don't like losing 99% battles.

So in short, civ5 has ZERO indications of actually being a simpler system with straightforward outcomes. All they did was lie about the odds past a certain point, which helps nothing except for people who really don't get math I guess. There's no evidence to suggest that at, say 80% odds there is less variance than before. If you can win 99% guaranteed, fine, but if at a reasonable 80% battle you could still have an invulnerable victory or crushing defeat the system is just as random (and annoying to players) as always.

A simpler system, someone else said something similar in this thread, would translate combat odds into a range of outcomes. Units not having to die is a good thing with the rest of the combat and war system at least. And if they have clear hitpoint totals, creating the range of outcomes would have been very easy. So then a battle would just result in "unit 1 takes a-b damage, unit 2 takes x-y damage" determined by final strength analysis. The system where any unit could win with no damage or be defeated by random chance is what should go. Or they should just leave it like civ4 and be honest about how random the game is.

Click to expand...

So you're arguing that you should be given a range of damage that a unit may inflict and a range of damage that unit may take before attacking. I assume you want a relatively narrow range since you didn't like the "randomness" of civ 4. Am I correct?

Earthling said:

So in short, civ5 has ZERO indications of actually being a simpler system with straightforward outcomes. All they did was lie about the odds past a certain point, which helps nothing except for people who really don't get math I guess.

...

The system where any unit could win with no damage or be defeated by random chance is what should go. Or they should just leave it like civ4 and be honest about how random the game is.

Click to expand...

I'm having such a hard time reconciling these two statements. It reads as if statement 1 means "it's bad to make a unit that would have had a 99% chance of success really just have a 100% chance," but statement 2 seems to mean "units with a 99% chance of success shouldn't fail."

I also don't get how CiV's system not being "a simpler system with straightforward outcomes." It seems simpler to me (not as detailed/precise, but that's a very different notion); instead of having a 55% chance of victory, you' have a slight advantage. Check simplicity. And for straightforward outcomes, I don't see what's not straightforward about "total victory" meaning "total victory" etc.

I still remember losing to a diplomat with my battleship in Civ I.
Considering that diplomats had 0 defense I'm still mystified how that even happened.

Guardian_PL said:

My goodness, finally. Thank you! In one post you've perfectly described my way of thinking, which took me several posts to explain and people still didn't get it!
...Probably because I'm not a native speaker and I'm unable to express myself clearly.
It seems you've seen the footage from the conference and get what's this about. Things like Schuesseled described above are exactly what Sid wants to avoid, because it ruins the experience of an Epic Journey.

Don't worry folks, Mr Meier knows what he's doing. Regardless how dumbed-down you might think it'll be, when it'll come to playing on Monarch+ you'll find that losing is still a perfectly valid option ;P

P.S. And don't get condescending on Schuesseled because of reloading. We can talk all day long about power of will, noobness and whatever but it's a game, not a life. I myself several times when lost Great General or lost Liberalism race thought to myself " f*** it, I'm reloading". Even the great Sisiutil himself worldbuilded barbarian city in one of his games, and then reloaded whole lot to basically, prevent barbs from settling his spot.

By eliminating incentive to reloading thousands of players won't be tempted to go for it, and everybody's happy

Click to expand...

Ok ...

I understand why Sid wants to do this, but that does not mean I agree with him. People in general do not understand odds and, because of that, they tend to be pissed when what they expect to be a certain does not happen ( OTOH they don't get half as happy when they got a unexpected luck out from a hopeless situation... ), and he wants to sell a game to those people.

I can relate to that, but I do not agree with the way Firaxis wants to implement that. There is another group of players that actually understand odds and/or doesn't want the game treating stuff unequal as equal ( I'm pretty sure you know TMIT ) and this implementations will surely piss a lot of people as well. The worst is that it would be relatively easy to make a combat engine that would produce very little of those 90-99,99 % odds ( in spite of there being people that think they should win all the battles with 50%+ odds ) that would not piss anyone in this regard. But they probably simply lazied it out ( no criticism, laziness is a major driving force in the history of mankind and probably the biggest responsible for civilization as we know it ) and decided to pretend that 4$98 =5$00 without thinking it to the end.

P.S Regarding trusting Sid .... well, it was him that introduced the in the first place, wasn't it ?

I strongly agree with the argument that just because some people can't understand probability is no reason to mess the game up for those of us who can.

I still remember losing to a diplomat with my battleship in Civ I.
Considering that diplomats had 0 defense I'm still mystified how that even happened.

Click to expand...

It must've been a veteran diplomat

Souron said:

@Tarkhan
The problem with what you describe is making good AI for the highest levels in the first place.
Ideally, with balanced odds, and 8 civs on the map, the AI should win 7 out of 8 games, against an average player. But this has never been achieved. Most players play at higher levels, and expect to win fairly often. So any time that would be spent on making the AI dumber on low levels, would be better spent trying to make the AI smarter, so that it could win more often with less of an advantage.

Click to expand...

While that may sound reasonable, I'm pretty sure that what you want is that the average player playing on average AI settings, should win 7 out 8, not lose that many.

Why? Because 'average' still means that you win most of the time. People would get pretty angry if it was not so.