That article is about psychology, not AI. Absolutely no mention is given to the difficulty of the game, simply the statistics on display. 'Sides, everyone knows the AI cheats (what this has to do with the RNG is anybody's guess).
The AI CHEATS! (Admitted to by Sid) lol
- Thread starter Dearmad
- Start date
That article is about psychology, not AI. Absolutely no mention is given to the difficulty of the game, simply the statistics on display. 'Sides, everyone knows the AI cheats (what this has to do with the RNG is anybody's guess).
Guardian_PL
Emperor
- Joined
- Nov 4, 2006
- Messages
- 1,231
People, how about you just watch the damned thing and THEN go on a nerd rage?
Yes, it's an hour long. But it's full of Sid and I didn't even noticed when it was finished. That man is pure wisdom and experience. And I agree with him on all counts.
All what was done in CivRev and Civ5 is to eliminate possibility of losing combat when having 95-99% chance to win. Why? Go and ask bazillion of players (including myself) that had 97.5% chance to win and lost three times in a row in CIV. It's frustrating. Very. To the point that a lot of people (including myself) reach for a earlier save or quits, and that is the one thing Sid wants to avoid.
Games are digital, zeros and ones, yes' and noes. But that's not how human mind or world operates .
That's why we now have this fascinating combat system when a single unit has to be attacked several times to die, and "Total Victory" means "you're pretty much safe" in combat outcome.
I could go on and on about it, but ignorance is always blind (I'm calling such behaviour "living inside your own head, in your little world"), while curiosity will lead to knowledge regardless of what I'll say
Yes, it's an hour long. But it's full of Sid and I didn't even noticed when it was finished. That man is pure wisdom and experience. And I agree with him on all counts.
All what was done in CivRev and Civ5 is to eliminate possibility of losing combat when having 95-99% chance to win. Why? Go and ask bazillion of players (including myself) that had 97.5% chance to win and lost three times in a row in CIV. It's frustrating. Very. To the point that a lot of people (including myself) reach for a earlier save or quits, and that is the one thing Sid wants to avoid.
Games are digital, zeros and ones, yes' and noes. But that's not how human mind or world operates .
That's why we now have this fascinating combat system when a single unit has to be attacked several times to die, and "Total Victory" means "you're pretty much safe" in combat outcome.
I could go on and on about it, but ignorance is always blind (I'm calling such behaviour "living inside your own head, in your little world"), while curiosity will lead to knowledge regardless of what I'll say
I don't know anything about programming, but from my experience playing video games, I think it's safe to say it's not as ridiculous as you make it out to be.
I'll give an example: Company of Heroes, a real-time strategy/tactics game, was released with adequate AI. However, once you went past normal difficulty, the AI immediately had to cheat by getting production bonuses and health buffs on its units. Then a community-made mod called Eastern Front was released, and one of the changes was that the AI was significantly improved even on Easy and Normal. The AI imitated human strategies to the point that sometimes I felt like I was playing online. If a bunch of non-professional programmers can do it, why can't the original developers do it (or at least add a patch later)?
ZimbuTheMonkey
King
People, how about you just watch the damned thing and THEN go on a nerd rage?
Yes, it's an hour long. But it's full of Sid and I didn't even noticed when it was finished. That man is pure wisdom and experience. And I agree with him on all counts.
All what was done in CivRev and Civ5 is to eliminate possibility of losing combat when having 95-99% chance to win. Why? Go and ask bazillion of players (including myself) that had 97.5% chance to win and lost three times in a row in CIV. It's frustrating. Very. To the point that a lot of people (including myself) reach for a earlier save or quits, and that is the one thing Sid wants to avoid.
Games are digital, zeros and ones, yes' and noes. But that's not how human mind or world operates .
That's why we now have this fascinating combat system when a single unit has to be attacked several times to die, and "Total Victory" means "you're pretty much safe" in combat outcome.
I could go on and on about it, but ignorance is always blind (I'm calling such behaviour "living inside your own head, in your little world"), while curiosity will lead to knowledge regardless of what I'll say
Now while I agree that losing 3 battles in a row for which you have a 97.5% to win could be VERY, VERY frustrating. I just don't want a developer messing with randomness to make it appear "more" or "less" random just for the sake of player psychology.
There are so many unknown variables in a strategy game; where the resources are, what the AI is going to do, where the AI is, etc. I just want the math portion of it to be left alone, I want it to be reliable.
So, maybe I might not notice this subtle change (eliminating randomness to make sure there are no astronomically unlikely unlucky or lucky streaks), I just don't feel right about it. Variance will ALWAYS be a part of any game that has a random element and that's totally okay with me because as often as you will get unlucky streaks, you will also get lucky streaks.
TL;DR: Random is random, the game isn't out to get you.
Guardian_PL
Emperor
- Joined
- Nov 4, 2006
- Messages
- 1,231
Yes. But you'd be surprised how paranoid people can be, a lot of my fellow gamers stated several times that the moment they had situation under control some random event has screwed everything up, or when they rushed an AI's capitol all of a sudden couple of Archers stopped eight units or sth. And many, many times I've heard or even felt myself that instead of having square odds (random is random) we have a "let's keep the player busy else he'll get bored" approach, and, yes, "The AI is out to get me".
Sid is right by saying that gamers are paranoid and delusional, and all that matters is that Epic Journey (at the end).
Think about it this way - we players are heroes of the story (think like Conan the Barbarian fantasy tale). We are the main character. Now, in how many books main hero dies right on the beginning because of Barbarian Uprising in 3000BC? How successful and popular such books would be? Throughout the whole story our heroes are facing dreadful challenges and guess what - they overcome them and everybody's happy, regardless of laws of probability.
So what Sid did is to ensure that this story will continue unless we'll repeatedly do something stupid. And even then we'll get explanation why we'd lost to get an incentive to not make the same mistake "the next time", which ensures replayability.
It's not dumbed down at all, that's what difficulty levels are for. I'm sure that from Emperor+ if you won't play carefully you'll be screwed regardless of the fact that when you have 3:1 odds you'll lose less times than you normally should, because the AI will have much, much bigger army than you
Again - this is not only zero-one yes-no relation, however random seed has been done I'm sure you'll be able to understand the exact mechanics behind it at some point, what with PieceOfMind already stripping down the combat chances even before the release
r_rolo1
King of myself
Well, that, for all it matters, is cheating
If they had made a combat function that made the 95+% odds battles extremely unlikely to avoid that enraging result ( btw , anyone noticed that this also erases the 5-% lucky wins ? 
), I would understand ( it is perfectly possible to do that). But saying that a battle has 97% odds and then erase the possibility of losing at those odds is atleast twisting hard the meaning of the word odds ... let's cut the politeness, it is a lie, pure and simple. Twist it in the way you like, 97% !=100% and if the interface says 97% when it means 100% , you can argue that the game is cheating, no?Guardian_PL
Emperor
- Joined
- Nov 4, 2006
- Messages
- 1,231
Well, that, for all it matters, is cheating
That's again thinking in Civ4 terms.
In Civ5 interface doesn't show percentage chance for win. You've got major, decisive and total victory outcomes, each with different repercussions, on top of that we no longer have at least one unit dying as a result of combat, unless there's a huge difference in strength and/or health of units. If you see a "Total Victory" in battle outcome you know you'll win, no need for percentages.
So there is no lie, at least in my opinion
r_rolo1
King of myself
You spoke of CivRev as well And , who knows if they will not add the battle odds again? Civ IV didn't had battle odds as well when it came out, just a breakdown of the bonuses of both parts and a vs of the modified strengths of both parts... And someone wisely at that time said that people would start complaining about losing at high odds just because it was shown as percentage
But again, if the game computes that the output has a 97% of happening and then it is trunked to 100% just because ( and this regardless of the program level where it happens ), it is cheating. That is a lot diferent of making a straight function that avoids giving results with 97% odds unlikely...
P.S The breakdown on major, decisive, and total victory makes no diference in this argument. It is just interface labeling. I'm arguing about the core concept of probability. Things that have less than 100% chance of happening mean that there is a chance of more than 0% of that thing not happening... That is the exact equivalent of trying to pay a 5$ bill with 4,98$
Worse , this things get cumulative.... the diference between the odds of 20 97% wins and 20 100% wins, for a example, is of 46% . Treating both as equal is making a almost half mistake ....and skewes things to the side of the more advanced troops more than it comes out of the combat engine.
And , who knows if they will not add the battle odds again? Civ IV didn't had battle odds as well when it came out, just a breakdown of the bonuses of both parts and a vs of the modified strengths of both parts... And someone wisely at that time said that people would start complaining about losing at high odds just because it was shown as percentage But again, if the game computes that the output has a 97% of happening and then it is trunked to 100% just because ( and this regardless of the program level where it happens ), it is cheating. That is a lot diferent of making a straight function that avoids giving results with 97% odds unlikely...
P.S The breakdown on major, decisive, and total victory makes no diference in this argument. It is just interface labeling. I'm arguing about the core concept of probability. Things that have less than 100% chance of happening mean that there is a chance of more than 0% of that thing not happening... That is the exact equivalent of trying to pay a 5$ bill with 4,98$
Worse , this things get cumulative.... the diference between the odds of 20 97% wins and 20 100% wins, for a example, is of 46% . Treating both as equal is making a almost half mistake ....and skewes things to the side of the more advanced troops more than it comes out of the combat engine.
Taking the content of this video, with screen shots and previous videos that show odds before the end of the battle, it seems safe to conclude that the battles will indeed be predictable before battle. If it says "decisive victory", then you know it's going to be a decisive victory before the battle ends.
Centurio
Warlord
I don't agree. From what I've seen so far, there have been battles that have differed from the expected outcome though not significantly (but remember absence of evidence is not evidence of absence and it would seem this is particularly true when considering screenshots and footage of an unfinished and unreleased game). Decisive victory and decisive defeat may be the exception, but for example a stalemate may result in an outcome that looks more like a minor defeat, or a minor defeat may turn out as more of a major defeat.
Of course, my disagreement here depends on what exactly you meant by "predictable". I mean, one of the simplest ways to define randomness is "that which can't be predicted".
Guardian_PL
Emperor
- Joined
- Nov 4, 2006
- Messages
- 1,231
I really see no problem if the combat system work like this:
50% - stalemate, both sides receive losses
60-75% decisive victory - loser gets losses, winner gets less damage
75%-90% major victory - loser has heavy casaulties, winner gets little or moderate damage
90%-100% - total victory - loser gets insta-killed or heavy losses, winner gets a scratch or nothing at all
In setup like that all that Civ4 percentage rating is not important anymore. Now, providing we have a similar mechanic like that "Rugged Defense" from Panzer General that Brawndo mentioned in civ5 Combat and Odds thread everything is dandy to me
50% - stalemate, both sides receive losses
60-75% decisive victory - loser gets losses, winner gets less damage
75%-90% major victory - loser has heavy casaulties, winner gets little or moderate damage
90%-100% - total victory - loser gets insta-killed or heavy losses, winner gets a scratch or nothing at all
In setup like that all that Civ4 percentage rating is not important anymore. Now, providing we have a similar mechanic like that "Rugged Defense" from Panzer General that Brawndo mentioned in civ5 Combat and Odds thread everything is dandy to me
Guardian_PL, when you write "75%-90%" what is that a percentage of? It seems contradictory that you say the percentage rating is not important anymore yet the very categories that are presented to the player are based on the percentage rating (whatever the percentage is you're measuring).
Right, there's some play in each category, so that the expected results may seem similar of the other two nearby categories, but in civ IV it was possible to have a unit die with >99% odds. It seems that Firexis has since concluded that it's better to guarantee a win. And not just for 99% odds, but for 75% odds.
Except it's not odds, but strength ratios.
On a related note, I speculate that part of the reason that the perception of odds differs from actual odds is that humans naturally think on a logarithmic scale. So a unit with 2 fighting 1 is perceived as close to 2:1 odds. A unit 20 fighting a unit 10 is perceived as 2^20:2^10 ~= 1000:1.
Schuesseled
Deity
- Joined
- Mar 31, 2008
- Messages
- 2,081
what sid was talking about is that no one ever likes to have a unit die if the percentage chance of victory was over about 60%, and you have to keep the customers happy so igf your units never die when the percentage is above 60% in this game, then sid will have achieved this in Civ V.
Guardian_PL
Emperor
- Joined
- Nov 4, 2006
- Messages
- 1,231
Sorry PieceOfMind, that's what's happening when one tries to post from work. 
All these percentages and stuff I've posted above have nothing to do with the truth, it's just an example of how can it work without splitting the hair between 2-3% difference for chance to win.
I was merely trying to point out that instead of showing to the player a difference between 95% to win or 97.5% (like it was in Civ4) we now have these victory/defeat types, and in my opinion if the difference in strenthgs between combatants classifies as Total Victory it doesn't matter anymore whether you would have 95% to win or 97.5% to win, because it's all within the same victory type class (Total Victory).
Therefore I don't think there is need to go for "omg the lying bastards" approach, since it's been done differently now and I think it'll be a much better experience than in previous Civ installments. Sorry for confusions, all these percentages are just an example - after all we don't know yet how it is done in Civ5.
Good stuff there's not many customers today
@Schuesseled
I still don't think that's the case, I think that unexpected losses will still happen, but like in Panzer General you won't necessarily have to see your hard-produced unit die, only get properly mutilated on a wimpy defender.
But it's different to a player to see "oh, tough bastards, these Riflemen not only didn't die, but took half of health from my Armour unit" (Civ5) then "omg I had 97.5% chance to win and my Swordsmen unit is gone while their Archers are not even scratched WTF!?" (Civ4).
What it's really going to be like we'll see in about a month
All these percentages and stuff I've posted above have nothing to do with the truth, it's just an example of how can it work without splitting the hair between 2-3% difference for chance to win.
I was merely trying to point out that instead of showing to the player a difference between 95% to win or 97.5% (like it was in Civ4) we now have these victory/defeat types, and in my opinion if the difference in strenthgs between combatants classifies as Total Victory it doesn't matter anymore whether you would have 95% to win or 97.5% to win, because it's all within the same victory type class (Total Victory).
Therefore I don't think there is need to go for "omg the lying bastards" approach, since it's been done differently now and I think it'll be a much better experience than in previous Civ installments. Sorry for confusions, all these percentages are just an example - after all we don't know yet how it is done in Civ5.
Good stuff there's not many customers today
@Schuesseled
I still don't think that's the case, I think that unexpected losses will still happen, but like in Panzer General you won't necessarily have to see your hard-produced unit die, only get properly mutilated on a wimpy defender.
But it's different to a player to see "oh, tough bastards, these Riflemen not only didn't die, but took half of health from my Armour unit" (Civ5) then "omg I had 97.5% chance to win and my Swordsmen unit is gone while their Archers are not even scratched WTF!?" (Civ4).
What it's really going to be like we'll see in about a month
It's my understanding they did indeed do this for civrev (I haven't been able to look at the code so I can't know for sure
) but I'm not convinced they've gone entirely down that route with civ5. It looks like they've tried to streamline the combat odds interface and have changed the combat mechanics such that major upsets might be rarer, but I don't they've gone down the "cheat for the player" mentality of fudging combat when the combat odds are good just so the player wins the battle.Such code would presumably be there for all to see when we get given the SDK. Even the person who was tasked to write that code would probably be thinking, "Oh, are we taking this too far, protecting the fragile player?"
Anyway, in short, I'm still expecting that Firaxis have left in occasional major upsets for combat outcomes. If combat becomes more predictable, it will become less challenging and probably less interesting.
r_rolo1
King of myself
That is because they are odds , duh. If they are odds, they aren't warranties of any kind. If there is a warranty of result, there aren't odds around. Simple as that.
@Guardian_PL
IMHO you are misunderstanding things. Things can't simply be like that because it would actually defeat the propose of having a combat
"Oh I have a strong enough unit, so I'll just go around and do some Rambo style skull bashing" would be the result if no actuall odds were introduced in the game, just he kind of thing you are advocating. If the labels are actually expected results ( a non-numerical version of Civ IV odds ), things are Ok, but delivering the result of the battle before the battle is denying the need for battle 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. @ PoM
I agree with you, most likely there will not be assured results and we will still see the ocasional
... because that is needed if you have actually a combat round involving the RNG It even does a more interesting ( and believable ) game: this is not a Doom for heavens sake, where the hero is supposed to kill every single moving thing without getting a scratch ever
Another thing. It's funny from my point of view that Sid was so fascinated by how players handled situations of 2:1 odds and 20:10 odds being somehow different. His solution was to let the player win the battles he thought he was supposed to win. This was in civrev where there was no proper combat odds interface - just a ratio of strengths.
But in civ4, the whole point of introducing the odds display was to remove/reduce the confusion of how the ratio of strengths played on combat outcomes. If people before were to incorrectly believe that 2:1 was not as good as 20:10, then they wouldn't have taken long to learn it was wrong when they experienced both types of combat with the actual odds display informing them of their chances.
Civrev tried to hide too much from the player IMO. Players still learn quickly that RNG can affect combat quite a bit, and they can get a feel for how much the combat odds depend on the strength ratios, but the interface fails to ever present that information to the player.
From what we've seen of civ5, they've tried to find a middle ground. They still want to give to give the player meaningful predictions of outcomes but they're trying to make it easier to quickly decide whether to go ahead or not.
But in civ4, the whole point of introducing the odds display was to remove/reduce the confusion of how the ratio of strengths played on combat outcomes. If people before were to incorrectly believe that 2:1 was not as good as 20:10, then they wouldn't have taken long to learn it was wrong when they experienced both types of combat with the actual odds display informing them of their chances.
Civrev tried to hide too much from the player IMO. Players still learn quickly that RNG can affect combat quite a bit, and they can get a feel for how much the combat odds depend on the strength ratios, but the interface fails to ever present that information to the player.
From what we've seen of civ5, they've tried to find a middle ground. They still want to give to give the player meaningful predictions of outcomes but they're trying to make it easier to quickly decide whether to go ahead or not.
Similar threads
