Luck in Wesnoth: Rationale

General feedback and discussion of the game.

Moderator: Forum Moderators

Locked
User avatar
Cackfiend
Posts: 559
Joined: January 28th, 2007, 7:36 am
Location: Florida, USA
Contact:

Re: Luck in Wesnoth: Rationale

Post by Cackfiend »

id like to give my 2 cents on this as well


the #1 reason i play wesnoth is for fun

I can easily have fun winning or losing, and I especially enjoy it when luck goes back and forth (also would like to point out theres two types of luck in wesnoth... Statistics and the ol' "my 1 hp unit didnt die after 8 failed attacks and i got to level it so now i win"


Long boring defensive games are not fun to me, and unfortunately these types of games happen the most at the highest level of competition

I still believe that at the highest level of competition in this game, and neither player can out strategize one another, it does indeed come down to luck.

this has just been my experience when I have played the best players in the game.




that said, wesnoth is a great game. I wish it didnt take 3-4 hours to complete some games because those tend to be rather boring. So i'll keep to my 1v1's and try to 2v2 on small maps only 8)

And of course, try to make it so luck does not determine the outcome of a game.
"There's no love in fear." - Maynard James Keenan

I'm the guy who's responsible for 40% Gliders in all hexes... I can now die a happy man. =D
Wesnoth Strategy Guide for competitive 1v1 viewtopic.php?f=3&t=54236
Dave
Founding Developer
Posts: 7071
Joined: August 17th, 2003, 5:07 am
Location: Seattle
Contact:

Re: Luck in Wesnoth: Rationale

Post by Dave »

JW wrote:
Dave wrote:There is no real comparison between the Wesnoth AI and the Fire Emblem AI. The Wesnoth AI isn't near as good as a strong player, but it's FAR and away much much better than the Fire Emblem AI.

Playing a Wesnoth campaign requires plenty of strategy and planning and lots of people enjoy it greatly. Fire Emblem doesn't play like this at all, and it's because the AI is very very weak.

David
Well I would request that you compare Wesnoth to a game with better AI and perhaps a different system as well, instead of only using "favorable matchups" if you will.
I honestly have never played a game that can be compared to Wesnoth that has an AI that is anywhere close to as good as Wesnoth's.

I haven't played Tactics Ogre, so I can't give you an analysis of its AI.

David
“At Gambling, the deadly sin is to mistake bad play for bad luck.” -- Ian Fleming
User avatar
JW
Posts: 5046
Joined: November 10th, 2005, 7:06 am
Location: Chicago-ish, Illinois

Re: Luck in Wesnoth: Rationale

Post by JW »

A couple more TBS that could be compared (don't feel like it myself at the moment):

Final Fantasy Tactics (PSX)
Gladius (XBOX)

Another great old game that is less similar, but could be compared:

Rise of the Pheonix (SNES)

-I don't remember the skill levels of the AI for any of these games ATM.
User avatar
jb
Multiplayer Contributor
Posts: 505
Joined: February 17th, 2006, 6:26 pm
Location: Chicago

Re: Luck in Wesnoth: Rationale

Post by jb »

Dave
I honestly have never played a game that can be compared to Wesnoth that has an AI that is anywhere close to as good as Wesnoth's.
Massive Assault has the best AI I've ever seen in a TBS game.

http://www.massiveassault.com/

It's a very fun game, which ironically uses zero luck.

I like wesnoth's luck system a lot, in fact it's one of the basic elements that I find attractive about the battle system. I guess unlike everyone else here, I seem to always get good luck. I'm either blessed, or the RNG fears me. :)
My MP campaigns
Gobowars
The Altaz Mariners - with Bob the Mighty
Dave
Founding Developer
Posts: 7071
Joined: August 17th, 2003, 5:07 am
Location: Seattle
Contact:

Re: Luck in Wesnoth: Rationale

Post by Dave »

JW wrote:A couple more TBS that could be compared (don't feel like it myself at the moment):

Gladius (XBOX)
Gladius is a good game; I enjoyed it (I enjoyed Fire Emblem too, for that matter).

However, Gladius's AI is also terrible. It does not compare to Wesnoth's. Wesnoth's AI can at least think reasonably well in tactical situations. Gladius's AI does not. It often moves in a seemingly random way. My wife, who is not much into strategy games, could easily recognize how bad Gladius's AI is, but after playing Wesnoth very often, she can't recognize any mistakes the Wesnoth AI makes.

Also, if you don't use the slider bars, I would say an individual battle in Gladius tends to depend on luck as much as a Wesnoth battle. If you try a combo attack and you miss, which is determined by luck, you will miss the entire combo! (Potentially exposing yourself to ripostes for the attack). This is a much huger effect than if you miss in Wesnoth.

Probably the only difference is that battles in Gladius are very very short and it's not usually considered a game over/loss if you lose one. So if you have some bad luck you can simply restart.

Also, Gladius has the option of using the slider bars, which luck-averse people can use.

I'm told that Final Fantasy Tactics and Tactics Ogre do not have particularly good AI's, though I have never played these games.

David
“At Gambling, the deadly sin is to mistake bad play for bad luck.” -- Ian Fleming
User avatar
TL
Posts: 511
Joined: March 3rd, 2007, 3:02 am

Re: Luck in Wesnoth: Rationale

Post by TL »

I've played FFT and one of the portable Tactics Ogre games. Both are RPGs first and foremost, so tactically they tend to be a bit simpler; their complexity mostly derives from character development options. FFT's AI in particular is downright stupid. The TO game I played at least had a minimally competent AI; it wasn't capable of doing much other than walking up and hitting people (or for ranged units, keeping distance and shooting people), but it at least performed those functions capably.

I think you may want to add a caveat to your horn-tooting though, Dave. Note that the default settings for the AI involve an incredible amount of stupidity. Given more sensible settings the AI isn't quite so bad although it still has some rather gaping holes, but a lot of peoples' experiences are flavored by those absolutely suicidal default settings.
Dave
Founding Developer
Posts: 7071
Joined: August 17th, 2003, 5:07 am
Location: Seattle
Contact:

Re: Luck in Wesnoth: Rationale

Post by Dave »

TL wrote: I think you may want to add a caveat to your horn-tooting though, Dave. Note that the default settings for the AI involve an incredible amount of stupidity. Given more sensible settings the AI isn't quite so bad although it still has some rather gaping holes, but a lot of peoples' experiences are flavored by those absolutely suicidal default settings.
I'm not "horn tooting". Simply pointing out that these other games aren't really built around having an AI at all. They are indeed more like RPG's.

When I talk of the AI, I mostly talk of campaigns. The AI doesn't provide a particularly interesting experience in multiplayer, it provides a much more interesting experience in campaigns where the campaign designer sets the AI up to play well in a particular scenario.

David
“At Gambling, the deadly sin is to mistake bad play for bad luck.” -- Ian Fleming
User avatar
TL
Posts: 511
Joined: March 3rd, 2007, 3:02 am

Re: Luck in Wesnoth: Rationale

Post by TL »

Dave wrote:When I talk of the AI, I mostly talk of campaigns. The AI doesn't provide a particularly interesting experience in multiplayer, it provides a much more interesting experience in campaigns where the campaign designer sets the AI up to play well in a particular scenario.

David
Ok. Interesting, yes, granted. But... campaigns routinely consist of superior AI forces with overwhelming terrain advantages being run into the ground with kill ratios of 3:1 or higher in the player's favor. How on earth does that constitute "playing well"?

(Again, I realize the AI is certainly capable of doing better than that--somewhat--but even in campaigns the default suicidal settings are often left alone).
Dave
Founding Developer
Posts: 7071
Joined: August 17th, 2003, 5:07 am
Location: Seattle
Contact:

Re: Luck in Wesnoth: Rationale

Post by Dave »

TL wrote:
Dave wrote:When I talk of the AI, I mostly talk of campaigns. The AI doesn't provide a particularly interesting experience in multiplayer, it provides a much more interesting experience in campaigns where the campaign designer sets the AI up to play well in a particular scenario.

David
Ok. Interesting, yes, granted. But... campaigns routinely consist of superior AI forces with overwhelming terrain advantages
Do they really? A typical scenario, I think, in a campaign, is The Siege of Elensefar, from Heir to the Throne. On medium difficulty level, the AI has 180 + 160 = 340 gold. According to stats.wesnoth.org, a player has over 400 gold arriving at this scenario. Additionally, a typical player will have a compliment of second level units by this stage, and perhaps some third level units. This all adds up to a large material advantage for the player. The player's main disadvantage is limited time, in addition to inferior terrain (but it's not that bad).

Even in a scenario where the AI is meant to have overwhelming force, such as Valley of Death, where the time is on the player's side, the AI has 'only' 1000 gold on normal difficulty. A player can have over 500 gold by this stage, and a substantial recall list to draw from.

Considering carried over player gold as well as the power of the recall list, I don't think the AI has that much of a material advantage in most typical campaign scenarios.

David
“At Gambling, the deadly sin is to mistake bad play for bad luck.” -- Ian Fleming
sam_waz_here
Posts: 77
Joined: November 13th, 2007, 7:15 am
Location: under a rock

Re: Luck in Wesnoth: Rationale

Post by sam_waz_here »

Dave wrote:
TL wrote: Ok. Interesting, yes, granted. But... campaigns routinely consist of superior AI forces with overwhelming terrain advantages
Do they really? A typical scenario, I think, in a campaign, is The Siege of Elensefar, from Heir to the Throne. On medium difficulty level, the AI has 180 + 160 = 340 gold. According to stats.wesnoth.org, a player has over 400 gold arriving at this scenario. Additionally, a typical player will have a compliment of second level units by this stage, and perhaps some third level units. This all adds up to a large material advantage for the player. The player's main disadvantage is limited time, in addition to inferior terrain (but it's not that bad).

Even in a scenario where the AI is meant to have overwhelming force, such as Valley of Death, where the time is on the player's side, the AI has 'only' 1000 gold on normal difficulty. A player can have over 500 gold by this stage, and a substantial recall list to draw from.

Considering carried over player gold as well as the power of the recall list, I don't think the AI has that much of a material advantage in most typical campaign scenarios.

David
That is all well and good if you are playing on the last level; but on most levels you won't recruit everything you can as soon as you can because:
A) Something valuable might die
B) The amount of money you lose from going all out and from upkeep outweighs the early turn bonus.

While the computer can use every coin it has because in the next scenario it will have a new army to use.

I am in favour of luck because when you get lucky it feels good rather then the feeling of being drained after playing a game like HOMM for hours. I don't know about anyone else but without the luck I would have gotten bored of Wesnoth long ago.
(Then again it is like the thrill of gambling but I'm not sure of your opinion on that.)
Where is the love?
F50
Posts: 48
Joined: April 22nd, 2008, 9:59 pm

Re: Luck in Wesnoth: Rationale

Post by F50 »

Wintermute wrote:
F50 wrote:I wouldn't mind altering the RNG, but its not a terrible issue for me. I don't think it would have that huge effect if the RNG was altered in the right way. Consider multiplying all unit's hp and number of strikes by 2. Any effect other than reducing luck would be psychological.
A simple example: The thunderer has one strike. Giving it two (even if every other unit gets their strikes doubled) makes it a very different unit from a probability point of view. Suppose a thunderer is shooting a unit that has 5 hp left and the unit has 40% def. That thunderer has a 60% ctk the unit in the attack. Now if we double hp and strikes, one strike will still kill the unit, but you have TWO chances to do that. Thus you have a 84% ctk the unit.

Now suppose we have a shaman attacking instead, and the suppose that the shaman is dealing 3-2. Currently, the shaman needs 2/2 strikes to hit in this case, so a 36% ctk. After doubling the hp and strikes, the shaman need 4/4 hits to kill a 10 hp unit, which is about a 13% ctk.

So your proposal would make the thunderer much better in this case (and many other cases), and the shaman worse in this case. Setting aside any benefits from such a change, my point (that hopefully should be obvious now) is that this is not as simple a change as you might think. :wink:
It is as simple as I think, though perhaps I did not make the definition of luck reduction: The distribution of possible consequences of an action are closer to the average consequences of an action. This will mean that things that have a small CTK have a smaller CTK, while things that have a good CTK, get a better CTK.

You provide a straw man by showing that I cause a disadvantage in one case and an advantage in another case when actually you simply show that the consequence of luck reduction holds true. If it is unlikely that the shaman will kill the unit (2/2 hits) then it will become more unlikely to make the distribution of possible consequences of an action closer to the average consequences of an action (no kill for you). In other words, the luck of the shaman is reduced. Other than the reduction of luck, the effect is psychological.

The situation with the thunderer is similarly flawed. Remember that doubling hit points is equivalent to halving damage. The problem with halving damage is that then not all damage amounts would be integers.

I am not underestimating the complexity of the game, but rather you are misunderstanding my idea.


I did forget healing though, that would have to be doubled as well.
User avatar
turin
Lord of the East
Posts: 11662
Joined: January 11th, 2004, 7:17 pm
Location: Texas
Contact:

Re: Luck in Wesnoth: Rationale

Post by turin »

F50 wrote:I am not underestimating the complexity of the game, but rather you are misunderstanding my idea.
I think you are - at least if you really think your change would require absolutely no changes for balancing purposes. Because you seem to assume that, even if luck reduction would change the usefulness of a given unit in a given situation, it would not have any effect on balance - which is clearly false. Changing the usefulness of a given unit in a given situation by definition changes balance.
For I am Turin Turambar - Master of Doom, by doom mastered. On permanent Wesbreak. Will not respond to private messages. Sorry!
And I hate stupid people.
The World of Orbivm
User avatar
Wintermute
Inactive Developer
Posts: 840
Joined: March 23rd, 2006, 10:28 pm
Location: On IRC as "happygrue" at: #wesnoth-mp

Re: Luck in Wesnoth: Rationale

Post by Wintermute »

turin wrote:
F50 wrote:I am not underestimating the complexity of the game, but rather you are misunderstanding my idea.
I think you are - at least if you really think your change would require absolutely no changes for balancing purposes. Because you seem to assume that, even if luck reduction would change the usefulness of a given unit in a given situation, it would not have any effect on balance - which is clearly false. Changing the usefulness of a given unit in a given situation by definition changes balance.
I agree, that is exactly the point. It would be simple enough to create an era that doubles everything, and edit a few maps to work the healing correctly. If you are really keen on the idea, I would be happy to help you test it. :)
"I just started playing this game a few days ago, and I already see some balance issues."
Dave
Founding Developer
Posts: 7071
Joined: August 17th, 2003, 5:07 am
Location: Seattle
Contact:

Re: Luck in Wesnoth: Rationale

Post by Dave »

jb wrote:Dave
I honestly have never played a game that can be compared to Wesnoth that has an AI that is anywhere close to as good as Wesnoth's.
Massive Assault has the best AI I've ever seen in a TBS game.

http://www.massiveassault.com/

It's a very fun game, which ironically uses zero luck.
I've never played this game, but it looks interesting; I will have to check it out.
“At Gambling, the deadly sin is to mistake bad play for bad luck.” -- Ian Fleming
Dave
Founding Developer
Posts: 7071
Joined: August 17th, 2003, 5:07 am
Location: Seattle
Contact:

Re: Luck in Wesnoth: Rationale

Post by Dave »

FS0 et al: "doubling everything" would be a big, fundamental change to the game. One of the key aspects of Wesnoth and tactical planning is luck management, and damage planning. A 6-4 attack is very different to a 8-3 attack which is very difficult to a 12-2 attack. Recognizing the differences between these attacks and planning around them is key. Doubling everything reduces this difference greatly. A 6-8 attack really isn't much different to a 8-6 attack, which is only a little different to a 12-4 attack: they all have very distributed results, and there are limited situations where difference between them matters.

If multiplying everything by 2 doesn't make a difference to balance, and is a solely positive change, why stop there? Why not multiply everything by a thousand? Then, there will be almost no luck at all. Combat results will be almost deterministic. Luck will be abolished.

Of course, I'm not claiming that our current setup is necessary optimum, though I personally think it is very good. You are welcome to try doubling things and see if it makes things more fun. I personally doubt I would find it more fun though.

David
“At Gambling, the deadly sin is to mistake bad play for bad luck.” -- Ian Fleming
Locked