Deterministic mode

[Jetrel]

The Knalgans have one in their Dwarvish Fighter, and the Elves have one in their Wose. Trolls can be pretty tough to kill quickly too, especially considering the price.

Yeah, I love those woses. They rock - except, there's just one little thing - YOU CAN'T GET THEM IN HTTT.

In fact, the only such unit you *can* get is the dwarvish fighter, and that's halfways into the game. Plus those guys crawl on most terrain.

[ImLittleJon]

That it's completely ridiculous nonsense. It's especially aggravating when you get cheated out of a lucky swing.

Try explaining why you think it's ridiculous nonsense. And also why my proposal would cheat you out of a lucky swing. Not to mention how you would know that you had been cheated out of a lucky swing...

wait, this algorithm doesn't even always work!
Suppose X=.5. Then the total is 2.0, and the part is 1.2. If the player misses, then the total is decreased by 1, leaving a probability of 1.2/1.0=120%! What's going on?

The algorithm is only valid for values of X >= 1.

The downside is that it makes 4 attacks each having 60% to hit, not really act like 4 attacks, each with 60% chance to hit.

True, it's not the same thing as sampling with replacement. Please try to explain to me why sampling with replacement is a priori better.

But i dont think messing the statistics is a good idea. Especially when the charts show one probability, but you actually get another probability.

This says to me that you don't fully understand the proposal. It's not messing the statistics, and it's not misrepresenting the probability. The 60% represents the expectation value, which is preserved. All that changes is the standard deviation.

[Dave]

The downside is that it makes 4 attacks each having 60% to hit, not really act like 4 attacks, each with 60% chance to hit.

True, it's not the same thing as sampling with replacement. Please try to explain to me why sampling with replacement is a priori better.

Simply because it's easier to understand. As well as it being less predictable, which is arguably a good thing, since among other things, battle really is unpredictable.

David

[ImLittleJon]

Simply because it's easier to understand. As well as it being less predictable, which is arguably a good thing, since among other things, battle really is unpredictable.

It is easier to understand, but users don't need to understand it. It's human nature to notice the unexpected, not the expected. As long as the value of X is chosen wisely, nobody is ever going to go "hey, how come I get the result I expect more often than I really should". Other than that, it's completely transparent for the user.

I'll give you that battle is unpredictable. However, Wesnoth is a strategy game, not a simulation, and in that context, I (and whoever else complained enough to get the topic into the sticky thread) prefer a little more predictability.

[Elvish_Pillager]

That it's completely ridiculous nonsense. It's especially aggravating when you get cheated out of a lucky swing.

Try explaining why you think it's ridiculous nonsense.

The game tells me I have 60% CTH. Actually I have some crazy probability that varies during the fight. I expect a fair combat with normal chance. I get a distorted combat fit to the purpose of making some guy less frustrated. If anything, it'll just make them MORE frustrated when such a thing does happen, because "How could it possibly happen with this system???"

And also why my proposal would cheat you out of a lucky swing.

When I hit with the first 3 of my elvish fighter's four swings, the last is much less likely to hit.

Not to mention how you would know that you had been cheated out of a lucky swing...

Every time my elvish fighter hits three times and misses the fourth, under your algorithm, I will assume I have been cheated.

[Jetrel]

I want to clarify something, before people misunderstand my last post.

I don't care what race it comes from, but I think the player in Httt should have some form of "tank" unit, sooner or later.

Scenario construction could well allow a long chunk of time before the player gets one. But, they need one eventually. Some strange side-track where they enlist the aid of the woses could work, or the dwarves could work.

However, they need one that can move more than 2 spaces per turn on most normal terrain, and for that the dwarves don't count. The dwarves are a smidge too slow in one terrain - forest primarily.

I know the dwarves are not meant to be good in forest, but moving even four spaces per turn is not good. Right now they only go half that. I think it would be perfectly acceptable to make the dwarves go "full speed" in the forest, considering that their top speed is pretty glacial. On top of that, the elves would still run circles around them, and would have much higher defense to boot.


Given that almost all the dwarves are currently modified to give a different, but consistent resistance set, I don't think it would be unreasonable to give the dwarves their own movetype.


Personally, I am of the opinion that if the dwarves were able to move at full speed in forest, I would have no further complaints about units dying. I would have my tank for HTTT, and I would be happy.

all the other scenarios have a workable tank:
SoTBE orcs: use trolls or warlord, preferably trolls.
TEI humans: Heavy infantry
TDH undead: skeletons/revenants

Not all scenarios should have such a tank, but the main scenario of the game should. The main scenario is not, and should not be too strange in terms of game workings. People have enough to get used to with wesnoth, it's nice to have some touchstones of what they are familiar with.

[Dacyn]

The algorithm is only valid for values of X >= 1.

My philosophy is: if it doesn't work for all values, it probably doesn't work very well for the others. Also, why is your algorithm called 'sampling without replacement'? I seem to recall something by that name that was different from this...

It is easier to understand, but users don't need to understand it. It's human nature to notice the unexpected, not the expected. As long as the value of X is chosen wisely, nobody is ever going to go "hey, how come I get the result I expect more often than I really should". Other than that, it's completely transparent for the user.

Wesnoth is Open Source; nothing is "completely transparent" to the user.
Given that, I don't think it makes sense to have the game rules say one thing and have the code say another. So if we put sampling without replacement in the code, then we also need to explain it as part of the game rules. So the users would need to understand it.

Jetryl: isn't your discussion a little OT?

[Dave]

nobody is ever going to go "hey, how come I get the result I expect more often than I really should".

I am going to do that. I like to analyze combat and work out my chance of killing, and chance of being killed. A system like this would make it difficult for me to do that.

I don't care what race it comes from, but I think the player in Httt should have some form of "tank" unit, sooner or later.

I'd actually kinda like this too. It'd allow more diverse strategies in many scenarios in the campaign.

David

[blackjack]

My thinking re the two subjects in this thread:

1. "Determinstic" Mode

I think that sampling without replacement etc or other forms of reversion to the mean would not work well in Wesnoth. It will make games boring, reduce the scope for strategy and reduce BfW to a bean-counting exercise. The AI would be much faster, and deadlier. (Right now, it does some form of Monte Carlo sampling, which would become much more accurate)

Mean reversion tends to drastically reduce the variance. I believe that part of the charm of Wesnoth is that you have to plan for contingencies. Bad luck? Live with it. Better still, expect it.

Any empirical observations about distributions in BfW (or rand) is probably flawed. If you flip a coin 2000 times, what is the expected length of the longest "run" of the same face? The vast majority of folks here will underestimate the correct answer. Yet, if you play HttT through once, I'll expect about 2000 calls to rand.

2. Tank unit

Sorry to disagree. In HttT, at the early levels, the emphasis is on using quick hitting, ranged units and retreating to heal. The Mobility of elven troops in forest and their high defence is the key. Not tanking. (though an elven fighter in forest is pretty close to a tank)

Tanking will just cause players to take greater risks, and make them save-load more to survive "one more turn" before running behind cover. Of course, this is just my opinion and I could be wrong about the ability of Wesnoth players to play properly.

[Neoriceisgood]

I just look at the map and enemies always and form a inpenetratable squad of strong fighters; may it be paladin and grand knight when it comes to long distance with little turns, to dwarven lords and dragonprinces when having to fight deep in caves.
my "tank" group practicaly always has my mage of light though as he ensures my lines of surviving longer as it sometimes heals them more than the damage they get.

[Dacyn]

The algorithm is only valid for values of X >= 1.

X=1
4*X*.6=2.4
Suppose the player hits 3 times:
-.6/4=-15%
Valid?

[Dave]

2. Tank unit

Sorry to disagree. In HttT, at the early levels, the emphasis is on using quick hitting, ranged units and retreating to heal. The Mobility of elven troops in forest and their high defence is the key. Not tanking.

Sure, but by adding slow, powerful units we can add more variety. Players can choose if they want speed, range, and healing, or if they want to use a slower, more methodical approach.

David

[Jetrel]

little note for blackjack. Do some math. The player only gets the dwarves on his/her side about, like, halfways into the game.

If we did woses, they would still likely be several levels into the game.


Elven swordsmen in forests basically are tanks, cause they're so hard to hit. The problem is, the early levels have a nice selection of forests to hide in. But the later ones don't.

Besides, those dwarves need something cool to do when bearing down on Asheviere's forces. Tolkien characterized dwarves as being nearly twice as hardy/durable as a human. They would make a crackin' good shield line.

[Elvish_Pillager]

Besides, those dwarves need something cool to do when bearing down on Asheviere's forces. Tolkien characterized dwarves as being nearly twice as hardy/durable as a human. They would make a crackin' good shield line.

Oh yes. More HP for dwarves!

[Dacyn]

maybe this topic should be renamed