Less random game

[Chris Byler]

Yeah, I think the objections are not so much to the system itself as to the small number of HP most units have and the commonness of attacks like 12-2. 12-2 attackers against a defender with 36 maximum HP are going to give VERY spiky results on any terrain whatsoever. 5-5 attackers (which are theoretically slightly stronger) are substantially less likely to kill the unit before it can react. (Both of those attacks are pretty strong for a L1 unit though - unless it is benefiting from time of day or leadership or negative resistance.)

I think some people would be happier with less spikiness in combat results, others would not. I can't see much of a solution other than refusing to play against people who play northerners or campaigns with lots of northerners in the scenarios, though. (Well, forking the whole project, but who wants to do that?)

[rrenaud]

Maybe there should be a "heroic" modifier to certain units that makes those units more reliable. After all, heroes are heroes because they come through when needed. The idea is that a unit will be able to trade off expected damage for higher probability to hit. Perhaps this feature can only be used when adjacent to a friendly unit that is in the red HP range. Though I guess changing probability to hit is just against the design (chance to be hit depends only on enemy units terrain); and increasing number of attacks for decreased damage per attack itself just doesn't seem heroic in itself.

[romnajin]

I'm not saying that I agree with your proposal, but increases to hit percent(for certain units) are not completely against the games design-see marksman and magical.

[Lugo Moll]

actually, its geared more towards realists who somewhat pessimistic.

they don't take unnecesary risks, but they do know when to take them.

optimists will lose too many unis due to bad, or moderate luck, and get a few extras kills from good luck.
pessimists will lose les units, but have less in the way of kills and xp.

for example, if I have an 83% chance of killing a target, I take it, no matter what.


( selfish weirdo knows what I'm talking about)

Realism=Pessimism

[jonadab]

there is no difference if you multiply the hit points by 4 but the attacks also by 4, the chance to die is the same, just now with higher numbers.

This is untrue. Drawing a larger number of random numbers decreases the standard deviation, which makes the results "feel" less random to most humans.

[Thrawn]

actually, its geared more towards realists who somewhat pessimistic.

they don't take unnecesary risks, but they do know when to take them.

optimists will lose too many unis due to bad, or moderate luck, and get a few extras kills from good luck.
pessimists will lose les units, but have less in the way of kills and xp.

for example, if I have an 83% chance of killing a target, I take it, no matter what.


( selfish weirdo knows what I'm talking about)

Realism=Pessimism

Pessimism=optimism w/ knowlege that hopes can't come true

[jonadab]

Actually, the discussion has nothing to do with trusting the computer or not.

But how can you speak for all users? I have seen many users more or less admit that they don't really trust the computer to generate fair random numbers.

This gets into the definition of what constitutes "random". Math majors and computer scientists define
"random" in a way that does _NOT_ match what most English-speaking people think the word means. In other words, when math geeks say "random", we're speaking a specialized math jargon, *not* standard English. Writing code that generates "random" numbers in the standard, non-math-geek sense of the word "random" is a VERY difficult problem. (In fact, substantially the easiest way to fake it is to revert to a pre-ordained "random" list.) A computer's "random" numbers are, to a human's way of thinking, horribly capricious and anything but "fair". (The definition of "fair" is another matter entirely, and it can be difficult to get two people out of a group of five to agree on a single definition.) Frequently the computer's random numbers will heavily favour one side. This is not because the deck is stacked, but rather because the deck is _not_ stacked, as it were, i.e., absolutely no thought is given (by the computer) to balance.

I work with a quizzing program ("program" in the non-computer sense of the word here), and in the course of working with it I wrote a program (in the computer sense) that generates quizzes, at random, from banks of questions. The algorithmic hoops my code had to jump through to get the quizzes adequately "random" to suit the organizers and participants defy a brief description. The code needed to assign specific numbers of each of several special types of questions per quiz, ensure that these special questions were distributed more-or-less evenly within a quiz, ensure that consecutive questions never come from too close together in the material, ensure that questions that are "too similar" to other questions are not used in the same quiz (which is *not* easy to code), and so on and so forth. Over the course of three years, about 30 such requirements have been imposed on the code. After all that, the results are finally getting to the point where they seem random enough to people that there aren't a lot of complaints about unfair quizzes. (The complaints about earlier iterations were about evenly distributed between people who knew that a computer program generated the quizzes, and people who did not know this.)

It is, of course, NOT reasonable to expect that kind of "random" out of Wesnoth. That would be like standing the KISS principle in the middle of a big cave with 20% defense and sending 500 Troll Warriors to attack it.

Nonetheless, saying "the computer's numbers are so random" is talking at cross purposes. When people say they don't trust the computer to generate random numbers, they're not using the computer scientists' definition of "random". If you tell them that fair results and unfair results are all thrown indiscriminantly together into one big pot and the computer just picks one out of the hat, they will probably on the whole believe this, but, they won't think it's fair.