AWKWARD ZOMBIE

a number generated by a computer can't truly be random ;)

Yes it can. There are RNG systems set up to generate numbers from real-world input like webcams or even physical dice. It doesn't matter anyway, because most pseudorandom numbers are based on a seed that changes so frequently that you an never take advantage of their lack of true randomness.

I've got a question, does Advance Wars use a random number generator? What do you guys think about RNGs in that sort of game?
I never minded it in Fire Emblem, though I assume that's because the range of outcomes was narrow enough that it still came up as you'd expect the vast majority of the time.

Yes, all attacks are affected by luck. Some characters even have luck-based abilities (Nell, Flak, Rachel, Jugger).

I'm about to start XCOM EU.
I've heard the RNG in this game is a bitch.

Those aren't truly computer-generated though, and a game couldn't use a source like that for common purposes since there would be too much latency before the signal reaches the client. Also, even if the seed changes the way the seed is generated itself isn't truly random. None of this matters in practical contexts though.

Well, games could hypothetically rely in it. For example, you could get a true random number at a time when speed doesn't really matter (just query a server during the title screen, for instance) and then feed that number as a seed into subsequent generations.

It's true that it doesn't matter a lick, of course. I'm not sure why princess brothel brought it up.

Using the same truly random seed is worse than using a changing pseudo-random one I think, especially in terms of exploitability.

That's a pretty awesome approach, but it does seem to show that creating a non-random system with just as much variety requires more effort.

Is it? As long as the player can't force the game to accept a specific seed, the numbers will still be impossible to predict. As long as you base the RNG on a truly random number, I'd argue that any results are still collectively random.
((I'm assuming the RNG works like the ones in programming languages, which can reuse the seed without getting the same result each time))

http://www.sirlin.net/ This is a nice site with a couple of pages on game design, such as this http://www.sirlin.net/articles/slippery ... eback.html

Sirlin isn't a person I particularly like but he wrote some excellent stuff on game design. His book Playing To Win is awesome.

I disagree that 'slippery slope' is a bad quality in games, especially competitive ones. The reason is that slippery slope allows a player to build up 'marginal advantages' that don't pay off immediately but will eventually stack up and lead to victory. This encourages more subtle and clever play. For a full explanation, please read this.

Conversely, I disagree that perpetual comeback is necessarily a good quality. To me it devalues the comeback. Comebacks in games like Starcraft and MvC do happen, and when they do they're much more exciting because the comebacking player had to really outplay or outsmart his opponent to make up for his earlier mistake or opponent's good play. Comebacks in Mario Kart aren't exciting at all because they weren't dependent on the player's skill or decisions. That's an extreme example of course, and I can see why some people would find this game element fun since it does preserve tension throughout the entire game.

On the other hand, games that 'snowball' very hard like League of Legends are less well designed imo as well because any small advantage quickly becomes game deciding. This prevents the need for many marginal advantages and instead makes the game feel too volatile which will often lead to players playing very conservatively to prevent the opposing team getting any small advantage. It also makes comebacks near impossible.

What do you guys think about balancing in single player games? I've always been pretty conflicted about it.

My two major opinions, which are a bit contradictory, are A] it should be possible to build an 'optimal' strategy and B] it should be possible to succeed by skill in most situations.

Of course, they can both work in a game, but the idea of the optimal build is weakened by making it not really matter.

I also think my views on enemy balancing are unpopular ones: I like enemies that are actively unfair, because I think it's better to have intimidating enemies. There's no tension to games if you know you can plow through an entire stage. It's why I dislike most RPG leveling systems in non-RPG games; they let you just increase damage rather than making you build skill, and you only rarely meet something you can't easily beat into submission.

In singleplayer, enemies that are unfair can be very good if done right, like creepers in minecraft. But in a multiplayer game, everyone would want to play the "creeper" that can go around one shotting people.

Ah, I meant for the single player distinction in the first sentence to carry through.

This very much depends on genre. In stealth, strategy, certain types of RPG, etc. strategy should play a large role, and can be argued to be the main "skill" the game is centered around. Action games should require more quick-thinking. It's not an absolute thing, but something that chafes depending upon the individual instance.


Also I just wanted to say that there is a big gap between "unfair" and "legitimate challenge." An enemy can be difficult without being unfair, and an unfair challenge should be the fail state of any game development.