Quazifuji wrote:dpeg wrote:(1) Games are basically not winnable, the best you can hope for is getting as far down as possible. In this setup, player skill is measured by dungeon depth: the better you are, the deeper you get (taking averages). This has approach has the potential for a lot of depth.
(2) Games are 100% winnable. All else being equal, in this setup there is a cap on depth: if you learn/figure out the skills necessary to win a game, you can effectively win every game (assuming you have the willpower).
I think case 1 makes sense in theory, but I'm not sure how it works in practice.
Note I said
extreme cases. I never said I want (1) outright, but rather that Crawl sits too far at the other end of the spectrum, for its own demerit.
Next, I don't believe that (1) is theoretically superior -- in fact, that very old philosophy section contains a clear statement on that: in an ideal world, Crawl games would be winnable by pure skill. My point is that this goal is delusional, and we get a better game, with greater depth, if we accept some more unavoidable deaths: by allowing more near-death situations, we provide more potential for skill. We almost certainly also provide more unavoidable deaths. All I say is that moving a little bit in this direction improves the game (in my opinion -- there is obviously no clearcut correct answer to this problem). What would constitute a good increase of challenge is then a major piece of "local design".
If you make a game essentially "unwinnable" by making these situations more common, then you're not really filtering for player skill, because these situations, by definition, can kill players who play optimally.
(Emphasis by me.) I think this is wrong. Only because
some games have this bad characteristic, this doesn't not mean at all player skill is pointless. Very good players would still stand out: they would win more very close seeds, and their winrates would show it.
The main effect such a change would have is that streaks become more rare, and so less a yardstick of player prowess. I am fine with that, as for me, the presence of arbitrary long streaks is a clear sign that Crawl is broken. I think my reasoning on this is clear by now. (Again, I am not saying I own the truth on this matter. I thought about this issue for a bit, and I believe that depth would be increased if Crawl was less forgiving.)
If you have a game where reasonable non-psychic optimal play can result in the player unavoidably dying on D2, then it doesn't matter whether the person playing is me, Elliptic, or someone playing their first game ever. And in that case, the fact that the person died on D2 doesn't measure their skill at all.
Yes, but it's one game in many, and over the course of several months, say, the differences in player level would clearly be seen. I am not arguing to turn Crawl into a die game, I want it to be a little bit harder!
If you want a game where even the best players can't win 100% of the time, and where a player's skill is measured by the average depth their characters reach and not by their win rate, then what you want isn't a game that isn't winnable 100% of the time with optimal play.
This is one of the more complicated sentences I've read recently.
I believe you think my position is much closer to (1) than I ever meant.
But a game where making every decision perfectly doesn't always result in a win doesn't filter for skill.
Like your sentence I emphasised above, I don't think this is true. Elliptic would perhaps stand out
more in a harder Crawl. Sure, he would lose some games without chance. But he would get farther in those games than the pack, and there'd be more games that are really close, so his skill can shine.
Because a situation that kills a player who is playing perfectly specifically does the opposite - it's a roadblock that levels the playing field by killing everyone no matter how skilled they are.
You are making this statement for the third time, so let me finish with another thought experiment about extremes:
(A) Crawl as now, basically always winnable.
(B) Same Crawl, but when entering Depths, the game rolls a die and you die upon a 6.
In (B), there are plenty of stupid, unavoidable deaths. Elliptic would be affected like everyone else (who manages to reach Depths). Would we will still see that elliptic is an extremely good player? Of course we would! This bizarre toy model shows that your reasoning is flawed.
I'm not saying that making every Crawl game theoretically winnable is an achievable or worthwhile goal. I think people have made very good arguments in this thread that the randomness of crawl will inherently make some games effectively unwinnable. And I think the goal of making wins dramatically rarer so that player skill is measured by average depth of death, not win rate or streaks or score, is a perfectly fine one. But making it so that not all games are theoretically winnable is not the way to accomplish that, in my opinion.
(Emphasis mine.) This is another very problematic detail: I claim that it is extremely hard to prove "theoretical winnability" in a complicated and random rule set like Crawl's. So if you want to go there, you have to err on the side of the player, consistently and often. I say that what you then get is a game with a doubtful slogan ("every game is winnable" -- people in this thread have argued that this cannot be achieved as long as combat is random) and is less deep than it could be.
Anyway, thanks for the thoughtful reply, the thread was a bit ... out there already