This is offtopic but I think this is actually a far more interesting discussion than the original topic so I'll go ahead and post instead of not posting. As I've said before I don't personally actually have any problem with crawl's difficulty, and in fact I think it's in a pretty good spot. This post is not going to have a coherent position or anything, it is more a collection of observations.
I'm not even sure what you are trying to query here:
- Code:
Sequell: 13202/193328 milestones for !greatplayers (br.enter=Lair): N=13202/193329 (6.83%)
Sequell: 5412/15456 milestones for greatplayers (br.enter=Lair): N=5412/15456 (35.02%)
Obviously worse players will win less often after reaching lair; they're worse (and while greatplayers does not specifically select players who are good at crawl, it contains a large number of the best and they are all at least good enough to have won 20+ times). You didn't even do a query that supports your position at all: you should note that "!@greatplayers" also reach lair much less often than greatplayers do!
Here's a better query:
- Code:
<crate> !lairratio !@greatplayers !boring rstart>2014
<Sequell> !greatplayers (!boring rstart>2014) has reached Lair in 39084 of 404003 attempts: 9.674185587730785 %
<crate> !lm * !@greatplayers br.enter=lair !boring rstart>2014 / won
<Sequell> 2432/39091 milestones for * (!@greatplayers br.enter=lair !boring rstart>2014): N=2432/39091 (6.22%)
<crate> !lairratio greatplayers !boring rstart>2014
<Sequell> greatplayers (!boring rstart>2014) has reached Lair in 1335 of 3752 attempts: 35.58102345415778 %
<crate> !lm greatplayers br.enter=lair !boring rstart>2014 / won
<Sequell> 552/1335 milestones for greatplayers (br.enter=lair !boring rstart>2014): N=552/1335 (41.35%)
This data at first glance does support your argument that post-lair is more deadly than pre-lair games for most crawl players.
I would suggest that this isn't necessarily meaningful, depending on what you actually care about; it's far, far easier to "luck" your way into reaching lair than it is to win an entire game: even really awful players and bots will reach lair sometimes (even rw has done so! and it has zero hope of ever winning the game). Perhaps the best illustration is this:
- Code:
<crate> !lairratio qw
<Sequell> qw has reached Lair in 172 of 676 attempts: 25.443786982248522 %
<crate> !lm qw br.enter=lair / won
<Sequell> 3/172 milestones for qw (br.enter=lair): N=3/172 (1.74%)
qw literally cannot think for itself (it is a bot); if you play crawl in this completely mindless fashion, then you can still reach lair 25% of the time (and win occasionally)! You might even argue that there is no difficulty at all in crawl if a bot can win: you just automate the whole thing and then keep trying until you succeed.
Personally I don't really find myself able to draw any conclusions about how difficult different parts of crawl are from looking at players who don't win with some sort of regularity. You can make it to lair purely by holding o and tab (and sometimes pressing G>) with MuBe occasionally, but does that actually give you a meaningful statement about crawl's difficulty? You're also unlikely to beat Super Mario Bros without ever pressing the run button, but that doesn't say much about how hard that game is, right? I see this as sort of the same thing. Maybe you don't.
Players who do win with some sort of regularity are almost certainly aware of the important crawl mechanics and are certainly outperforming qw.
I find it pretty telling that very nearly every single good player (even using either the "goodplayers" nick or "greatplayers" nick) dies less often after reaching lair. Using all of a player's games here is problematic: it will obviously bias the sample toward dying more often anywhere (as players get better at crawl over time, and I'm trying to select players who are already reasonably good). But even if you do strip out early games (with rstart) then the same pattern holds: pre-lair is more deadly (or very close to the same, for a few players like N7) for every single player I've checked except for a few who are specifically not trying to win every game (e.g. 4tharra).
This tells me that even if you play reasonably well, pre-lair has deaths that are much more likely to happen. However, if you compare the really good players (elliptic, mikee_, bmfx, etc.) to merely "goodplayers" or "greatplayers" then you still see that they reach lair much more often, so there is absolutely still skill in avoiding these deaths. To me this says that pre-lair crawl is definitely more difficult, and this is the only group of players where you can study data to compare pre-lair games and post-lair games in my opinion (as I said above; I am not going to elaborate on this more; as I said this post is not trying to convince anyone of anything, really). Of course, a large majority of crawl players never win even once.
Having to learn such a large number of monsters and items and such that crawl has is something that strikes me as more of "
fake difficulty" than anything. Dying to something once because you don't know what it does doesn't necessarily mean that said monster is difficult to deal with (Mara, for instance, will almost certainly kill an unspoiled player the first time if the player chooses to fight, despite the fact it's usually quite easy to just avoid Mara if you know he's that dangerous). I don't have a problem with having to die to learn how to win crawl, but I do have a problem with the number of different things that crawl includes, but as I said earlier I'm not going to talk about that.