ZAngband is winnable.
That’s the allure.
It’s wicked hard. It’s not always completely fair. There are certain intrinsics that are required before venturing past certain points in the game (If you see a little blue “e” and you don’t have Free Action yet, you’re too deep).
Since the documentation can’t really cover every scenario, the learning curve is nearly asymptotic.
And even when you have a character that’s ready to win the game, there are still creatures in the dungeon that you *run* from (Ancient Multi-Hued Dragons come to mind).
But it’s possible to win. And that’s either enough for you, or it’s not, to keep you playing.
Should one spend the effort to learn the game well enough to win it? That’s not for me to say. Plenty of people seem to spend lots of time learning to execute perfect headshots in Halo. Being an expert at roguelikes is a skill that will probably be equally useless in “the real world”.
If you’re hung up on ASCII, yuhjklbn, or characters that you name carefully, you may want to quietly take a pass on these games.
Also, if you think baseball batting averages are a tough sell for your self-confidence, then you may be a bit disappointed with your “win/loss” record on the RL high score list.
I’ve been playing roguelikes for over 25 years. I estimate that I’ve started at least 1000 characters. I’ve beaten the game (in one variant or another) four times. Bait the Eighteenth (an Elf Mage in IMoria) was the first. But that’s a record of (approximately) 4 – 996. That’s a lot of Yet Another Stupid Deaths.
Bait VII the Daywalker (a Vampire Chaos/Sorcery Mage in Zang 2.7.3) just gained Light Resistance, and is doing pretty well right now, but, as always, “if he dies, he dies”.
The game is winnable. And it’s very, very fun. And so I play.