Early game (pre-Temple or pre-Lair)

[bel]

[These are my tentative thoughts. I have no idea if I'm on the right track. This thread is related to the "starting attributes" thread elsewhere on GDD.]

What is the purpose of the early game? It's not clear to me that there are some sort of general principles. I see the following aspects:

• Weaker versions of threats you'll encounter later in the game.
• Teaching you about skilling, threat assessment etc.
• Choosing a God.
• Determine your initial playstyle. Basically due to sunk costs and starting items.
• Gain items and experience for later.

The early game is very hard, but, as far as I can see, this state of affairs is neither necessary nor thought out, nor can it be derived from other principles (if there are any). All players (but especially newer players) see a lot of the early game; not so much of the later game: therefore it is important. I see several bad things about the early game:

• Things like adders, which can randomly kill you. Orc priests are another such monster. Monsters with elec or venom short blades are bad too, but usually they can be avoided.
• Threats way out of proportion with how strong you are at that point in the game. Sigmund or Grinder on D:2 or D:3 are much tougher than any late-game threat.
• Few or no consumables, so few options. Low XL, HP and MP are obviously given.
• You are pretty much forced to skill in a certain way because you have to survive. There's very little margin for error or experimentation.
• There are lots of different kinds of monsters in the early game. Is the player supposed to die to every monster to see how they work?
• Early game monsters have very high EV. This is because their AC has to be low, since the player does so little damage. But this makes for very swingy combat.

I suggest making the early game easier (perhaps start with more consumables, as I've suggested elsewhere, or more XP), reducing the different kinds of monsters and make the loot more varied (perhaps one can borrow the early vault idea from Brogue). The early game should be easier because you need breathing room if you want variety (and the early game is too hard anyway). Fewer kinds of monsters (and fewer very tough monsters) reduce cognitive load and unfair deaths at the point of the game where you have few options to deal with them. More variety in loot would go towards meeting the "starting package" aspect of backgrounds instead of "you have to skill in this way to survive" aspect.

[Siegurt]

I disagree that the early game is significantly harder. It's a popular trope, but I think that results from people who aren't good at the game seeing more of the early levels than the later ones.

I happened to install crawl on a new computer earlier this week, and played through a couple dozen games so far, my deaths are:
D1-2: 5
D3-4: 5
D5-6: 6 (One was a shafting from D:3 though score that how you will)
D:7-8: 5
IceCave: 1
D12: 1
Shoals3: 1
Swamp4: 1
Vaults3: 1

It's been my experience generally that my deaths from portion to portion (as a percentage of survival from the previous portion of the game) remain fairly constant throughout.

[bel]

Spoiler: show

My argument was not solely based on early game hardness, but the point is interesting. Here's how I approach the issue.

First, we need to define a reasonable metric for the hardness at each point in the game: A floor's hardness is the probability that you'll die on the floor, given that you have survived to see that floor.

I ran Sequell queries about tenpercenters (who are presumably not too bad, though I suppose one can find bad players even in this group).

(Wait wait, doing it again, with tenpercenters recent !boring.)

"!lg tenpercenters recent !boring / place=D:1" and so on. I got 8.3%, 8.1%, 8.9%, 6.9%, 4%, 2.4%, 2.9%, for the first seven floors respectively. So, for hardness, I get:

D:1 -- 0.083
D:2 -- 8.1 / 91.7 = 0.088
D:3 -- 8.9 / 83.6 = 0.106
D:4 -- 6.9 / 74.7 = 0.092
D:5 -- 4 / 67.8 = 0.058
D:6 -- 2.4 / 63.8 = 0.037
D:7 -- 2.3 / 61.4 = 0.037

D:1-4 seem to be about the same hardness, then it drops sharply at D:5, and goes down even more. The trend seems downward. Later calculations are harder because of Lair/Orc branches, but one can probably do it with some reasonable approximations.

Gah I realized that I forgot portals like Sewer. Anyway. I would have to fall back on folklore till I figure out a way to do it better.

[duvessa]

What is the purpose of the early game? It's not clear to me that there are some sort of general principles.

I mean...

  Code:

Major design goals
  * challenging and random gameplay, with skill making a real difference
  * meaningful decisions (no no-brainers)
  * avoidance of grinding (no scumming)
  * gameplay supporting painless interface and newbie support

Minor design goals
  * clarity (playability without need for spoilers)
  * internal consistency
  * replayability (using branches, species, playing styles and gods)
  * proper use of out of depth monsters

• Weaker versions of threats you'll encounter later in the game.
• Teaching you about skilling, threat assessment etc.
• Choosing a God.
• Determine your initial playstyle. Basically due to sunk costs and starting items.
• Gain items and experience for later.

Like, these are all true, but I think your list is missing a pretty big item, namely "challenge the player and get them to make interesting and meaningful decisions". The early game's not supposed to be a tutorial, it's supposed to be a challenging and fun part of the actual game.

Things like adders, which can randomly kill you. Orc priests are another such monster. Monsters with elec or venom short blades are bad too, but usually they can be avoided.

I agree that adders (and some other monsters) have a noticeable chance of causing deaths that aren't reasonably avoidable, but I disagree with putting orc priests in this category. This probably sounds like a nitpick, but I think it's pretty important. They aren't fast and they're not a common sight on d:1.

Few or no consumables, so few options. Low XL, HP and MP are obviously given.

I don't think this is a bad thing about the early game at all. I think having consumables available makes the game less interesting, not more. Suppose you wake up a centaur, and may not be able to kill it before it kills you. Without consumables, you need to decide whether you're more likely to survive by fighting the centaur in melee, or running up the stairs (or behind another monster) and taking some arrows. Making the right choice here could mean the difference between dying and not dying.
With identified consumables, you're not going to choose either of those options, because you're not stupid: you have consumables that virtually guarantee not dying. Instead you decide whether you want to definitely lose a potion of might, or possibly lose a potion of heal wounds. The stakes are low and the consequences probably won't matter. The decision isn't nearly as meaningful. If you have a scroll of blinking, pretty much the only threats to your character are one-shots and forgetting to use your scroll of blinking, neither one of which are satisfying ways to die.

• You are pretty much forced to skill in a certain way because you have to survive. There's very little margin for error or experimentation.
• There are lots of different kinds of monsters in the early game. Is the player supposed to die to every monster to see how they work?
• Early game monsters have very high EV. This is because their AC has to be low, since the player does so little damage. But this makes for very swingy combat.

These are definitely important weaknesses in the early game. I have to point out that the rest of the game suffers from having too many monsters even more, though, and learning what things do by dying to them on Dungeon:2 is a lot faster than learning what they do by dying to them on Depths:2. (yeah Zot definitely needs 48 different draconian monsters)

I don't agree that the early game is too hard. There's a large set of characters that almost any spoiled player can stomp the early game with: most Bes, Ces, DDs, Trs, and a bunch of more specific species/background combos like HOFi. Almost all the remaining combos can consistently win if played well. I think if the game were easier to learn - i.e. simpler and better mechanics, better tutorials - the difficulty of the early game wouldn't pose an approachability problem.
Perhaps something in the vein of chess puzzles would be useful for Crawl? There's already Sprint for mostly unrandomized dungeons, but the Sprints aren't built to be instructive by any stretch of the imagination. It's a sexy pipe dream that anyone would want to devote the time to actually implementing that, though.

I definitely wouldn't want Crawl's early game to play like the rest of it. In the rest of the game practically all of the deaths are due to impatience/laziness (not stopping for the few dangerous monsters, because there are so many monsters), or straight-up not being informed ("I didn't know hydra simulacra could do 150 damage", "I didn't know infernal demonspawn had AF_FIRE", "Oh, draconian knights have a spell that actually does damage", etc.). Whereas even the best players often have early-game deaths that they can trace back to deliberate decisions that they made.

[nagdon]

Early game is the good part of the game because there you can be killed by real challenges and not just by boredom. In the late game the game design disallows causing unavoidable deaths, and because the challenges are randomly generated this means that most situations are very far from challenging (because even the most unlucky situations must be survivable by everybody who got there).

You are pretty much forced to skill in a certain way because you have to survive. There's very little margin for error or experimentation.

This is a good thing. If skilling did not affect the chance of survival, then it would be meaningless.