Point-based stat distribution is too powerful

[Derakon]

Also, I don't honestly find the first 20 dlvls or so all that interesting-there's not enough variation. Later on in the game, the equipment you find can make a big difference in how you play and what you do, but the first 20 levels go about the same almost without exception.

This says to me that the early game monsters need an overhaul, personally.

[jevansau]

A flatter curve would be fine by me. Though, since that would mean moving more power to earlier in the curve, we might need to worsen starting stats to have equivalently powerful characters in the early game, which means that stat gain gets extended more.

I was making this suggestion in line with lowering starting stats. The idea was to make it that lowering starting stats would not mean an increase in the number of stat gain potions required.

It might also make getting stats up to 18 a bit quicker once a character does get to stat gain depth, but as previously pointed out, most of the power gain comes after that.

[Nick]

I would really recommend that contributors to and interested readers of this thread try Competition 105 to see what a challenging early game with varied monsters can be like.

[jens]

I'm not really seeing the motivation in changing the early game...I already find it disproportionately hard relative to the rest of the game, probably 90% of my deaths take place in the before 1000'.

Yes most deaths happen in the early game, and that is as it should be. If people lost 90% of their characters in the late game we would see a whole lot fewer winners... However I believe most of these early deaths are due to just that; players have not invested that much in their character yet, so it is much more efficient to take great risks. If you did play as carefully as you would if it was your own life at stake you would die a whole lot less, probably just if you happened on Grip and Fang too early and too far from a staircase...

[fizzix]

I've thought about this problem a bit. While I think combat changes are probably in order, I'm not convinced that these will be ready in the near future. However, I think several easy to code changes in concert can solve some of the inherent problems in the game, that have been outlined in this thread. These are:

1: The early game is trivial. It should be trivial for a half-troll warrior or a high-elf ranger. But it should not be for many other classes.

2: The game goes from being able to kill nothing to being able to kill everything somewhere between dlevel 30 and dlevel 50. This is due to an explosion of stats and gear.

From the stats perspective, I would propose:
1: Weakening starting characters. The approach of the OP is as acceptable as any other.

2: Changing stat gain structure to be +2 points until 18 and +1 post 18 (meaning 18/10 to 18/20) A stat potion at 17 will bring you to 18.

More difficult changes could consist of:

I'd also like to tweak monster frequency and gear frequency, but those changes are harder. Specifically, I'd like to make finite but rare possibilities for very dangerous monsters. Essentially the idea would be that if you continue to hang around dlevel 1 indefinitely, you will eventually run into something that can kill you.

As for gear changes. I'd like to make artifacts rarer. Low level egos (slays) more common. Sleep/slow/confuse wands/staves more common low in the dungeon where they are useful. These are harder changes.

[jens]

I've thought about this problem a bit. While I think combat changes are probably in order, I'm not convinced that these will be ready in the near future. However, I think several easy to code changes in concert can solve some of the inherent problems in the game

Precisely! There are lots of easy changes that can be made to improve the current game balance (current game as opposed to the coming game that we will have in a few years).

And I like your suggestions.

I'd also like to tweak monster frequency and gear frequency, but those changes are harder. Specifically, I'd like to make finite but rare possibilities for very dangerous monsters. Essentially the idea would be that if you continue to hang around dlevel 1 indefinitely, you will eventually run into something that can kill you.

As for gear changes. I'd like to make artifacts rarer. Low level egos (slays) more common. Sleep/slow/confuse wands/staves more common low in the dungeon where they are useful. These are harder changes.

These tweaks are all easy to make, but I admit they can be hard to balance. In any case these are statements that most seem to agree on, so I can't see much opposition to small tweaks. Small enough that they could not be unbalancing. Then we can tweak more after 3.3 is out.

[fizzix]

These tweaks are all easy to make, but I admit they can be hard to balance. In any case these are statements that most seem to agree on, so I can't see much opposition to small tweaks. Small enough that they could not be unbalancing. Then we can tweak more after 3.3 is out.

The changes to monster frequency are slightly more difficult. It Depends a bit on how you want to go about it. I agree the changes to gear is easy to do, but the difficulty is in balance.

However, I don't think *any* of these changes are appropriate for 3.3 The current mechanic is not broken, it's just not ideal. Let's get the 3.3 bugs out of the way and then view the game as a sandbox for more changes.

[jens]

The changes to monster frequency are slightly more difficult. It Depends a bit on how you want to go about it. I agree the changes to gear is easy to do, but the difficulty is in balance.

I've always felt there are too many low level monsters hanging around in the deep dungeon. I feel it is both thematically wrong (how would they survive to get that deep), and makes those levels easier.

That is easy to fix though: just change rarities in the monster.txt file so the earlier monsters are a lot rarer, then gradually become more common. Since the rarities are only relative, it would not affect the rarities on their native levels, but the deeper you get, the fewer low level monsters you would get.

And when I say it is easy to fix, I mean I have already done it. In fact I made a small program that converts the rarities. However I have not had time to test it, which is why I have not suggested this already. And probably it can be done better within the code.

Oh, and I do not propose this for 3.3, changing monster rarities will affect drops, which might affect game balance too much.

However, I don't think *any* of these changes are appropriate for 3.3 The current mechanic is not broken, it's just not ideal. Let's get the 3.3 bugs out of the way and then view the game as a sandbox for more changes.

Here I disagree. We still have time to tweak. I am sure we can tweak in such a way that the sandbox of 3.3 will be a better game. Thus making further changes and tweaks easier because we have a better reference to compare with.

[Magnate]

1: The early game is trivial.

I disagree with this. Lots of people have posted saying that they die a lot in the early game. If anything, the consensus seems to be that the later part of the game, should you get there, is closer to trivial (i.e. less chance of dying).

2: The game goes from being able to kill nothing to being able to kill everything somewhere between dlevel 30 and dlevel 50. This is due to an explosion of stats and gear.

Here I agree with you. There is a long-neglected ticket to reduce the size of the dungeon to address this, though personally I would prefer your solutions of rebalancing monsters and gear (and stat gain) to smooth things out until the very end.[/QUOTE]

[fizzix]

I disagree with this. Lots of people have posted saying that they die a lot in the early game. If anything, the consensus seems to be that the later part of the game, should you get there, is closer to trivial (i.e. less chance of dying).

Let me elaborate.

dlevels 1-10 are trivial for all but mages and priests. The spell casting classes are a little tricky.

dlevels 10-30 are relatively difficult only because many of us speed through them. There are lots of things that appear here that can kill an unprepared character. Still monsters tend to be more annoyance than deadly (think Zephyr hounds). A _teleport will keep you safe. So will spending some time and leveling up as new players should.

dlevels 30-50: These are very difficult until you hit the one or two good items or stat jumps that make things trivial and you go into super-dive mode.

[Magnate]

Let me elaborate.

dlevels 1-10 are trivial for all but mages and priests. The spell casting classes are a little tricky.

dlevels 10-30 are relatively difficult only because many of us speed through them. There are lots of things that appear here that can kill an unprepared character. Still monsters tend to be more annoyance than deadly (think Zephyr hounds). A _teleport will keep you safe. So will spending some time and leveling up as new players should.

Hmmm ... I'm still really not with you. Even veteran players lose tons of characters to Grip/Fang/Bullroarer/Wormy appearing at a bad time (e.g. just after a fight), or to breeder explosions, or to teleporting into a pack of hounds, or to carrion crawlers they've mistaken for orc captains. New players have all this to contend with and a whole lot more besides.

So I just don't get that the early game is too easy. It's certainly easier than it used to be, but I'm not convinced it's in need of serious rebalancing (though more interesting monsters are always welcome). Some tweaks to starting stats or monster distribution could help though - things can always be improved.

I think the main issue with difficulty is the "explosion" you have identified, where a huge number of good items drop in the middle fifth of the dungeon. I have three huge stats databases at http://www.terminalarrogance.com/phpliteadmin.php which contain data from over 100,000 dungeons (i.e. ten million levels). What we need is a SQL wizard to query them and plot this explosion visually for us, showing the levels where different things become available (damage, speed, ESP, etc. etc.). Without any coding changes we could then adjust rarities and depths and alloc_probs to smooth out these curves so that stuff wasn't available quite so quickly in the 2000'-3000' range, and more was only found in the 3000s and 4000s.

If anyone wants to do this, PM me for the password to the stats. (I haven't protected them out of any particular fear, but because the software requires write access to the data, so an irresponsible user could wipe it.)

EDIT: in case anyone is interested, some info on the stats. Each "run" simulates one "level-clearing ironman" game, in which all 100 dungeon levels are completely cleared once. Every floor object and monster drop is catalogued, but no level is revisited, skipped or partially cleared (so there is no accounting for trapdoors or tlev etc.). There is as yet no adjustment for breeders, summons, clone farming, acquirement or chest contents. Nevertheless, it should provide a reasonable approximation to what you *could* find in a game (if you cleared every level). I suspect that most people's real (complete) games will only net 10-50% of the objects catalogued in a single "run" (perhaps a little more for obsessive level-clearers and farmers).

[buzzkill]

The early game is only dangerous since characters have little room for error because of limited offence and low hit points (also, no fail safe escapes and limited detections). The monsters aren't particularly fearsome, quite the opposite, but the players have to actually plan ahead and think a little in order to avoid a possibly deadly convergence of events. Point based supermen are able to effectively side step most of this hazard with superior HP's and offensive power.

[Derakon]

I'd be interested in getting my grubby mitts on that database, though I can't promise I'll find anything particularly noteworthy.

For what it's worth, there's plenty of monsters that even "point-based supermen" have trouble with in the early game. Particularly group monsters -- gallants, orcs, and especially orc uniques -- can easily overwhelm a young character and require care (or at least lots of healing) to deal with.

[jens]

I'd be interested in getting my grubby mitts on that database, though I can't promise I'll find anything particularly noteworthy.

+1, but changing the last statement from 'find' to 'do'

[dos350]

Point based supermen are able to effectively side step most of this hazard with superior HP's and offensive power.

please, no rage~ !