Point-based stat distribution is too powerful

[Nick]

What is wrong with skipping the first 20 levels of the game? IMO, the game isn't very interesting before stat gain (DL 30) to begin with.

I disagree completely. I think the first part of the game can be the most exciting part. Spells are unreliable, devices are few and far between. Careless play can lead to death from an infestation you can't escape, or a slightly OOD monster. I find that much more compelling than constantly crushing the skulls of endless demons using a zillion blows from your weapon of uber-destruction.

[bulian]

What is wrong with skipping the first 20 levels of the game? IMO, the game isn't very interesting before stat gain (DL 30) to begin with. IIRC there has been sentiment before about an option to start CL20/DL20.

This is an unneeded change.

[PowerDiver]

I came up with the points system, and not surprisingly I really like the way it does the tradeoffs. I think if you want to weaken it, it might be better to start with a base of 9 instead of a base of 10, and leave the points for improvements as they are.

I'm not convinced starting stats need to be lessened. That probably adds another 10% to stat gain, and stat gain was one of the banes of 3.0 gameplay for many players. You may really be opposed to the combat system, and are taking it out on stats instead of addressing the real issue.

[PowerDiver]

There was a reason to make it easy to get 3 blows. If you don't get 3 blows, you should use a launcher instead of melee to get +3/round from each damage scroll you buy in town, compared to +1 or +2 per round with fewer blows. If you don't want everyone to use a longbow, you have to make melee competitive. When the primary force behind increasing damage is enchant scrolls, that means 2 blows isn't competitive.

Now that enchant scrolls are black market only, I think that makes the argument less relevant. But it might make the argument more relevant if people scum for enchant scrolls from the market. I think watching what happens with 3.3 for a while is necessary before knowing how people are going to react.

[d_m]

PS.
Implementation is really easy, in birth.c, change the line:
static const int birth_stat_costs[18 + 1] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 2, 4 };
to:
static const int birth_stat_costs[18 + 1] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 3, 4, 5 };

PSS.
I did my first actual code change, with subsequent compile and testing today

Thanks for thinking about this. I will give your change a shot and see how it plays. I am actually happier with an easier early game (as I've said before) so I don't want to make things murderously difficult, but it may well be too easy now and I agree that characters start with pretty amazing stats at the moment.

Incidentally, if you are someone who's interested in modifying Angband's source and can compile changes I strongly encourage you to sign up for a github account and submit pull requests.

It's more convenient for us developers to pull in someone's request than to do anything else (even for these little one-line changes, it's easier). Plus it also helps us (or at least me) keep tabs on who's working on what.

Thanks!

[Derakon]

Congrats on compiling your first change.

Your suggestion sounds reasonable to me. Toss it in!

Would it also be fair to simply reduce the number of points available to buy stats with? Or, to get sidetracked only one post into the thread, even make that be an aspect of the race you play; high-elves start with good stats across the board but can't customize them very well, while humans have excellent customizability but a low starting point.

[buzzkill]

I been working up a rant (in my head) on what I'd like to refer to as "birth scumming" for a while now. You may have just saved me the trouble of actually putting it (the long version) into words.

The advantage bestowed by point based character creation is tremendous, every bit as advantageous as starting with, lets say an early artifact, or an ego weapon. If you were to stair scum for such thing it would be rightly be called scumming. Gain a similar advantage at birth and suddenly it's acceptable.

IMO it's even worse than stair scumming as it doesn't rely on random generation, you get to choose exactly the advantage you desire.

[jens]

I stated somewhere that the early game of Angband was a bit too easy. I was wrong, I was basing that statement mostly from playing mages. After starting a bunch of new characters with a wide variation in class and race, to test the stream of nightlies, I've come to the conclusion that the early game is way too easy.

The problem is that the first fifth of the game can basically be skipped. Just dive as fast as you can, gaining levels on the way, and before you need to stop you are at dlvl 20. This means we miss the fun that could be had from progressing a bit slower through those levels. Finding ego items that would be useful in the first levels, but could be discarded already at dlvl 20-30, etc. The game simply becomes shorter.

The main reason for this is that it's very easy to get multiple blows these days. Damage output for a lvl 1 melee character is way out of balance with the monsters he will face.

So, combat needs to be redesigned, well that's nothing new, and we are closing in on 3.3, so that issue is for another time. We can however make one small adjustment that would help balance this issue a bit. We can give starting characters less overpowering stats.

Time to switch viewpoint: How come we have the stats we have today?

Once upon a time there was a basic roller. This was balanced so the player would have a fun start in the game. When I mention balanced, I mean that the roller was (and is) designed to give the player very good stats compared to the 3d6 D&D system it was based on.
On the mission to reduce tedium an auto-roller was introduced. The auto-roller just rolled the roller until some minima you specified had been achieved. All was not fine though, the auto-roller stopped after 100 000 rolls IIRC, so you still had to reroll several times to get the best stats you could get.
Then the point based system was introduced. You now had three ways to get your stats. To ensure that the tedium of rolling was finally killed, the point based system was designed to allways beat the auto roller (well, it is still possible to get better results rolling, but not in any practical sense).
Later, when noone used the auto-roller any more, the auto-roller was removed, and thus we arrive at the current situation: One roller, that gives challenging stats, and a point based system, designed to always beat the best stats you could possibly roll. Meanwhile, the creatures in the dungeon have stood still in their development.

I suggest a small step backwards for the players, while staying far away from the tedium of rolling. The innovation of the point based system was great. The only problem is that it was compared to the auto-roller. Today the auto-roller does not exist, so it should be compared to the roller. I propose that we change the price/stat from (1,1,1,1,1,1,2,4) to (1,1,1,1,2,3,4,5). So, how does this compare to the standard roller? A few stat configurations could be beaten 1-2 times in 10 000 rolls (I've run a simulation for 100 million rolls). I do not believe that will risk introducing tedium.

The two main effects will be that you cannot get a few stats as high as before, and if you go for a few high stats, you will have a lower total count than before. You will still be able to get 24 points in stat increase if you spread them among all stats, so if you feel it's important to have a high total you can still get that. You will have to think more about each point and what you really need, and your number of blows/round will probably be a bit lower.

Some examples of stats you could buy:

Today				Proposed
(+8, +8)			(+8, +5) or (+7, +6, +2)
(+7, +7, +7)			(+6, +6, +5)
(+6, +6, +6, +6)		(+5, +5, +5, +5)
(+5, +5, +5, +5, +4)		(+5, +5, +4, +4, +4)
(+4, +4, +4, +4, +4, +4)	(+4, +4, +4, +4, +4, +4)

In most games you can choose a difficulty level from a wide range. This range needs to be wide to accomodate a wide variety of players and playstiles. In Angaband even the 'challenge' races/classes are too easy with the point based system, while the standard roller is a big jump harder. I propose to smooth that gap, and make room for a bit more challange in the early part of the game.

PS.
Implementation is really easy, in birth.c, change the line:
static const int birth_stat_costs[18 + 1] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 2, 4 };
to:
static const int birth_stat_costs[18 + 1] = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 3, 4, 5 };

PSS.
I did my first actual code change, with subsequent compile and testing today