Point-based stat distribution is too powerful

[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

[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.

[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.

[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!

[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.

[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.

[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.

[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.

[Philip]

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.

You might be thinking in terms of O with a shadow-fairy mage. In V you are constantly crushing the skulls of endless orcs, kobolds, yeeks and such using a zillion blows from your dagger of flame +5 +8.

[CJNyfalt]

It is stats that is the real problem, or is it the combat system?

[jevansau]

I think the idea of lowering the starting stats - whether by lowering the starting base, increasing the rate at which upgrades cost or lowering the number of points available would be good as long as it is combined with a way of speeding up stat gain. Maybe stat gain potions could increase stats by 2 or 3 points up to 18. That would keep the number of potions required similar but allow for more character growth.
Having varying customization amounts between races sounds neat.

[Hariolor]

Aren't most of the arguments in favor of skipping (by diving or by birth option) the first few dozen levels actually arguments about those levels being too easy?

Making the game harder from the word "go" might add interest back into those levels, making the skipping argument moot?

What seems to have happened here is that a few disproportionately powerful mechanics introduced a long time ago have created an arms race of exploitation/patching that has led to the innumerable posts these days about X Y or Z being too powerful/too useful. I think dragging players kicking and screaming back a few notches down the power-scale might be a good thing for the game.

I, for one, certainly get the most enjoyment out of the first twenty minutes of the game - at least up until I am prepping a likely winner. This enjoyment comes precisely because those first few levels require caution, forethought, etc. Not that this ever goes away entirely, but certainly CL50/DL 98 plays much differently from CL1/DL 2.

[Jazerus]

The main issue here is that before 1000' the game is pretty uninteresting. The monsters require no tactics beyond making sure to face groups one at a time, you have very few interesting options even as a mage, who is supposed to be the most varied and tactical class, and the loot is uniformly awful unless you find a *thanc or Forasgil...in which case you immediately dive to 1000' to start having more fun. The early game isn't terrible, but it's not as good as the rest of the game.

I'd be all for reducing character power if it made things more interesting, but my feeling is that it would just increase early sudden death, making a new character (and new players) less likely to get to the interesting floors. Fewer blows wouldn't result in melee characters having to use more tactics but rather necessitate increased consumable usage to tank through longer fights.

I'm also not really sure why the game picks up at 1000', but the difference is fairly noticeable for me. Stat-gain depths are also a spike in the interest curve.

[artes]

Yes, the first levels are very easy with melee characters that have multiple blows. Not as easy with mage. I'm not saying anything has to be changed though. That depends on what kind of game we want.

Most people like a character with optimized stats. Rerolling 100000 times with the auto-roller is kind of ridiculous though.

Point-based stat distribution has the downside that all characters will be the same, like warriors will always have maximized str and dex, and a high con, while minimum on the other stats. That's boring.

Recently I've played a couple of characters with completely random stats, but even then I've rerolled a few times.

Since most people seems to want to have optimized stats, I don't see a reason to take away auto-rolling or point-based stats, though of course the system can be adjusted if there is a need for the sake of balance.

[dos350]

blah blah cry cry , i want to ruin game more!

anyway if u want to see real abuse of stat, try old autoroller~!

please, no rage