Point-based stat distribution is too powerful

[buzzkill]

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

I actually go one step further and use the standard roller, where available. The problem as I see it is that point based has become the norm, and therefore the game will eventually be balanced to cope with the genetically engineered supermen, causing further 'inflation'.

[jens]

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.

I actuall got a GitHub account a while back, it's the easiest way to keep track on what you guys are up to without having to wait for the nightlies

I've been thinking about contributing through pull requests. This was sort of a start, checking that I can actually make and see the results of a change.

I'm still working very primitively though. What I did here was to (after getting the latest source from GitHub) copy the contents to a new folder, change one line of code with notepad, and then compile with MinGW (I'm on Windows 7). Too much hassle for any serious work, but Eclipse does not cooperate when it's time to compile

[jens]

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.

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.

I agree that today the first 20 levels are not very interesting, and I have (without knowing it has been discussed) thought about suggesting an option to skip them. But I, just like Nick, know I can get a lot of enjoyment from those levels. But to get there the starting characters need to be nerfed.

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.

We have now at least one new mechanic to help adress this (we can now soon allocate ego items to early dlvls, introducing new items that can be interesting at those depths, without cluttering the rest of the levels). We can do other design changes to make the early levels more interesting, without affecting the rest of the game. But all of these possible changes are moot if the starting characters can just plow through the first 20 levels.

[jens]

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.

Exactly!

Reducing the possiblity to maximise with the point-based system will force people to think more on what is important for them right now, and in the long run.

[jens]

You may really be opposed to the combat system, and are taking it out on stats instead of addressing the real issue.

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

I feel that the combat system is the bigger problem, it sure needs to be modified. However, modifying the combat system will take some time to design, implement, and balance. We are closing in on 3.3, so we do not have the time to do that now.

We do have the time to make a small change to starting stats though.

Besides, the starting stats being too good is also an issue, and it is important to adress this before fixing the combat system. If a new combat system is based on stats (which I assume it will be), it is much better that the stats have been balanced before the combat system is changed.

[jens]

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.

I really like this idea! In the birth code it's just one constant that sets that you have 24 points to play with. It should not be very difficult to have this set by race instead.

[jens]

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.

This depends on how a nerf is implemented, and then on player choice. With the price increase I propose players will get a more interesting choice to avoid prolonging stat gain.

Would it also be fair to simply reduce the number of points available to buy stats with?

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.

Here is a comparison between some different nerfs.
0) Current system.
1) My proposed price increase.
2) Use current system, but all stats start out at 9 (instead of 10 that is used today).
3) Same price, but only 18 points to buy with (6 points less than today).

Base stats achieved in the different systems, (sum of stats allocated)

0) current system	1) price increase	2) base stat = 9	3) 18 points to allocate
18,18,10,10,10,10 (16)	18,15,10,10,10,10 (13)	17,17, 9, 9, 9, 9 (10)	18,16,10,10,10,10 (14)
18,16,16,10,10,10 (20)	17,15,14,11,10,10 (17)	17,15,15, 9, 9, 9 (14)	17,15,15,10,10,10 (17)
17,17,17,10,10,10 (21)	16,16,15,10,10,10 (17)	16,16,16, 9, 9, 9 (15)	16,16,16,10,10,10 (18)
16,16,16,16,10,10 (24)	15,15,15,15,10,10 (20)	15,15,15,15, 9, 9 (18)	15,15,14,14,10,10 (18)
15,15,15,15,14,10 (24)	15,15,14,14,14,10 (22)	14,14,14,14,13, 9 (18)	14,14,14,13,13,10 (18)
14,14,14,14,14,14 (24)	14,14,14,14,14,14 (24)	13,13,13,13,13,13 (18)	13,13,13,13,13,13 (18)

Both 2 and 3 have the problem that they hit stat gain very hard, at least in the sense that the player cannot choose to get a higher total of stats. In both 0 and 1 the player can choose to spread out his stats and thus gain a higher total of stats, shortening the stat gain grind. To me, a steeper trade off for this gain is more interesting. I believe more players would consider lowering the best stats to gain quite a few points in a less important stat, compared with the current system. Another smaller problem with both 2 and 3 is that you can conceivably roll to achieve better stats when spreading them (the roller can achieve 14 in each stat).

[Magnate]

I feel that the combat system is the bigger problem, it sure needs to be modified. However, modifying the combat system will take some time to design, implement, and balance. We are closing in on 3.3, so we do not have the time to do that now.

We do have the time to make a small change to starting stats though.

Besides, the starting stats being too good is also an issue, and it is important to adress this before fixing the combat system. If a new combat system is based on stats (which I assume it will be), it is much better that the stats have been balanced before the combat system is changed.

There is no dispute that the combat system needs some fundamental fixing - fractional blows was merely a step on a long road. As well as blows and stats we need to look at weapon weights, dice, plusses/brands/slays (on- and off-weapon), criticals and AC/to-hit. There is a branch called "AC" at http://github.com/fizzix/angband which people can playtest - this splits monster AC into separate components for evasion and absorption. IMO this is the way we should be going, if not in 3.4 then 3.5 or whatever is next. It will need a lot of development and balancing, of course.

I agree that the stat system needs at least thinking about in tandem with this - perhaps we should just move to a linear scale (3 to 40), perhaps we should do something else. We should strive to eliminate breakpoints where possible and make any impacts of stats smooth across the entire range. We should also revisit which things are affected by stats and which aren't, as there is scope for improvement there too.

All this is for 3.4 or later. I am not yet convinced that there is merit in changing the birth system for 3.3, though I could be persuaded that we should make a small change now (such as changing the point costs, or reducing the number of points available) and spend the next month testing it, not being afraid to revert it if it doesn't work.

[Derakon]

One thing to keep in mind is that the early game needs to be forgiving. Newbies need to have a place to figure out the basic game rules without being killed over and over again; otherwise they're too likely to simply give up. Unfortunately, "forgiving to newcomers" readily translates to "too easy for veterans". I'm not aware of any way to reconcile those two so that veterans can have a challenging early game, short of expecting veterans to use weaker race/class combinations (or intentionally not fully exploit the point-based stat buyer).

Now, there's some question as to how much of the game needs to be intentionally forgiving -- where does it stop? 300'? 500'? 1000'? I think you could make reasonable arguments for any of those points. Whatever you end up with, though, unless you can resolve my first point, you're going to end up with a section of dungeon that veterans basically end up skipping, or at least blazing through as fast as possible.

[CunningGabe]

I really like this idea! In the birth code it's just one constant that sets that you have 24 points to play with. It should not be very difficult to have this set by race instead.

Seconded! We could give Humans 24 points and everyone else, say, 18 or 21 points. I think that fits nicely with the "humans can be anything" theme that is found in many RPGs.

[jens]

One thing to keep in mind is that the early game needs to be forgiving. Newbies need to have a place to figure out the basic game rules without being killed over and over again; otherwise they're too likely to simply give up. Unfortunately, "forgiving to newcomers" readily translates to "too easy for veterans". I'm not aware of any way to reconcile those two so that veterans can have a challenging early game, short of expecting veterans to use weaker race/class combinations (or intentionally not fully exploit the point-based stat buyer).

Now, there's some question as to how much of the game needs to be intentionally forgiving -- where does it stop? 300'? 500'? 1000'? I think you could make reasonable arguments for any of those points. Whatever you end up with, though, unless you can resolve my first point, you're going to end up with a section of dungeon that veterans basically end up skipping, or at least blazing through as fast as possible.

I've been thinking that the monsters in the first levels need to be more differentiated when it comes to how dangerous they are. What I was thinking was that 50' could be made easier, 100' a bit easier, 150' as it is, then start getting tougher than currently. So, I'd say the first 2 levels can be forgiving oh, and the farmer's dogs would then need to be moved to 150', and Smeagol to 200'...

Another option would be to intentionally lead newcomers to easier starts, say by making human a better race with relatively better stats allocation.

[CJNyfalt]

I've been thinking that the monsters in the first levels need to be more differentiated when it comes to how dangerous they are.

I would say that the the problem is more that the monsters are so boring the first levels: yellow centipede, white centipede, white icky thing, large brown snake, large white snake, small kobold, rock lizard & soldier ant are all quite similar and lacks anything that makes them stand out from the rest.

The worst boring monster problems are in my opinion:
- the centipedes, of which only the carrior crawler does something interesting.
- the snakes, of which only the poisonous are dangerous. Would work better modeling them after real world snakes. They are after all nastier than the *band ones.
- the ants, which seems to curiously be loners in *bands. Giving them FRIENDS flag would make them more interesting.

[Derakon]

Stegocentipedes are pretty terrifying for many of my characters.

I have to admit that all of the snakes and ants are pretty unmemorable though. But if you make the ants show up in groups early on, they basically become jackals.

[jens]

The snake that spits to blind can be dangerous, at least I always feel in danger when I am blind...

Making a level forgiving for newcomers does not have to mean boring. We can introduce a few status ailments that are not really dangerous, but can scare, and differentiate among them.

[Derakon]

How about an early frog that bites to cause hallucinations, then?