The Tavern

This was suggested by friendlybee, thanks friendlybee.

Right now the average aptitude is somewhere around -1. While this is fine for me, a dispassionate robot, a lot of players are turned off from training things when they see a -1, despite the fact that it is a fine aptitude. So, I propose that the aptitude scale be shifted by 1, simply to shift perceptions for players starting out that something is bad because it is -1. A side benefit is that humans' advantage is now apparent by them having all 1s, and demigods/demonspawn's average aptitudes would now appear as 0, which strikes me as a more elegant/harmonious baseline than -1.

I'm (hopefully) obviously not proposing any actual change in gameplay or behavior here, just what numbers are displayed in the skill screen.

I know this would be a little confusing for experienced players, myself included. I think they could probably take the hit though, and the help it would give to new players trying to get over a mental block would be worth it.

It makes sense that 'average' would be +0, and 'good at learning' is +1.

The average aptitude is closer to 0 than it is to -1. So I don't find that argument very persuasive at all.

duvessa wrote:The average aptitude is closer to 0 than it is to -1. So I don't find that argument very persuasive at all.

Shows what I get for believing oft repeated mantras. You're right, I get about -0.30 average if I substitute NAs for 0, and count all draconians, and -0.38 without NAs and with only plain draconians.

It's psychological. You see a creature with a bunch of -1 aptitudes, Joe Newbie immediately thinks it's terrible, when that's not all that bad.

OK so let's add 0.3 or 0.38 to all aptitudes then it's fine ;)

XuaXua wrote:It's psychological. You see a creature with a bunch of -1 aptitudes, Joe Newbie immediately thinks it's terrible, when that's not all that bad.

This is the main point to me. Look at the recent thread about a tanky caster race where someone lamented that fact that races with good HP and/or Armour aptitudes all have few to no positive spell aptitudes. As people pointed out, the wealth of +0 and -1 magic aptitudes with the occasional +1 on races like gargoyle, hill orc, or formicid are actually very respectable caster aptitudes, but to a new player they look poor to mediocre. People instinctively expect races to bad bad at anything they have a negative aptitude in and meh at something they've got a 0 in, but really those are perfectly respectable aptitudes for most skills.

I think a key thing is that it's not just about averages, but just about quality. You can make a good case that "0" should be the closest to average, which makes the current numbers the best choice, but that doesn't necessarily convey the fact that a -1 is still a good enough aptitude to use most skills effectively. Just looking at the numbers without having the experience playing the game to have an intuition for what different aptitudes really mean in practice, Hill Orcs, gargoyles, merfolk, etc look like somewhat poor magic-users, when really they're very solid spell-casting races who just don't excel at it to the extend that races like Elves do.

In the case of magic, there's also the issue of new players over-estimating the importance of spellcasting, which i a different problem, although bumping up all aptitudes would still remove the perception that races with a -1 in spellcasting are below average magic-users.

So if you don't care about the scale being centered at 0 (since you propose adding 1 to all aptitudes), and don't care about the scale making sense (since you propose making all aptitudes act as if they were 1 lower), and your entire problem with the status quo is "negative numbers are too negative-sounding", why not suggest increasing all aptitudes by 5 instead? That gets rid of all the negative numbers and isn't any worse in other respects.

The current setup has the advantage that we can write in ?%

Code:: -4 terrible aptitude (learning half as fast as at 0 aptitude) +4 outstanding aptitude (learning twice as fast as at 0 aptitude)

which is meaningful information. You could write something like this after an arbitrary transformation of the aptitudes, but I claim it looks worse afterwards.

That's a valid point. I'd forgotten that you lose the symmetry if you shift everything. I'll admit that the current system is cleaner, and it may make more sense to just let new players learn that a -1 aptitude isn't that bad rather than switch to an uglier system just to make the numbers potentially slightly more intuitive i some ways.

dpeg wrote:The current setup has the advantage that we can write in ?%
Code:
-4 terrible aptitude (learning half as fast as at 0 aptitude) +4 outstanding aptitude (learning twice as fast as at 0 aptitude)

which is meaningful information. You could write something like this after an arbitrary transformation of the aptitudes, but I claim it looks worse afterwards.

-5, -1, and 3 share the same relationship, though, so you can say the same thing after adding 1. Or has this changed?

dpeg wrote:The current setup has the advantage that we can write in ?%
Code:
-4 terrible aptitude (learning half as fast as at 0 aptitude) +4 outstanding aptitude (learning twice as fast as at 0 aptitude)

which is meaningful information. You could write something like this after an arbitrary transformation of the aptitudes, but I claim it looks worse afterwards.

You could say the same thing unmodified I think - correct me if I'm mistaken but isn't the +4 doubling is true anywhere along the line, not just centered at zero?

duvessa wrote:So if you don't care about the scale being centered at 0 (since you propose adding 1 to all aptitudes), and don't care about the scale making sense (since you propose making all aptitudes act as if they were 1 lower), and your entire problem with the status quo is "negative numbers are too negative-sounding", why not suggest increasing all aptitudes by 5 instead? That gets rid of all the negative numbers and isn't any worse in other respects.

I actually mentioned this elsewhere when the discussion came up - it's just as valid. I don't know in what sense it "makes sense" right now other than that the current numbers correspond to some amount of skill xp versus those same numbers corresponding to a different amount of skill xp.

johlstei wrote:
dpeg wrote:The current setup has the advantage that we can write in ?%
Code:
-4 terrible aptitude (learning half as fast as at 0 aptitude) +4 outstanding aptitude (learning twice as fast as at 0 aptitude)

which is meaningful information. You could write something like this after an arbitrary transformation of the aptitudes, but I claim it looks worse afterwards.

You could say the same thing unmodified I think - correct me if I'm mistaken but isn't the +4 doubling is true anywhere along the line, not just centered at zero?

This is correct.

So if we upped the aptitude display of all races by 1 that specific chunk of text in the help wouldn't have to change at all.

In fact the only thing that upping the aptitude display of all races would have to change, is the racial aptitude chart in the ?% screen and in the skills menu...

Maybe it could be a toggleable option in the init.txt

I guess part of the trouble comes from background recommendations for newbies (MiBe, DEFE, SpEn,...) always having huge positive aptitudes in the relevant skills

But overall I think its not worth it, having two different inconsistent notations for the same thing complicates more than it helps. And I think everyone learns a negative aptitude is not the end of the world after trying a handful of combos

Let's make it really confusing and set an aptitude to be 100 for 'normal' with lower numbers meaning better aptitudes and higher worse.

Let's make it really confusing and set an aptitude to be 100 for 'normal' with lower numbers meaning better aptitudes and higher worse.

In all seriousness, when the old system was being replaced, was the option of inverting it and getting rid of the factor of 100 ever considered? Then a 'normal' aptitude would be 1.00 - which makes sense. And higher aptitudes would be better, and moreover the system would be fairly intuitive - an aptitude of 1.18 means that the skill trains 1.18x faster than one at 1.00. IMO one of the problems with the current system is that player's don't know *how* much worse a -1 aptitude is than a 0 aptitude, and probably often assume that it's worse than it really is.

Also, this system would avoid negative numbers - bad aptitudes would simply get closer and closer to zero.

ion_frigate wrote:
Let's make it really confusing and set an aptitude to be 100 for 'normal' with lower numbers meaning better aptitudes and higher worse.

In all seriousness, when the old system was being replaced, was the option of inverting it and getting rid of the factor of 100 ever considered? Then a 'normal' aptitude would be 1.00 - which makes sense. And higher aptitudes would be better, and moreover the system would be fairly intuitive - an aptitude of 1.18 means that the skill trains 1.18x faster than one at 1.00. IMO one of the problems with the current system is that player's don't know *how* much worse a -1 aptitude is than a 0 aptitude, and probably often assume that it's worse than it really is.

Also, this system would avoid negative numbers - bad aptitudes would simply get closer and closer to zero.

It's really not all that hard to figure it out. Relative to skill level 0 we get:
-5: 42.0% as fast
-4: 50.0% as fast
-3: 59.5% as fast
-2: 70.7% as fast
-1: 84.1% as fast
0: 100% as fast
1: 118.9% as fast
2: 141.4% as fast
3: 168.2% as fast
4: 200.0% as fast
5: 237.8% as fast

nilsbloodaxe wrote:It's really not all that hard to figure it out. Relative to skill level 0 we get:
-5: 42.0% as fast
-4: 50.0% as fast
-3: 59.5% as fast
-2: 70.7% as fast
-1: 84.1% as fast
0: 100% as fast
1: 118.9% as fast
2: 141.4% as fast
3: 168.2% as fast
4: 200.0% as fast
5: 237.8% as fast

I'm not making any statement on the broader question of this thread, but I have to point out that most people are better at understanding percentages than logarithms.

ion_frigate wrote:In all seriousness, when the old system was being replaced, was the option of inverting it and getting rid of the factor of 100 ever considered? Then a 'normal' aptitude would be 1.00 - which makes sense. And higher aptitudes would be better, and moreover the system would be fairly intuitive - an aptitude of 1.18 means that the skill trains 1.18x faster than one at 1.00. IMO one of the problems with the current system is that player's don't know *how* much worse a -1 aptitude is than a 0 aptitude, and probably often assume that it's worse than it really is.

For a couple versions after the change in aptitude format, the "118" values were still shown in ?% next to each new aptitude. Later those numbers were removed from the screen so that the meaning of aptitudes other than -4 and 4 was hidden. This was clearly a deliberate design decision. (IIRC part of the reason for the original change was to force aptitudes to be more discrete).

PleasingFungus wrote:I'm not making any statement on the broader question of this thread, but I have to point out that most people are better at understanding percentages than logarithms.

Logarithms? For a human to convert a "new" aptitude to a meaningful value they perform one exponentiation. For old aptitudes, it was even easier, you just had to divide by 100.

IIRC part of the reason for the original change was to force aptitudes to be more discrete.

I gather that's correct, but personally I don't get why the display has to be changed for that. The devs can definitely just say "These are the aptitudes we're using, deal with it" to anyone who starts calling for finer aptitudes (which I agree are pointless for the most part).

It's really not all that hard to figure it out.

You have to memorize all these values or have a calculator handy with the formula. The formula may not be complex, but for aptitudes that aren't 0, 4, or -4 it gives irrational numbers - so no chance of being able to do it in your head. Some people might know 2 and -2, but that's not actually as common as you might think, and still leaves over half the aptitudes as unknown. The point is that a newer player is not that likely to have Google/Wolfram Alpha open, and is not likely to memorize these numbers - and they are likely to just assume "-1 is kinda crappy, right?", when it's in fact perfectly acceptable for most purposes.

ion_frigate wrote:You have to memorize all these values or have a calculator handy with the formula.

Or you just ignore the details and move on. It's not like the exact percentages matter.

nilsbloodaxe wrote:It's really not all that hard to figure it out. Relative to skill level 0 we get:
-5: 42.0% as fast
-4: 50.0% as fast
-3: 59.5% as fast
-2: 70.7% as fast
-1: 84.1% as fast
0: 100% as fast
1: 118.9% as fast
2: 141.4% as fast
3: 168.2% as fast
4: 200.0% as fast
5: 237.8% as fast

Is there a logical reason why it's done this way instead of:

-4 50 %
-3 62,5 %
-2 75%
-1 87,5 %
0 100%
+1 125 %
+2 150 %
+3 175 %
+4 200%

Would seem far more accessible & easy to remember this way, to me at least.
(I know its % based decrease/increase in the current one but it just seems overcomplicated that way)

The way it is, an aptitude difference of 1 always means the same thing, in a certain sense. The way you suggest is just completely arbitrary. Yes, the percentage numbers are prettier, but that doesn't matter because they are not shown anywhere.

Galefury wrote:The way it is, an aptitude difference of 1 always means the same thing, in a certain sense. The way you suggest is just completely arbitrary. Yes, the percentage numbers are prettier, but that doesn't matter because they are not shown anywhere.

But they are shown, inside the game, as numbers.
And if i had not searched this out from somewhere, i would still think that +1 is +25% exp, +2 +50% exp etc, just because the game only tells me that -4 is 50% exp and +4 is 200% exp, and not the numbers in between them. (aptitudes.txt)
But maybe this is just me, and other people conclude that is isn't increased on steady increments of 25%.

And in my example, the differences mean the same thing as well, in a certain, different sense.

In the current system, going from an aptitude of x to an aptitude of y means you spend (2)^((x-y)/4) times as much xp to train it. Thus, decreasing an aptitude by 1 always means you spend 1.18 times as much xp as before, regardless of where you started.

Or you just ignore the details and move on. It's not like the exact percentages matter.

This is exactly the problem. People are doing this. But 'ignoring the details' is leading to them imagining that a -1 aptitude is a lot worse than it actually is.

Another option would be to get rid of the numbers altogether and replace aptitudes with adjectives, or a visual display like spell hunger.

ion_frigate wrote:Another option would be to get rid of the numbers altogether and replace aptitudes with adjectives, or a visual display like spell hunger.

I like this idea provided we have less adjectives/aptitudes. As extreme let's have only 3 adjectives: good (+4), average (0) and bad (-4), though maybe having also +2 and -2 could be better. If difference between 0 and -1 is minor, why is it still in the game? Does it affect any decisions?

Sandman25 wrote:
ion_frigate wrote:Another option would be to get rid of the numbers altogether and replace aptitudes with adjectives, or a visual display like spell hunger.

I like this idea provided we have less adjectives/aptitudes. As extreme let's have only 3 adjectives: good (+4), average (0) and bad (-4), though maybe having also +2 and -2 could be better. If difference between 0 and -1 is minor, why is it still in the game? Does it affect any decisions?

In isolation, a difference of one is minor, but three or four small differences of the same type do add up, and have a meaningful impact, and thus can affect your decisions. Even if the difference between 0 and -1 is relatively small, in one aptitude, what would be gained by suddenly making everything 0, +4, or -4? Extreme aptitudes tend to railroad species into small niches, and this change would bifurcate literally everything into either "as good as merfolk are with polearms" or "as bad as trolls are with stealth." That's an extreme change that would not be good for game play.

At most, I could see aptitudes being rescaled from -3 to +3, and having it pretty much work out, but I don't think that's needed.

It should also be considered that no matter what you do, many new players will initially shy away from investing in skills when there is a bad aptitude, especially if they don't have a lot of previous roguelike experience and/or they are most familiar with games that encourage min-maxing.

Yeah, but they'd think 'humans are good at everything!' and 'demonspawns are average at everything.' rather than
'humans are average at everything.' and 'demonspawns are bad at everything!'

Since players are supposed to think the first thing rather than the second thing, having the aptitude system more clearly express that would be good. Since few people understand what the aptitude numbers even mean currently, there's not much lost by changing it, and there is something gained.

Maybe a crawl math tutorial would be a good idea. You could flavour it up or whatever. "You hear the screech of chalk off in the distance! A portal to Sif Muna's Remedial Math Clinic has opened somewhere on this level! Find it before it closes!"

damiac wrote:Yeah, but they'd think 'humans are good at everything!' and 'demonspawns are average at everything.' rather than
'humans are average at everything.' and 'demonspawns are bad at everything!'

Since players are supposed to think the first thing rather than the second thing

What? Why would you want them to think the first thing? Humans are closer to "average at everything" than demonspawn are.

"Average" is a useless concept, though. The fact that humans are slightly better than average at everything and demonspawn are slightly worse than average at everything is meaningless, because you're not competing against other races, you're fighting against the monsters in the dungeon. What matters is that demonspawn are competent at everything and humans are pretty good at everything, because the average aptitude is actually pretty strong.

Whether this is something that should be conveyed to a new player or is better (or at least perfectly fine) as something that they can just discover for themselves through experience (or learn from other players) is another matter. But I don't think the notion of above or below average is relevant here. -1 does a perfectly fine job of conveying "slightly below average", what it does not convey is that the average is quite good and slightly below average does not mean the species is bad at something.

Perhaps adding 5 to all apparent aptitudes (so that they were a range from 0 to 9) might convey that more accurately (it doesn't seem like as big of a deal if humans are 5s and ds's are 4s)

This would also mean no minus signs in the charts making them easier to read.

Quazifuji wrote:What matters is that demonspawn are competent at everything and humans are pretty good at everything, because the average aptitude is actually pretty strong.

The average aptitude is "pretty strong" compared to what? Not the average, obviously. The median aptitude? Oh wait, that's 0, the average aptitude is worse than the median.

duvessa wrote:
Quazifuji wrote:What matters is that demonspawn are competent at everything and humans are pretty good at everything, because the average aptitude is actually pretty strong.
The average aptitude is "pretty strong" compared to what? Not the average, obviously. The median aptitude? Oh wait, that's 0, the average aptitude is worse than the median.

I think they meant something along the lines of "pretty strong when you make heavy use of it relative to the difficulty curve of the dungeon" but that's pretty nebulous.

johlstei wrote:
duvessa wrote:
Quazifuji wrote:What matters is that demonspawn are competent at everything and humans are pretty good at everything, because the average aptitude is actually pretty strong.
The average aptitude is "pretty strong" compared to what? Not the average, obviously. The median aptitude? Oh wait, that's 0, the average aptitude is worse than the median.

I think they meant something along the lines of "pretty strong when you make heavy use of it relative to the difficulty curve of the dungeon" but that's pretty nebulous.

That is what I meant, and the reason I didn't respond earlier was because I realized it was pretty nebulous. That's part of the issue here, I think. Comparing races to each other is objective and easy, but somewhat meaningless since you're not playing against other races, you're playing against the dungeon. But judging what an aptitude means in terms of your ability to use that skill to battle the dungeon is vague and subjective enough that it might make more sense to just drop it.

The Tavern

Add one to all displayed aptitudes (visual only)

Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)

Re: Add one to all displayed aptitudes (visual only)