Stat Gain Question (spoilers possible)

[bpleshek]

In one of tables in the spoilers file for character creation list the following table:

skill lvl Gain
-----------------------
to 18 +1/+2
18/01 to /69 5+d15
18/70 to /89 2+d6
18/90+ 1

Is this the same formula that is used when drinking stat gain potions as well?

Brian

[PowerDiver]

In one of tables in the spoilers file for character creation list the following table:

skill lvl Gain
-----------------------
to 18 +1/+2
18/01 to /69 5+d15
18/70 to /89 2+d6
18/90+ 1

Is this the same formula that is used when drinking stat gain potions as well?

Brian

I have no idea what that is. It looks kind of similar to gains from stat potions, but except for the "to 18" I think you can gain more. E.g. I konw I have gone from 18/97 to 18/99 with one potion.

[bpleshek]

I found it in Ability.SPO.

But my question is, are those values correct for gaining stats with potions or is there another table instead?

Thanks,

Brian

[bpleshek]

i went from 18/01 to 18/31 in 1 pot. Is that a good roll ?

Brian

[aeneas]

i went from 18/01 to 18/31 in 1 pot. Is that a good roll ?

Yep- not all that unusual, I don't think, but I think that's about as much as you can gain at that point. I don't know the exact formula but as PowerDiver pointed out the above is definitely not it (for stat potions)- you can certainly gain more than 1 point even when over 18/90, and you just gained 30 when over 18, so... The progression is quite similar to that though, but the values are a bit larger.

[zaimoni]

In one of tables in the spoilers file for character creation list the following table:

skill lvl Gain
-----------------------
to 18 +1/+2
18/01 to /69 5+d15
18/70 to /89 2+d6
18/90+ 1

Is this the same formula that is used when drinking stat gain potions as well?

Brian

Almost. A stat gain potion first rolls whether to apply that table once or twice (1 in 4 chance, assuming recent versions have not deviated in this from when I last checked).

[bpleshek]

Well this is what i found in the code:

bool inc_stat(int stat)
{
int value, gain;

/* Then augment the current/max stat */
value = p_ptr->stat_cur[stat];

/* Cannot go above 18/100 */
if (value < 18+100)
{
/* Gain one (sometimes two) points */
if (value < 18)
{
gain = ((rand_int(100) < 75) ? 1 : 2);
value += gain;
}

/* Gain 1/6 to 1/3 of distance to 18/100 */
else if (value < 18+98)
{
/* Approximate gain value */
gain = (((18+100) - value) / 2 + 3) / 2; /* Paranoia */
if (gain < 1) gain = 1;

/* Apply the bonus */
value += randint(gain) + gain / 2;
/* Maximal value */
if (value > 18+99) value = 18 + 99;
}

/* Gain one point at a time */
else
{
value++;
}

/* Save the new value */
p_ptr->stat_cur[stat] = value;

/* Bring up the maximum too */
if (value > p_ptr->stat_max[stat])
{
p_ptr->stat_max[stat] = value;
}

/* Recalculate bonuses */
p_ptr->update |= (PU_BONUS);

/* Success */
return (TRUE);
}

/* Nothing to gain */
return (FALSE);
}

Based on this, my max gain should be:

((118-19)/2 + 3)/2 = 51

RandINT (51) + 25 (51/2) or 26 - 76 stat points in the first potion if starting at 18/01(19).

Then it gets harder each other time. Assuming that i roll(or cheat to roll) the max of 76, then for my next potion my start value is 18/77 or 95. Applying the formula, the next potion will give me:

((118-95)/2 +3)/2 = 7(.25)

RANDINT(7)+3(7/2) would give me a range of 4-10 for the next potion.

Am I understanding this correctly ?

Approximately how many stat gain potions do i need of each stat once I get to 18 so I can gauge how long i need to stay at 1600'.

Brian

[bron]

> Am I understanding this correctly
The idea is correct, but your math is a little off.

> ((118-19)/2 + 3)/2 = 51
> RandINT (51) + 25 (51/2) or 26 - 76 stat points in the first potion if starting at 18/01(19)
That should be ((118 - 18)/2 + 3)/2 = 26 and
randint(26) + (26)/2 = 14 - 39 for the first potion if starting at 18

There is also a bug here (well, at least I think it is a bug). If your stat starts out
at 17, and the RNG decides to give you and extra +1, your stat goes to 18/01
rather than going first to 18 and then to 18/14 - 18/39.

I am also personally annoyed by the comment in the code. If they want the stat to
increase by between 1/6 and 1/3 of the distance to 18/100, then that's what the code
ought to do. Instead, it is actually increasing the stat by between 1/8 and 3/8 of
the distance to 18/100 (with a little rounding error thrown in). Probably someone
had the mistaken idea that dividing by something other than a power of two would be
slow. Either the code or the comment should be changed. Preferably both.
[I actually wrote code to fix both these issues in my personal version of the source,
but never submitted it.]

[takkaria]

> Am I understanding this correctly
The idea is correct, but your math is a little off.

> ((118-19)/2 + 3)/2 = 51
> RandINT (51) + 25 (51/2) or 26 - 76 stat points in the first potion if starting at 18/01(19)
That should be ((118 - 18)/2 + 3)/2 = 26 and
randint(26) + (26)/2 = 14 - 39 for the first potion if starting at 18

There is also a bug here (well, at least I think it is a bug). If your stat starts out
at 17, and the RNG decides to give you and extra +1, your stat goes to 18/01
rather than going first to 18 and then to 18/14 - 18/39.

I am also personally annoyed by the comment in the code. If they want the stat to
increase by between 1/6 and 1/3 of the distance to 18/100, then that's what the code
ought to do. Instead, it is actually increasing the stat by between 1/8 and 3/8 of
the distance to 18/100 (with a little rounding error thrown in). Probably someone
had the mistaken idea that dividing by something other than a power of two would be
slow. Either the code or the comment should be changed. Preferably both.
[I actually wrote code to fix both these issues in my personal version of the source,
but never submitted it.]

I think by far the best solution here is to make stat gain potions gain one point up to 18, and 10 points at a time after that. This would simplify a fair amount of source with it. I expect to do this in 3.1.1.

[Narvius]

Can't we just have linear stats...?

[PowerDiver]

Can't we just have linear stats...?

That requires rebalancing time attacks, but IMO they need rebalancing anyway.

I hate the 18/xxx thing.