by SaltyCrate » Thu Jan 24, 2019 12:29 am
First, as was noted before the +10% random quality roll is the entire point and enabler of this whole process. (Are you sure that it is actually +10%? Wiki says that it is 10% chance for 20% increase in quality and this is also what I recall from memory, but maybe both wiki and my memory are outdated.)
You already have a pretty good solution for your second problem, so this leave us with the first one. But here is the thing: while it is true that metal spiraling exagerrates it, even without spiraling at all the fact that "qualities of metal (or iron, at least) seem to depend mainly on the equipment used, rather than the ore" is still very true simply by virtue of relatively long crafting chain. Let's take Chainmail Shirt for simple crafting example: applying wiki formulas to the whole process we get that ore quality are factored with the coefficient of 162 / 1024 or 0.158 in the quality of final product (other coefficients are: smelter - 81/1024, coal - 81/1024, anvil - 400/1024, hammer - 300/1024). So will even removing spiraling entirely solve the problem you want to solve?
That is not to say that there couldn't be other problems found with metal spiraling itself. Depends on what one want in the end, really. I think either of your solutions seems works for removing metal spiraling. Another possible and maybe simpler solution which came to mind is to halve the percentage value of possible quality increase. From what I recall of participating in spiraling process, qualities of finery forge and coal lag behind quality of metal, and thus already apply softcap at creating blooms part. But the current potential quality increase when forging the bloom outvalues this softcap. Which would not necessarily be the case with the numbers changed.
Let's say we have q1000 anvil, q1000 hammer, q100 cast iron and q500 average of finery forge and coal. Here is comparative process of quality increase with the +20% modifier:
1. cast iron -> bloom: q100 (no change, since average of finery forge and coal only softcaps)
bloom - > cast iron: q493.7 (simply by the formula, assuming no crit in quality)
2. cast iron -> bloom: q493.7 (still the same)
bloom - > cast iron: q715.2 (still assuming no crit)
3. cast iron -> bloom: q607,6 (softcap applies)
bloom - > cast iron: q935,1 assuming the crit (q779.2 with no crit)
4. cast iron -> bloom: q717,5
bloom - > cast iron: q1009,4 assuming the crit (q841.2 with no crit)
Which is already an increase of quality compared to anvil and hammer. Usually at least one more step is made for more significant increase at the cost of cast iron bars, but I will omit it.
Now let's see what will happen with +10% modifier:
Steps 1 and 2 are the same.
3. cast iron -> bloom: q607,6 (also the same)
bloom - > cast iron: q857,1 assuming the crit (q779.2 with no crit)
4. cast iron -> bloom: q678,5
bloom - > cast iron: q901,4 assuming the crit (q819.2 with no crit)
5. cast iron -> bloom: q700,5
bloom - > cast iron: q914,4 assuming the crit (q831.2 with no crit)
And it marginally increases afterwards, but at a very slow rate, reaching equilibrium before q1000.
In this scenario simple reducing of percentage high roll from +20% to +10% allows avoiding infinite spiraling of metal. Obviously, it depends on how far qualities of finery forge and coal actually lag behind. However I feel that the gap in qualities in this scenario is somewhat realitistic, if not smaller than it usually is.