A 28+ Attack Brown Work Glove

Purpose:

I've always wanted to see one of these made. However, to this day, I've YET TO SEE anyone even bother to try. Why not? People have chaos'd every other item to death, including Zhelms, Belts, and HTPs, yet they refuse to stick these wondrous scrolls on their +3ATK 8slot Brown Work Gloves. Why is perfection limited to 10% scrolls? Why stop short of the sky?

Method:

Calculations, because they're more exact and it's impossible to obtain data through actual experimentation or simulation.

Materials:

Table 1: Items Needed and their Average Costs

#	Variable	Item	Cost	Calculations	Amount Vs. Control
1	Control	+6atk 5slot BWG via 30% scrolls	2.222b	(200)/(0.3)/(0.3)	1x
2	Experimental Group 1	+6, +7, or +8atk 5slot BWG via Chaos scrolls	5.244b	(200/0.3+120)/((1+5+10)/64x0.6)	2.36x
3	Experimental Group 2	+9 or +10atk 4slot SCG via Chaos scrolls	12.8b	(600+120)/((1+5)/64x0.6)	5.76x
4	Negative Control	+27atk 0slot BWG via 2 30% scrolls and 7 50% scrolls with White scrolls (cheapest way to make this)	14.822b	(200)/0.3/0.3+900(7)(2)	6.67x

Calculations:

Assuming that you have the exact funding to make a perfect +27atk BWG and decide to make your BWG using Experimental 1, then you would basically have enough money to scroll, under statistical conditions, the following number of slots with 50%s + white scrolls, which is the cheapest way to make such an item:

(6.67-2.36)(200)/(0.3)/(0.3)/(2(900)) = 5.321 slots

What this means is that:
1. You have a 67.9% chance to end up with a +22atk or +23atk 2slot BWG
2. You have a 32.1% chance to end up with a +25atk or +26atk 1slot BWG

To understand this further, since these chances are applied for both (perfect+1) and (perfect+2) BWGs, we need to consider that the chance of getting a +4 with a single chaos scroll is modeled by 5/64 x 0.6. The chance of getting a +5 with a single chaos scroll is modeled by 1/64 x 0.6. This means that in addition to the 68% and 32% chances, we can further break up these two chances as follows:

(5/6)(6-(6.67-2.36)(200)/(0.3)/(0.3)/(2(900))) = 0.566
(1/6)(6-(6.67-2.36)(200)/(0.3)/(0.3)/(2(900))) = 0.113
(5/6)((6.67-2.36)(200)/(0.3)/(0.3)/(2(900))-5) = 0.267
(1/6)((6.67-2.36)(200)/(0.3)/(0.3)/(2(900))-5) = 0.0535

1. You have a 56.6% chance at getting a +22atk 2slot BWG
2. You have a 11.3% chance at getting a +23atk 2slot BWG
3. You have a 26.7% chance at getting a +25atk 1slot BWG
4. You have a 5.35% chance at getting a +26atk 1slot BWG

Now, we calculate the average of these two costs for the "marginal cost" of making this hypothetical BWG:

2(2)(900)(6-(6.67-2.36)(200)/(0.3)/(0.3)/(2(900))) + (2)(900)((6.67-2.36)(200)/(0.3)/(0.3)/(2(900))-5) = 3.022b

This is confirmed by the reverse calculation, which is the difference in the cost between the Control and the Experimental Group 1:

(200/0.3+120)/((1+5+10)/64x0.6) - (200)/(0.3)/(0.3) = 3.022b

Data:

Marginal Cost: 3.022b

Table 2: If the Chaos adds +4 or +5

#	Item	Chance (%)
1	+22atk 2slot BWG	56.6
2	+23atk 2slot BWG	11.3
3	+25atk 1slot BWG	26.7
4	+26atk 1slot BWG	5.35

Discussion:

It costs ~3.022b more to get either a +28 atk BWG (83.3% chance) or a +29 atk BWG (16.7% chance).

It's well worth the effort in my opinion if you have that kind of money, because you're only spending ~20.4% more than what you would have spent making that definite +27atk. It's important to keep in mind that the 3.022b is only a calculation on the average cost, which would only apply if many people were trying this and we only accounted for the total costs. Individual costs can vary greatly, and if you do get lucky, you'd score a huge profit.

Now that you know the cost and probabilities associated with this, you should go make one.