Question About Maximum Experience - English

Placio (atys)

I'm not sure what analysis you did, but the craft actions are independent. In essence if your success rate is 5% than you have a 95% degrade chance on every item, even if you make a million, and because the actions are independent you will have much lower than 5% on average.

I'm not sure you analysis of high levels should be applied to overcrafting; if overcrafting you gain 1.6% for every skill level you get closer to your action level- under your levels hypothesis this would give 1.6x the amount of points stated by the scrolls and craft window. However if the crafter's level is above their action level they only get 0.5% for every level above- this might explain you lower observations, or it was just random.

Finally, scrolls are not some rare thing. they are pretty easy to replace. and the 'minor' xp improvements can add up, because a boost of 8 or 10% might not get you your full action q but will increase the q of the majority of degrades.

The analysis on my part involved several thousand crafts using q60 scrolls, the highest level that is "easy" to get. Daomei's analysis involved a similar number of crafting actions using q70 scrolls, because she is even more bloody-minded than I am. The percentages matched the 0.5% per level figure that you note.

Your analysis of the statistics, however, is flawed. The actions are indeed independent, but a 5% success chance means that *on the average* one crafting action in 20 will be a complete success. This will be true whether you craft a thousand times or ten thousand times. There is no reduction in percentages for large numbers of actions; on the contrary there is a closer and closer approach to the 5% value as random variations even themselves out.

In addition there is no indication that the potential frequency of a certain amount of degrade if there is not success is affected by the level of crafting skill. As a master I have only a 15% chance of degrade when crafting at q250. However, I seem to have just as much chance of producing a q200 as I have of producing a q245 if I do fail, and this does not change even when I am using scrolls.

Without looking at the code I cannot be certain, but it very much looks as if the calculation is :

xx=rand(1,100)
if xx<=success_threshold
then q_result = q_attempt ; (success!)
else
yy = rand(1,100) * 0.20
q_result = q_attempt * (1 - yy) ; (degrade by random amount up to 20%)
endif

Whether or not there are extra decimals involved in the calculation is a detail.

The real question in my mind is whether or not the increase in success factor will produce a decrease in time to master that is significant in terms of requiring significantly fewer bags of materials or hours on line. Obviously it will have some effect, but I doubt very much that it is a large one.