Post by nozzle117
From
EJ, 11/3/09.
As for the ongoing crit cap discussion: there's a fair amount of misinformation and confusion present, so let me see if I can shed some light on the situation. First, let's get this out of the way right at the beginning: Fight length has no effect whatsoever on the crit cap. RNG has no relevance to the determination of the crit cap. You don't need to worry about being "near" the crit cap - either you're above it, or you're below it, and that's all the matters.
The important aspect of the crit cap is that it's the point at which it is no longer possible for a white attack to be a normal hit. If you have (say) a 24% glance chance, a 0% dodge chance, a 16% miss chance, and a 60% crit chance, that's 100% of the hit table - there is no chance of an attack being anything other than a miss, a glance, or a crit. So the probability of a white hit is 0%. And regardless of RNG, that will be the case - it is theoretically impossible for you to hit. So with this hit table, if you go hit something for a while, yeah, you may wind up with only a 20% glance rate... but your crit and miss rates for that fight will then be correspondingly higher, as no other outcomes are possible. This is the key point being glossed over: yes, RNG is possible, but these aren't independent events we're looking at. So if one of your outcomes is higher or lower due to RNG, the other possible outcomes must compensate.
Thus, if we're at the crit cap, and white hits are impossible, adding more crit doesn't change that fact - white hits are still impossible, and our crit, miss, and dodge rates are still... whatever they are. Hence additional crit doesn't increase our white damage, making crit rating and agility less valuable - hence, a crit cap.
On the other hand, if we're *below* the crit cap, and white attacks *are* possible, crit rating still does something for white attacks - it reduces the chance of a regular white hit. If our hit table is instead 58% crit, 2% hit, 24% glance, 16% miss, yes, it's entirely possible that we'll have an observed crit rate of 60% on some fight - but those 2% aren't necessarily going to come all out of the hit budget. The constraint in that case is merely that hit + glance + miss rate will need to be 40% instead of 42%, and this can involve more or less misses, more or less hits, and more or less glances.
Now, if our fight is short and our white hit rate is low, it's entirely possible that we won't see any white hits. But that doesn't mean that we *can't*. Sometimes, we'll see none at all. Sometimes, we'll see more than we expect. But on the average, we'll see the 2% (or whatever) we expect. And the important piece is: crit rating still increases the number of crits we expect to see, and reduces the number of white hits. And as long as it's still doing so, it's just as valuable as it always was, and, as such, we're still below the crit cap.
Look at it this way: the hit table is like rolling a 100-sided die. Normally your hit table is something like:
1-58: crit
59-60: hit
61-76: miss
77-100: glance.
Now, if I roll that die a hundred times, I may get more or less than 58 crits. But that doesn't change the underlying table - it merely affects what I see on that one fight. And the important piece is: I will roll some number of 59s. And they will be white hits. But if I had another 1% crit, they would be crits. So that crit rating is doing something for me. On any particular fight I may roll more 59s or less 59s, so the advantage I get from that crit might matter more or less - but on average, I expect to roll a 59 about 1% of the time, so my expected damage is that much higher if I get an extra 1% crit. And that argument applies no matter how close to or far from the crit cap I am, so the value of crit to white attacks does not change with how close I am to the crit cap until I actually hit it.
