Northrend weapon enchant

It's actually quite easy to disprove that then. If you change the next hit on the attack queue to a dodge when mongoose procs and you get hit, then that would guarantee that the system does not work that way. It most certainly does not work like that and would be a pretty poor way to run a supposedly random combat system.
_________________
Shadow Council - US - Horde
Xandreth (74) - Prot Paladin - MT (going ret until 75)
Tyrenn (80) - Frostfire Mage
Saege (70) - Resto Druid in training
Staub (63) - DK backup tank

If it's actually done in the way Obrimos suggests, there would be significantly fewer and shorter streaks involved than a random system would produce.

I could definitely see something like this system being done for quest drops. That's a nice solution to the '5% drop rate, 10 million people doing quest, 1 poor bastard almost guaranteed to need to kill 500 before seeing it' problem.

In that case, you really do not want randomness. You only want the appearance of randomness. And most people do not associate randomness with streaks (when, as most people reading this forum know, streaks are one of the hallmarks of randomness).

A purely random system would likely produce a greater variety in encounter difficulty for the exact same encounters. So you could view the lack of even pseudo randomness as being desirable from a game design perspective.

Obrimos is right, or is uncovering an important point that people are missing.

Agility gives Dodge which contributes to Uncrushability/Block Capping.

Mongoose gives AGI gives Dodge which is not useful for determining Uncrushability/Block Capping.

The entire philosophy of Uncrushability is that by filling up the table with Miss/Dodge/Parry/Block, we will never take a normal hit. We're not looking to get a favorable probability here - we are looking to eliminate the possibilty in an absolute manner.

Because Agility is there all the time, it is useful for determining whether you are block capped or not. Because Mongoose is only up a minority of the time, it is not.

That said, Mongoose does result in more dodges than, say, Potency, and a dodge is generally better than a block. So Mongoose does lead to less damage taken, but agility leads to less damage taken AND contributes to becoming block capped (uncrushable).
_________________

If I recall correctly, and this is debatable as it was back in vanilla wow, some blizzard developer actually posted that they maintain a pool of pregenerated random numbers from which they pull numbers out of as needed. The pool is constantly being refilled from some kind of source which they claimed was not pseudo-random.

In regards to the attack table, a different developer outright confirmed the one roll, one table model on which all theorycrafting has since been based. I'm not sure which specific issues you referred to with rogues, but is is known that player combat mechanics are different from pve mob mechanics and are much more detailed/complicated.

It seems this discussion has shifted wildly from the original topic, if anyone would like to pursue it further perhaps we should start a new thread.

Whether Blizzard's RNG is truly random or not doesn't matter at all, it is statistically random enough, particularly when you only see a tiny subset of them. With the threading, clustering, and multiple tier models they would have, they could easily be relying on a pool of them. If CPU is a constraint, the database is likely the last place you want to be generating them from. It might not be a problem though.

Really, there's a ton of different solutions, particularly with their resources that are available to deal with the design, so it's very hard to speculate from this end. Too much is unknown, making all of this pretty unreliable conjecture.
_________________
http://armorylite.com/us/frostmane/fridmarr/