User blog comment:Krankenheimer/Critical Strikes/@comment-4881935-20140418223214/@comment-1330314-20140424144115

In a case-by-case scenario, dodge and critical strike chance don't look the same, but in the long run they're both percent damage reduction/amplification: 1% of crit is a 1% increase to your autoattack-based DPS, and the same applies to dodge: if you have 1% dodge chance, you'll dodge 1 attack out of a hundred on average, so the average damage dealt to you is 99% of what it would be without dodge. Effectively, 1% dodge equals 1% damage reduction, so the two stats are quite similar.

As for "what else?", the randomness of crit, while being bad itself, also entails a lot of other secondary issues: on one hand, it severely skews gold efficiency when it comes to crit: when you invest in crit, unless you have 100% crit chance you're going to end up with autoattacks that make your crit chance worth zero gold (which you can't predict) and others where you crit and get 5000 gold's worth of stats, far more than you likely put money into. It's the issue of the 1% crit rune in top lane: sure, most of the time it won't do anything, but if you do happen to land a lucky hit you instantly get a lot more gold's worth of power over your opponent. Sure, the gold value averages out over time, but the issue with crit is that every individual autoattack matters, and can potentially define the course of the rest of the game. It also creates a weird itemization curve, where you need to buy a lot of a single stat just so that it becomes remotely useful. Even attack speed brings a reliable bonus when you buy just a bit of it.

The point here isn't that crit should be deleted. It's a defining aspect of marksmen, and the idea of increased autoattack damage isn't bad in itself. It just needs to be made reliable, which is the problem with it right now.