Actual Numbers
Generic brutal rogue with 18 dexterity and 14 strength, using a dagger, with the backstabber feat, with combat advantage, using piercing strike, attacking a target with a 14 reflex: Expected damage per hit is 17.5. Expected damage per critical hit is 24. Chance of hitting is 80% with an additional 5% chance of critically hitting. Overall expected damage is
therefore 15.2.
Generic ranger with 18 strength using two bastard swords with bastard sword proficiency, with combat advantage, attacking his quarry with twin strike versus an AC of 16: Expected damage per sword is 5.5. If at least one sword hits, an additional 3.5 expected damage is added in. If at least one critical occurs, that's 6 instead. I'll save you the math, since it gets really involved, but the overall outcome is going to be 11.587875.
Generic warlock, assuming similar stats and combat advantage, assuming Implement Expertise as a feat since I don't know what else to give them, and assuming Prime Shot for good measure, using Eldritch Blast versus a cursed target with a 14 will is going to hit on a 6+. With a 70% chance of hitting, 5% chance of critically hitting, and expected damage per hit
of 1d10+1d6+4, we've got a total expected damage adjusted for accuracy of 10.1.
Generic avenger with 18 wisdom, using a fullblade with the fullblade proficiency feat, with combat advantage, using any generic at will power he chooses, attacking an oath target with a 16 AC: Expected damage per hit is 10.5. Expected damage per critical hit is 22.5. Chance of hitting requires some math, beginning with a 70% chance of hitting in some fashion per attack roll, and therefore an overall chance of hitting of 91%. Inside that chance is a chance of hitting critically of 9.75%. So the chance of hitting without a critical is 81.25. Overall expected damage is therefore 10.725.
The sorcerer is a tough one, because he has area of effect and multi target at will attacks. If we're only counting single target hits, Acid Orb is probably the go-to choice, and its lame. Its a 1d10 attack, like the warlock above, except with additional damage equal to your secondary stat instead of +1d6. That's usually a reduction, since 1d6 averages to 3.5. And you don't get Prime Shot. And you are ill equipped to go flank in melee to get a flanking bonus, so you probably won't have combat advantage like a wise fey or infernal pact warlock. On the other hand, you have at will attacks that hit multiple targets, and that can dramatically increase your overall damage. But since there's no good way
to quantify the advantage or disadvantage of hitting multiple targets versus single targets, I'm going to leave the sorcerer out.
The barbarian is also a tough one, since rage changes things significantly and the amount of time you spend raging increases as you advance in level. I'm going to leave him out.
So our totals are
Rogue: 15.2
Ranger: 11.587875
Warlock: 10.1
Avenger: 10.725
But what does that tell us? Well, the first thing we should note is that it leaves some other details out.
Rogue: The rogue's attacks grow very poorly as he adds [W]s. He shows the best in comparisons of low [W] at will attacks where his large sneak attack damage can overwhelm the higher damage dice other classes receive. When he has no combat advantage he suffers an enormous in effectiveness, though this rarely happens in most groups. He has no class abilities or effects which add any other details to this comparison. This comparison is probably the perfect storm for the rogue- there may be no other comparison possible that would make him look any better than this one.
Ranger: The ranger's attacks grow well with the addition of magic items and feat damage, which will apply to both hits. He has no other relevant class abilities or effects.
Warlock: The warlock in our example is a lucky warlock indeed to have both combat advantage and prime shot. A real warlock would probably not have this, unless it was an infernal pact warlock shielding himself with temporary hit points while casting spells at point blank, or a fey pact warlock doing the same with eyebite. The warlock's damage is likely lower than this in actual gameplay. The warlock has no other relevant class abilities or benefits except for his pact boon if his cursed foe dies.
Avenger: The avenger in our example has significant benefits that are not listed here. First, he is the only character using an attack that does something other than pure damage. He may be shifting himself and sliding his foe about, or granting attack bonuses to his allies, or some other utility benefit. He also has an automatically applied class ability related to his oath that is included, such as bonus damage if his foe flees, or bonus damage if he is doubleteamed by his enemies. He also has the highest AC present, and possibly the highest hit points, though the ranger's free access to Toughness will mitigate that at lower levels. His attacks grow the best as [W]s are added.
So... I think he's fine. The math seems to hold that up. If we were to redo the comparison at a higher level, not much would change. The rogue would drop back a bit, as his incredible frontloading would recede in importance and as he swiftly ran out of feats to improve his damage. The ranger would gain a bit in damage as feats like Weapon Focus add static damage twice per round for him. The warlock would stagnate a little, unfortunately, but at the same time gains some significant utility effects. The avenger would grow slightly in damage, but not quite at the same rate as the Ranger, but with the addition of Armor Proficiency: Leather and Improved Armor of the Faith he will match the paladin for armor class and leave the rest of this group in the dust. Since I don't believe in evaluating any role, even that of striker, purely on one dimension, I am inclined to think that these are fair tradeoffs.