I think what's overlooked in optimization all the time is:
1) Encounters end. When they end, it doesn't really matter how many hit points you have left. The earlier the encounter ends, the lower the chance of losing hit points, and thus dying.
2) Fun. Hitting more often is more fun than dealing another point of damage. Conditions prevent you from hitting, or using your most effective attack. Again, since encounters have an end, you want to use the rounds you have as effectively as possible. Ending most conditions whenever you want allow you to do so.
3) Some conditions deal 5 ongoing damage. Investing 10 to prevent this means you prevent the first 5 damage, and avoid the 50% chance of losing another 5. If I understand math correctly, you therefore invest only 2.5 hit points on average. Of course if the ongoing damage is 10, it becomes a no-brainer.
4) Let's assume a condition immobilizes you. As a melee fighter, with no enemy adjacent (because all the critters shifted away from you) you are left with your inner rage and your backup javelin to throw. One round lost. You could have dealt 60 damage with your daily, but you don't. The monsters live longer, will hit you more often, and thus most likely deal at least 10 points of damage. Having the ability to invest 10 hit points to prevent this outperforms almost every other optimization if there is at least one condition per encounter. Again, encounters end.
The only reason I see to not go Dreadnought is if the alternative path is to have multiple attacks per round. In that case, the multiple attacks also increase the chance of ending an encounter earlier, perhaps earlier than the dreadnought is able to.