I enjoyed the recap and your observations. Thanks!
BTW, what kind of dragon was it?
White dragon.
I don't understand this, though. The rogue could just stand back and make ranged attacks with sneak attack damage (+4d6 at 8th level) every round. Stealth would be nice for the advantage, but not necessary. It seems like that should be pretty effective, as long as one of the rogue's compatriots is in melee with the dragon. He/she also has a good shot at being unaffected by the dragon's breath weapon.
I was wrong about that part. A rogue could be very effective if the party planned to make him so. He might be even more effective than most of the party. I was wrong about that after doing some thinking. In our particular scenario, the rogue would have had a hard time given no one made it into melee. Doing an ambush, the rogue might have been amazing if he had hidden farther away than the dragon can see with blindsight. I'm very wrong about the rogue.
I'll explain some of the difficulties the rogue has to face, which planning can eliminate:
1. Assassin archetype: Rogue can't get surprise within range of a dragon's blindsight. That eliminates the assassin archetypes best opening ability: critical hit sneak attack on the surprise around. He could do it with a bow from far away by ambushing the dragon when he exits the cave.
2. Sneak Attack: When no one makes it into melee, very hard for the rogue to sneak attack. A rogue can sneak attack by gaining advantage on attack rolls or when an ally is within 5 feet threatening the dragon. In our particular scenario, neither of those events happened.
3. Dragon was a white dragon. Breath weapon was a Con save. That would have hammered the rogue. Most cold attacks are Con saves. Rogue is not good against cold attacks until he picks up Resilient Con.
After spending some time thinking of how my own rogue would handle the scenario, planning could have led to a win:
1. Rogue is a great switch hitter. He can use ranged attacks with bow and switch to melee with no loss of efficiency.
2. With Uncanny dodge he could have halved the damage of the dragon's strongest melee attack. This would have made it easier to heal.
3. He could have kited the dragon or stayed hidden if over 100 feet away, while still attacking. Put a rogue in a forest full of trees where he can hide, anything will have problems winning.
4. He would have done effective single hit damage while kiting.
Rogue maybe very good against dragons. Only problem being his fighting style wouldn't have been great for an immobile fighter and paladin. If the fighter and paladin had been able to get into melee, he would have done very well. I was flat out wrong about rogue effectiveness. They would be better than a melee based fighter against a dragon, far better. If the dragon had a Dex-based breath weapon, way, way, way better. That breath weapon ate us up.