I've wrote (again) a long answer that got ditched because of the crappy connection. Short:
I didn't noticed the 5 feet disadvantage, but that doesn't really affect the account, because the original 34 didn't take it.
In a featureless, empty white room without positioning or companionship, my base damage with lance is as follows-
6.5 (average damage) + 6 (strength + dueling) x 2 (extra attack): 25
4 Superiority Dice. 4.5 average damage.
Action Surge four double attacks
Average fights during 4-5 rounds
CR 6 Av. armor Class: 15, hit points: 146-160. I hit with 8, average attack result is 11.
Round 1: 25+9 = 34
Round 2: 25+9 = 34
Round 3: 25x2 = 50
Round 4: 25
Round 5: 25
Total damage in 5 rounds: 168; DPR 33.6 (34). Total Damage in 4 rounds: 143. Average DPR 35.75
In a featureless, empty white room without positioning or companionship, my base damage with sword is as follows-
4.5 (average damage) + 6 (strength + dueling) + 0.45 (10% crit) x 2 (extra attack): 21.9
4 Superiority Dice. 4.5 average damage. Assume 1 crit in 4 rounds for extra 4.5.
Action Surge four double attacks in 1 round.
Average fights during 4-5 rounds
CR 6 Av. armor Class: 15, hit points: 146-160. I hit with 8, average attack result is 11.
Round 1: 22.9+9 = 31.9
Round 1: 22.9+9 +4.5= 36.4 (I put the critical hit SD damage here, although it could be in any of the first 4 rounds)
Round 3: 22.9 x2 = 45.8
Round 4: 22.9
Round 5: 22.9
Total damage in 5 rounds: 159.9; DPR 31.98 (32). Total Damage in 4 rounds: 137. Average DPR 34.25
A short rest is needed after every complete fight, but DPR sustains. Not a bad assumption, especially when I'm not counting anything that could possibly benefit me, such as terrain, OAs (10.95 or 12.5 extra damage as a reaction), basic poison (2.5 extra damage for one minute, PHB), the fear caused by Menacing Strike (possible routings or withdrawals of enemies... unless undead), alchemist fire to the blade (let's say +1 fire damage, the same as a torch does), hot coals on the ground, spikes, caltrops (1 piercing damage when you step in them, probably more if you fall prone on them -say 1d4 or 2.5-), magic weapons or falls from mountains, bridges, windows or rooftops that automatically win a battle.
Lance is still RAW and RAI a dueling weapon. You need to weild it two handed when you are on foot, because is a horseman weapon, not because is a two handed weapon. It is the weapon of choice in jousts, that is duels between knights, and it is always one handed then; their reach and damage are to reflect the fact that you are on horseback (in fact, a sensible ruling would be that you can't even attack a prone creature on horseback if you have not a lance; and I must add that a knight should have the choice of using the horse trample instead of its weapons, but I'm not advocating this, because it is already powerful as it is).
And battles are never in a vacuum unless you have a crappy GM. Terrain, cunning, tricks, diplomacy, intimidation, cover, morale and such should play a heavy role on combats. Surrender, capture and retreat are legitimate ways to end a fight for the losing side. People should be aware that goblins won't fight to the last men alive, unless there is a very powerful reason. Only fearless or very desperate creatures should fight to the end. Wild animals, intelligent foes and monsters should retreat when they are badly injured or they think that they can't win, specially when their main force is dead or badly injured (killing spellcasters work wonders on their morale; and a dragon won't die over a pitiful fight: he is an intelligent enough monster to retreat and return when not expected).
About the horse thing: Most of my fights are on horseback. Not everyone, of course, but most yes. If a troll enters in a dungeon, a horse too. And if they not, luring the enemy outside does the trick. And the troll is at disadvantage. And yes, they die easily, that's why I have four horses and I'm making a plate barding for all of them.
And you centered on one piece of my argument, ignoring the sheer versatility that comes for a good build, even the "boring fighter": 6 skills, one of them with Expertise (athletics), 3 tools (Smith, Carpenter, Tinker), 3 languages, and heavy use of equipment such as crowbars, caltrops, and traps beat a white room theorist. As I've said earlier all that I've mentioned is sustained during real gameplay. Multiclassers are often beaten by dead levels, tier feature delays, lack of ASI to remain competitive, and bogged down in real play because they are often centered on an objective instead of a path. They suffer until they reach their desired synergy power, and then they get dissapointed when said power isn't all that great. I've witnessed this multiple times. I'm not saying that it is impossible to make it work, I say that is really difficult and not all that necessary.