Expected number of rounds per encounter?

[Corran]

So, does it say anywhere in the DMG if there is good estimate for the number of rounds that most encounters tend to last for?
I am trying to piece together a dpr comparison and this info would be quite handy right now. Thanks in advance for any answers!

[Ninja_Prawn]

So, does it say anywhere in the DMG if there is good estimate for the number of rounds that most encounters tend to last for?
I am trying to piece together a dpr comparison and this info would be quite handy right now. Thanks in advance for any answers!

I don't know of anything official, but the consensus seems to be that most fights are decided (if not always over) within the first three rounds.

[DivisibleByZero]

I have no idea where NP got that "concensus" from.
Most combats in our games last somewhere around 6 to 8 rounds, depending on the number and types of monsters being fought.

[Fflewddur Fflam]

Depends on if any of your PC's have Great Weapon Master or Sharpshooter feats.

Wish I was joking....

[EvilAnagram]

Depends on if any of your PC's have Great Weapon Master or Sharpshooter feats.

Wish I was joking....

I really disagree with this sentiment. These are solid features, but they cannot completely undermine a well-crafted encounter.

[Specter]

All of the groups I've played with have fights that range from 2 to 6 encounters.

[MrStabby]

I find about 3 rounds then mopping up covers most encounters. Some may be 2, some may be 5. Some may be longer but it kind of relies on the encounter being engineered that way with reinforcements or a lot of purely defensive abilities.

Party composition counts for a lot as well. With two barbarians, a paladin a warlock and a sorcerer in the party you will (usually) end fights quicker than if you have two lore bards, a wizard and a druid in the party.

[D.U.P.A.]

The maximum should be 10, as many 'encounter' abilities last for 1 minute, which is 10 rounds. Although most of the time it is less than half of that, but that also depends of intensity of the encounter.

[quinron]

The DMG's monster creation rules implicitly assume 3 rounds - it recommends that when creating creatures whose damage output varies across rounds, you should use the average damage dealt in the first 3 rounds of combat (p. 278, under Overall Damage Output).

Considering the DMG uses this to calculate monster DPR, it's probably the best standard to use when calculating DPR in other circumstances.

[BW022]

So, does it say anywhere in the DMG if there is good estimate for the number of rounds that most encounters tend to last for?
I am trying to piece together a dpr comparison and this info would be quite handy right now. Thanks in advance for any answers!

There is no rule...

In terms of practicality. Players take about 30-60 seconds per turn. NPCs more like 30 seconds per turn. This includes calling initiative, declaring an action, moving your figure, making attack roll(s), casting spells, saving throw, adjusting hit points, the occasional lookup, drawing areas on the battlemap, etc. Assuming the PCs are facing say 2 enemies each, you are looking around 5 * 60 + 10 * 30 = 10 minutes per turn. This is highly variable... often faster at low levels (less choices and actions which affect multiple creatures/saves) and longer at higher levels. Assuming a number of creatures gets reduced over time, a typical combat might run 5-7 rounds... likely taking 45 minutes to an hour. Fewer creatures or less choices may make the combat significantly faster... complex areas, opponents with spells, etc. a lot longer. Most combats appear to be designed around 45 minutes to an hour. Now... single or more powerful opponents can easily run 15+ rounds... but typically turns are faster as folks settle into standard sequences fairly quickly and fewer enemies are in initiative.

A good rule might be 5-7 rounds with a typical PC able to focus 75% of the time. 25% of the time, they'll be using actions to heal, drink potions, move, open doors, go defensive, etc.

[Kane0]

Most of my combats last around 3-6 rounds plus one or two for mop-up, with larger fights adding in an extra 2-3 rounds to that. Big multi-part battles will usually be split into these smaller sections.

[Fflewddur Fflam]

I really disagree with this sentiment. These are solid features, but they cannot completely undermine a well-crafted encounter.

Sure......

[EvilAnagram]

Sure......

The key words there are, "well-crafted."

If 10-20 extra points of damage per round completely alter an encounter, the encounter was poorly made. If the DM can't overcome a few HP and adjust the difficulty of encounters to make them challenging and fun, maybe someone else should DM.

[Fflewddur Fflam]

The key words there are, "well-crafted."

If 10-20 extra points of damage per round completely alter an encounter, the encounter was poorly made. If the DM can't overcome a few HP and adjust the difficulty of encounters to make them challenging and fun, maybe someone else should DM.

Always love this one: "The feat isn't broken, the DM is!!!"

Tell me another one.

[JNAProductions]

Always love this one: "The feat isn't broken, the DM is!!!"

Tell me another one.

That's putting words in his mouth. Perfectly good DMs will sometimes mess up an encounter, or just not have their A-Game when designing one, and Sharpshooter or GWM will make the encounter a snap when it wasn't meant to be. But a well-thought out, interesting encounter can easily be so even with SS and GWM on the table-it could be as simple as upping the monsters' HP to accommodate for it, or having reinforcements come in partway through, or interesting terrain that can't be affected by HP damage.

[Isidorios]

Even monsters with a slightly higher AC provide a challenge to characters taking a -5 penalty to hit, I've found.

[EvilAnagram]

Always love this one: "The feat isn't broken, the DM is!!!"

Tell me another one.

I regularly accommodate a Fighter and Paladin who go nova for 60+ damage each at level 8, and the major battles still regularly leave someone on death's door because I balance my encounters against what my party can do.

If a feat that adds ten points of damage (actually less than that on average) broke the game, the game would already be broken. Complain about it if you must, but if a feat with a major drawback to any damage boost is breaking the balance, your DM should start dedicating time to balancing the game.

[Malifice]

A lot will depend on how much firepower your party can bring to bear (how many encounters they expect to face/ how drained of resources they already are) and the difficulty of the encounter.

I generally work around a 6(ish) encounters of around hard(ish) difficulty as a very rough default. Party is currently 11th/ 12th.

Anecdotally, most encounters run for around 5-6 combat rounds.

[Wymmerdann]

Those arguing against the impact of feats/character selection, specifically the example of GWM should broaden their consideration to include the bonus action attacks, which increase the damage output of a level 1-4 character by a significant amount on the turns that it is procured [which should be fairly regularly in any fight including mooks]. Saying that a DM can work around this by selecting encounters with fewer, tougher and better armoured enemies does nothing to undercut the impact of this feat against a run-of-the-mill encounter. This isn't a rant about the system. I love the feat, I use the feat, and I think it behooves us to be realistic about the impact of this feat at the table.

I would absolutely expect this feat to shave a round or two off the length of many standard fights, especially if the character with this feat is playing in a DPR or striker role [just as I would expect the Inspiring leader feat to add a round or two to a fight in which the part gets TPK'd].

[MrStabby]

I think if a player wants to play an aggressive character who feels the best form of defence is a good offence they can. I think it is fine for a player to build a character with the aim of ending combat a turn early through massive damage at the expense of defence or control options.

A bad DM won't let them play the character they want. A bad DM will decide to invalidate their concept by adding HP or otherwise making what sets their character apart unimportant.

I think that there is something wrong with the idea that a good encounter has to be a long one. Grinding out a 10 turn encounter may not be better than a more brutish and short encounter. What matters is the players having fun, the right length is the length that they find rewarding and engaging. It is probably true that some encounters should be longer and some shorter - to give opportunity to both nova and endurance players.

[Giant2005]

Always love this one: "The feat isn't broken, the DM is!!!"

Tell me another one.

To be fair, in the case of those feats, it most certainly is the DM's fault.
Those feats are extremely powerful against low AC enemies, moderately useful against medium AC enemies, and all but useless against high AC enemies.
If a DM is complaining about those feats, it is because he is limiting himself to the lower end of the AC spectrum (or more likely just reacting to the fact that he is noticing the increased damage when they hit, but failing to notice the many times they aren't hitting because of the feat).

[Citan]

So, does it say anywhere in the DMG if there is good estimate for the number of rounds that most encounters tend to last for?
I am trying to piece together a dpr comparison and this info would be quite handy right now. Thanks in advance for any answers!
Not sure if that is the right way to go, but good luck!
As for number of rounds, I'd suggest you consider a mediane of 5 rounds. Simpler encounters will finish in 3, more complex in 6-7... Only extremely tactical encounters may last more (aka, it's in fact several small encounters but chained in such a way you are "in warring state" from start to finish, or an encounter specifically designed to encourage stand-still/attrition tactics).


Depends on if any of your PC's have Great Weapon Master or Sharpshooter feats.

Wish I was joking....

I really disagree with this sentiment. These are solid features, but they cannot completely undermine a well-crafted encounter.
Agreed with EvilAnagram. GWM is very easy to limit, because you have many ways to hamper its efficiency: limiting movement, disarming, blinding, fearing, restraining, etc...
Sharpshooter is much "worse" in this way for a DM, because of the fact that it is a ranged attack which ignores most covers. So you have to completely block his sight or disarm from a distance, or use an ability that inflicts condition from a distance, or Stealth a guy to create an opening from behind... So it's difficult. But it can still be done. Just more painful.
(I'm thinking this could actually be worth a dedicated "DM help" thread, what would people say?).

I also agree on the above comment that not all encounters should be hard, the same as not all encounters should be steam-rolled.
I tend to make usually 3-4 encounters at most (no more because my players are not THAT fond of fighting) with varied style: one steamroll, one deadly, one attrition...
Alternating encounter styles helps keeping tension and pressure on players I find, because it plays with their mindstate...

[Kryx]

To be fair, in the case of those feats, it most certainly is the DM's fault.
Those feats are extremely powerful against low AC enemies, moderately useful against medium AC enemies, and all but useless against high AC enemies.
If a DM is complaining about those feats, it is because he is limiting himself to the lower end of the AC spectrum (or more likely just reacting to the fact that he is noticing the increased damage when they hit, but failing to notice the many times they aren't hitting because of the feat).
Sorry to dive into this rabbit hole further, but if the proposed solution is just raising the AC above the expected levels for CR to compensate for GWM/Sharpshooter then you might as well directly address those issues instead of punishing every player and effectively making that part of the feat a trap.

In your opinion good DMs alter the enemies which impacts every PC and strengthens saving throw classes (read spellcasters) and effectively makes that part of the feat a trap option.
In my opinion good DMs will directly fix the issue without impacting other PCs or creating trap options for PCs.


On topic: I find it best to view encounter duration based on the difficulty. In my spreadsheets I assume an Easy encounter will average 2 rounds, a Medium encounter will average 4, a Hard encounter will average 6, and a Deadly encounter will average 8. That results in about 4.3 encounters per day, with around 25 rounds per day assuming PCs lean more toward hard and deadly encounters. Spreadsheet based on xp tables from DMG (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=2091322934)

[Giant2005]

Sorry to dive into this rabbit hole further, but if the proposed solution is just raising the AC above the expected levels for CR to compensate for GWM/Sharpshooter then you might as well directly address those issues instead of punishing every player and effectively making that part of the feat a trap.

In your opinion good DMs alter the enemies which impacts every PC and strengthens saving throw classes (read spellcasters) and effectively makes that part of the feat a trap option.
In my opinion good DMs will directly fix the issue without impacting other PCs or creating trap options for PCs.

That isn't what I was saying. I was saying that the DM should be creating varied encounters. That means some encounters will be composed of goblins that play to GWM's strengths, while some will be composed of stronger enemies that don't. Variation is a good thing.

[Kryx]

That isn't what I was saying. I was saying that the DM should be creating varied encounters. That means some encounters will be composed of goblins that play to GWM's strengths, while some will be composed of stronger enemies that don't. Variation is a good thing.
So a rephrased version of my critique above: "You can use GWM to smash things that pose little challenge, but come boss time he'll be sure to have a lot of AC". That still poses the exact problems I outlined above - it's a trap for martials and prioritizes saving throws from spellcasters.

[Giant2005]

So a rephrased version of my critique above: "You can use GWM to smash things that pose little challenge, but come boss time he'll be sure to have a lot of AC". That still poses the exact problems I outlined above - it's a trap for martials and prioritizes saving throws from spellcasters.

That is just how the game is. A boss isn't much of a boss if he has mook level stats.

[MrStabby]

So a rephrased version of my critique above: "You can use GWM to smash things that pose little challenge, but come boss time he'll be sure to have a lot of AC". That still poses the exact problems I outlined above - it's a trap for martials and prioritizes saving throws from spellcasters.


Well it needn't be the "boss" that is hard to hit. An encounter where all attacks are at disadvantage vs only moderate armour can do the same and be a strongly thematic encounter vs some not too high level enemies.

But generally I agree with you. It is a feet for someone who wants to sacrifice other things to do more damage. If the feat seems too good it may be a sign that there is too much combat in the game rather than the wrong sort. If everyone is taking that feat not feats like the actor feat then consider running a less combat intensive game (consider not just do - a dungeon crawl is fine if thats what your players want). If someone wants to be good at something, find a way for them to be good at it. If what they want to be good at is cutting down an opponent in record time then why adjust the game so they can't play that character. It doesn't mean subtracting fun from others as long as you let their characters be good at the things they want to be good at as well.

Tying this back to the OP, combat may be quicker if parties optimise for combat. If wizards prepare fireball over invisibility or in the case of feat selection being martial over utility.

A combat CAN be pretty much as long as you engineer it to be, but it can seem a bit contrived, especially if every medium combat involves reinforcements or healing effects, or perfect spaces to run away to and so on.

[mgshamster]

So a rephrased version of my critique above: "You can use GWM to smash things that pose little challenge, but come boss time he'll be sure to have a lot of AC". That still poses the exact problems I outlined above - it's a trap for martials and prioritizes saving throws from spellcasters.

That's not what he said. You've put words in his mouth and changed his argument to make your own look better.

More on note, it doesn't follow that having varied encounters using a spectrum of defenses and offenses will make martials worse and Spellcasters better, nor does it follows that using varied encounters makes certain feats a trap option.

[EvilAnagram]

So a rephrased version of my critique above: "You can use GWM to smash things that pose little challenge, but come boss time he'll be sure to have a lot of AC". That still poses the exact problems I outlined above - it's a trap for martials and prioritizes saving throws from spellcasters.

I would honestly say it isn't a trap. Sure, a boss might have a higher AC to pose a higher threat, but the moment the martial character puts themselves in position to use their abilities (e.g., some buff or debuff mitigates the -5 to attack) they can nova for a solid chunk of damage and feel badass.

The increased difficulty and the pressure pushing them away from using the -5 +10 simply creates the need for tactical creativity, which helps to create a more engaging fight.

And against a run of the mill enemy, or even an enemy with a lot of HP but little AC, the feat is fully exploitable.

[mgshamster]

That is just how the game is. A boss isn't much of a boss if he has mook level stats.

Not necessarily. You can have some good boss fights where the boss itself is a weakling, and his mooks are the stronger ones. Or where the fights leading up to the boss are really tough only to one-shot the boss itself. And you can have the classic one where the boss is the hardest fight in the scene.

More to the point though, you don't have to wait for the boss to "make it tough" and have all the other fights be piddly. You can have varied fights at any time, just by making some have lower defenses, some have higher defenses, some have greater or weaker offense. You can have some that are a total cakewalk for anyone, and some that are so difficult your players shouldn't even be engaging in the fight.

It's about variation. Mix it up.

[Citan]

So a rephrased version of my critique above: "You can use GWM to smash things that pose little challenge, but come boss time he'll be sure to have a lot of AC". That still poses the exact problems I outlined above - it's a trap for martials and prioritizes saving throws from spellcasters.

First, that's not what he said.

Second, I really don't see how such an event could be viewed as shocking anyways. There are many kind of bad bosses in any story, some relies on minions, others rely on themselves. The latter, if built as a martial, is expected to have strong defenses (even casters NPC could have Shield ready).

Third, that does not mean your feat suddenly becomes moot. It just means either you coordinate with teammates to compensate, or you let another one with the right tool/spell take care of this high AC target while you dispatch other enemies.
Which, incidentally, has a fair chance to trigger the "bonus action attack" benefit.
Leading in turn to the fourth point.

Fourth, GWM and Sharpshooter bring other strong benefits, which some could say would make good enough feats even without the -5+10 (for Sharpshooter, one could argue that the real problem is rather "voiding third-quarter cover" AND "no long-range disadvantage" ;)).

Fifth, whatever the kind of encounter you make, unless it's designed to be real easy, there should always been creatures with a strong defense, some AC, some DEX, some WIS... Because it's realistic and forces players to always keep their wits "online".

Finally, I would daresay that bumping the AC or a target as a "-5+10" counter is actually one of the nicest thing a DM could do, because there are many (many) ways to circumvent that. Just less so as a solo player.
Qualifying that as bad DMing would be justified only if a DM used that constantly (aka every enemy of every encounter) because that would indeed punish player and his teammates in the end by shoehorning them in "non-AC" attacks, making fights just tedious and boring. But a few enemies with extra high AC is just normal DMing, if not good.

[Kryx]

If the feat seems too good it may be a sign that there is too much combat in the game rather than the wrong sort.
I think it's a bit unreasonable to put the issue on campaign and PC choice when the system isdesigned to prioritize combat. That is why I split out feats like Actor to be taken in their own progression (I call them Traits).

There's some truth to what you say about design, but it shouldn't be up to DMs to fix the inherent issues in the design of the system (hello 5e!)


That's not what he said. You've put words in his mouth and changed his argument to make your own look better.
I didn't mean it as a trap - I sincerely assumed he was implying so.

The problem with -5/+10 is it is either OP or a trap option with the line between being very thin. IMO it's horribly designed. In a typical game it'll crush minions (typically not too much of a challenge) and fail against bosses.

The problem with your assumption of mixing it up is most GMs won't do so. It'll most likely play out as I put above.

We should stop discussing -5/+10 in a thread about number of encounters. I apologize for my role in that - I won't reply to further discussion so we can hopefully move on. My contribution above that might have been missed in the -5/+10 discussion:

On topic: I find it best to view encounter duration based on the difficulty. In my spreadsheets I assume an Easy encounter will average 2 rounds, a Medium encounter will average 4, a Hard encounter will average 6, and a Deadly encounter will average 8. That results in about 4.3 encounters per day, with around 25 rounds per day assuming PCs lean more toward hard and deadly encounters. Spreadsheet based on xp tables from DMG (https://docs.google.com/spreadsheets/d/1d-9xDdath8kX_v7Rpts9JFIJwIG3X0-dDUtfax14NT0/edit#gid=2091322934)

[MrStabby]

Finally, I would daresay that bumping the AC or a target is actually one of the nicest thing a DM could do, because there are many (many) ways to circumvent that. Just less so as a solo player.
Qualifying that as bad DMing would be justified only if a DM used that constantly (aka every enemy of every encounter) because that would indeed punish player and his teammates in the end, making fights just tedious and boring.

Even this is not that clear-cut (although I do generally agree with you). If the part of the campaign is about fighting a bunch of rogue knights then a lot of platemail is in keeping with the setting and plot (although even then not a problem to work in squishy squires). Used sparingly high AC is good, but it should be done as part of an organic campaign and rarely in response to the builds of your players (unless a dedicated party is being sent out to get them).

If a player builds their character to be exceptionally focused rather than versatile then I don't have a problem with them sometimes being useless. If a fighter takes GWM for dealing with high HP low AC targets then they can also take magic initiate (as just one example) to get sacred flame to assist with things they can't just hit with a sword. If players take abilities they should get to benefit from them, if they skip others they should from time to time wish they had selected them. This is what makes building characters fun - your choices are meaningful.

[Logosloki]

No more than five rounds is a goal I aspire to. The first three rounds are where the bulk of the encounter happens and what happens after that is sweeps, tactical retreats and/or diplomacy. For a big set piece I move to ten rounds at most for all the main action since players are more motivated and it allows for me to throw in some interesting mid act twists. I find that if you keep using long round encounters people get a little tired (unless there is some sort of narrative impetus). This applies equally to combat, social and exploration encounters.

As for everyone's favourite feats...I don't see them as an issue. I already try and vary things as it is so sometimes the conditions are a gift to a player using power attack and sometimes the player has to create their opportunities.

[Plaguescarred]

Expected number of rounds per encounter is calculated by the number of combattants for each side times the average constitution modifier divided by 2. Its not an exact science but it's really good approximation.

[MrStabby]

Expected number of rounds per encounter is calculated by the number of combattants for each side times the average constitution modifier divided by 2. Its not an exact science but it's really good approximation.

Is this the blue for sarcasm thing? Or am I missing something?

So on one example you have 4 goblins vs 4 level 2 PCs. A reasonably quick fight. If we add 4 more PCs and make the PCs all level 10 (both likely to raise average Con) then we expect it to take longer for them to kill the goblins?

[Plaguescarred]

Yes blue is sarcasm :P

[Plaguescarred]

There's no expected number of rounds per encounter as there is too many variables to put down rigorous expectations.

You can have a idea once you know your party and the monsters a bit, but even then, so many unpredictable things can happen in combat that can affect the lenght of a fight that its hard to say really.

[MrStabby]

There's no expected number of rounds per encounter as there is too many variables to put down rigorous expectations.

You can have a idea once you know your party and the monsters a bit, but even then, so many unpredictable things can happen in combat that can affect the length of a fight that its hard to say really.

Yeah. i find myself having to sketch out some rules for my fights as DM to keep myself fair sometimes. If reinforcements will arrive after 3 turns then it will be 3 turns - not when it looks like the fight is beginning to be over. This ensures that things like limited use items deliver full value by ending encounters quicker and getting the breathing space for the PCs that is appropriate. Yes, sometimes it can make planned encounters easy - but that is just wise use of resources.

[gfishfunk]

My players' PCs are not optimally created, and they generally tackle hard encounters in 4-6 rounds, even when they go nova.

[Pex]

I would honestly say it isn't a trap. Sure, a boss might have a higher AC to pose a higher threat, but the moment the martial character puts themselves in position to use their abilities (e.g., some buff or debuff mitigates the -5 to attack) they can nova for a solid chunk of damage and feel badass.

The increased difficulty and the pressure pushing them away from using the -5 +10 simply creates the need for tactical creativity, which helps to create a more engaging fight.

And against a run of the mill enemy, or even an enemy with a lot of HP but little AC, the feat is fully exploitable.

Alternatively you can have the BBEG be such the character needs to use those feats because his offensive abilities are devastating so he needs to brought down fast. Doesn't even necessarily need high hit points. The threat of his offensive abilities is enough to warrant take down as fast as possible.

[Xetheral]

There's no expected number of rounds per encounter as there is too many variables to put down rigorous expectations.

You can have a idea once you know your party and the monsters a bit, but even then, so many unpredictable things can happen in combat that can affect the lenght of a fight that its hard to say really.

I agree with this completely. Even something as simple as the enemy making tactical retreats can double or triple the length of an encounter depending on if (and how) the PCs decide to give chase.

[CursedRhubarb]

My group has had a wide variety of encounter lengths. I'm not sure you can put it down to a specific range because we've had a fight of our 4 guys vs 6 enemies, and it took about 15 rounds because everyone, even the DM, rarely rolled over a 10 and the enemies seem to almost always have around 16-18 AC.
We've also had ones where we took on 20+ zombies and finished in 3 rounds due to AoE and good rolls.
Environments also effect it. Trees have been a blessing and a bane due to needing LoS and taking potshots while trying to get in position while the DM has the enemies use ready actions until we get close then we wind up in a n overlapping Spirit guardian/sanctuary deathtrap so we spend more time using movement actions than attacks.

It tends to come down to encounter build and if your dice like you or not that day.

[EvilAnagram]

There's a lot of talk about HP and AC adjustment, but those are really only the beginning of how you adjust an encounter to fit the players.

Example time:

At the time of this particular battle, my players were a Vengeance Paladin 6/Sorcerer 1, Eldritch Knight 7 with Polearm Master, and Valor Bard 7 with a necklace of fireballs. Thanks to their abilities and magic weapons, they have very high damage ceilings when they nova, but the Paladin and Bard both have 12 Con and moderately low HP for their level.

The combat they entered started out with a running battle with some Lizardmen. The Lizards were mostly standard, but they maintained a fighting retreat to take advantage of their ranged javelin throws and set up an ambush. This kept the players from sneaking up and fireballing the all of them at once, and more importantly were sensible tactics for the situation, fulfilling both game and story justifications for their actions.

Once they sprang ambush, the players took some initial damage, but were able to rebound and slay them. Unfortunately, the Young Green Dragon (CR 8) that had been using the Lizardmen as minions objected to this, and they ended up fighting him in a relatively open area. I had adjusted the dragon's HP up to ensure that the fight would be worthy of the buildup, and I started him out flying in the air and out of reach of the Paladin's smites and the Fighter's magic glaive.

So here we have a high HP, moderately high AC enemy with a powerful ranged attack and flying speed against a party with middling ranged capabilities. Odds were intentionally stacked against the players because the rewards for this fight were quite substantial, including magic items and a level increase. My players, expecting challenging fights, immediately set about trying to solve the problems in front of them.

Round 1

The players all prepared actions in case of hostile action on the part of the dragon, and the Paladin intentionally goaded the dragon into melee combat while placing his Vow of Enmity on it.
Dragon swoops down, triggering attacks and taking decent damage from the Paladin's smite. All three attacks hit and deal solid damage to the Paladin, who goes down.


Round 2

Bard heals Paladin.
Paladin attacks again.
Fighter positions himself to attack the dragon and novas, blowing his action surge and hitting five times, dealing nearly 100 points of damage thanks to some very lucky rolls and a crit.
At this point, the dragon is bloodied.
Dragon attacks Fighter and flies up forty feet.


Round 3

Bard and Fighter make ranged attacks, Bard heals.
Paladin, unable to smite, pole vaults and misty steps up to dragon, grabbing onto it.
Dragon, Paladin in tow, delivers breath attack on Fighter and plunges deep into nearby lake.


Round 4

Fighter uses Chromatic Orb, Bard uses Dissonant Whispers. Despite disadvantage, Fighter hits.
Paladin smites one last time despite disadvantage, killing the dragon



So let's analyze this. I knew they were weak defensively, but strong offensively, so I adjusted the enemy's HP to keep them from killing it before the fight got interesting, and I included a preliminary battle to drain resources before the fight. Despite this, they managed to severely reduce the dragon's HP by the second round thanks to very lucky rolls. Knowing this, I used my second means of prolonging the encounter and took to the skies again, stopping the entire third round from yielding any profitable results for the party. Despite this handicap they still managed to think creatively and make awesome decisions (teleporting to grapple a dragon in mid-air).

Note that up to this point, every time the dragon had turned the field to its favor and negated the uses of the PCs' most powerful abilities, it merely prompted my players to act creatively. The Paladin used his charisma to goad the dragon into attacking on the ground and teleported to it in mid-air when it took back to the skies. The Fighter used his positioning to take advantage of a boon giving him advantage on attacks when he had no conscious allies in a certain radius. The Bard used spells that kept the others in the fight and played off of their strengths.

Also note that the dragon was tough enough to survive some lucky rolls, but it was not so heavy a hitter that the lucky rolls couldn't significantly reduce the number of total rounds. Had luck not been so firmly on the PCs' side, the battle could conceivably have lasted until the end of round five or start of round 6.

For its last actions, the dragon then stuck to its principal priorities: pettiness and self-preservation, and it used its breath attack before plunging into the nearby water to keep everyone involved at a distance. At this point, the battlefield was once again entirely in the dragon's favor, despite its health. Had the disadvantage worked out at all, it could have easily knocked out the Paladin or made another breath attack on the now-clustered Bard and Fighter. However, some lucky rolls once again turned the tide, and it fell towards the end of Round 4.

Still, careful balancing meant that the dragon was a considerable threat at every point of the battle, and its flight and swim speeds allowed it to alter the terms of the battle and force the players to think outside the box. Even with a ton of luck on their side, it took four rounds of consistent damage to bring it down, and they could last through five or six with luck against them.

Note that this is all immediately after seven rounds of running combat.

Now, bringing it all back to the question of how long combat should last... I forgot to come up with a meaningful conclusion. Feel free to make one yourself!

EDIT: Now I remember! My point was that the DM has a lot of tools in his kit when it comes to prolonging encounters while keeping them engaging. You don't have to simply rely on things like increasing AC and HP, but you should feel free to occasionally force your players to act outside of their optimal zone because it encourages creativity. Things like keeping opponents mobile (with the Lizardmen), setting up ambushes, or using the terrain can lengthen encounters, while allowing your players the opportunity to ambush their enemies can drastically shorten them. I don't think there is really an optimal zone, though there may be an optimal number for a given encounter (i.e.: They should drop these mooks in 1-2 rounds, but this badass should keep them on their toes for 4+).