Argothair
2018-08-30, 02:00 PM
D&D is a fantastic game that's bogged down by having too many rules that don't accomplish anything important. I know the designers have spent thousands of hours playtesting the 'balance' so that you can send a Level 8 party against a Level 8 encounter and have a statistical expectation that the party will be wounded but survive -- and I don't think that kind of balance is nearly as important as making the game simple, intuitive, and fun. If your Dungeon Master is halfway decent, she'll be able to adjust difficulty on the fly, by adding more monsters or having monsters run away or bringing in NPCs or gifts or cursed loot or divine blessings. So these are my suggestions for how to get the rules right, in the hopes that the rest will follow:
1. Tie ability scores directly to game mechanics.
2. Always let players roll their own dice.
3. Replace spell slots with a mana pool.
4. Use 3d6 instead of 1d20.
5. Give players a clear choice of tactics in combat.
1. Tie character stats directly to game mechanics.
D&D has arguably the best set of core character stats ever developed: Strength, Dexterity, Constitution, Intelligence, Wisdom, and Charisma really do an excellent job of describing a fantasy character's abilities. They fit together well, they don't overlap too much, and they give you an interesting set of tools for thinking about why characters are different from each other.
The problem is that instead of just *using* these character stats, D&D asks you to put them through a series of Rube Goldberg exercises before the stats take on any meaning in terms of the game mechanics. If you want to know how many points of damage you can absorb, you've got to take your Constitution stat, subtract 10, divide by 2, round down, multiply by your current character level, and then add that value to the total shown on a number of dice of varying sizes that's one less than your current level. What? No. Bull****. This is a game, not math class.
A better way to set this up would be to...say that you can absorb one point of damage for each point of Constitution. Got a CON of 14? Congratulations, you're super-buff and you have 14 hit points. Got a CON of 8? Man, that sucks, you only get 8 hit points. If you really want characters to become invincible as they level up, you can add a flat bonus per level -- maybe your hit point maximum is your CON stat plus your current level. That's it. Keep it simple. All that math isn't adding anything to the gameplay value for 95% of the players, and the handful of geniuses who actually like this kind of thing would rather see the complexity applied to something more interesting, like spells.
Speaking of which...your Intelligence can be the number of spells you're capable of memorizing when you prepare your daily spells. Got an INT of 15? You're a smarty-pants, so you can learn to cast 15 different spells before you start forgetting your old cantrips. Your Wisdom can be the total mana cost of spells that you can simultaneously concentrate on maintaining in any given hour. Got a WIS of 13? Great, you can maintain up to 13 mana points' worth of buffs for your allies without slipping up and letting the oldest spell fade away. Your Charisma can be the target that opponents need to beat in order to break free of your magical charms. Got a CHR of 17? Guess those Orcs will need to roll a 17 or better to shake off your extremely persuasive Sleep spell.
Dexterity can represent the difficulty of hitting you in combat when you're not wearing armor. Got a DEX of 14? Cool; you're reasonably shifty, so your opponents need to roll a 14 or better to hit you. If you're wearing leather armor that gives you a +2 to dodging, they need a 16 or better to hit you. This stuff should not be hard. If I never hear the phrase "Armor Class" again it'll be too soon.
Strength can represent the number of different things you can carry, or the number of kilograms you can carry, depending on whether your friends are capable of thinking in the metric system. Got a STR of 15? Cool, that's pretty strong; you can carry 15 things or 15 kilograms (40 pounds) of weight on your back and still be able to thrust and parry and dodge for an hour and then hike 10 hours over a hilly trail.
2. Always let players roll their own dice.
For some reason, D&D 5e tells players to keep track of the differences between "skill checks," "saving throws," "attack rolls," and "magical attack rolls," like these are somehow four separate things. Bull****. As the Angry GM teaches, there is a player, there is something she wants to do, and there is somebody or something that makes it possible for her to fail. That's all. The active player -- the person trying to do something -- gets to roll dice. The DM sets a difficulty target based on how hard the thing is, and if the player rolls equal to or better than the difficulty target, she succeeds. Otherwise she fails.
While this is happening, nobody else gets to roll any dice. They can roll dice later, when it's their turn. That's how taking turns works.
Also, saying that your character is "dodging out of the way of a poison arrow" or "not getting flame-broiled by a Red Dragon" or "resisting the seductive call of the magical Sirens" does not count as "trying to do something." None of these are excuses for wrecking the turn order and taking an extra turn. The person trying to do something in these three examples is, respectively, the orc archer, the Red Dragon, and the mermaid Sirens, respectively. They're probably all controlled by the DM. Let the DM roll dice for them. You can roll dice when it's your turn again. If you really can't wait, consider investing in a fidget spinner.
3. Replace "spell slots" with a mana pool.
Nobody other than D&D players has any idea what a spell slot is or how it's supposed to work. They don't match up with anybody's intuitions, and even after someone explains the mechanics of the rules to you, they still don't make any sense. What does it feel like to be a Wizard who has three Level 2 slots left and one Level 4 slot but no remaining Level 3 slots? I have no frigging idea. Obviously magic isn't real, but this kind of magic isn't even a thing that shows up in books or comics or movies about fantasy worlds. Thor doesn't have spell slots. Gandalf doesn't have spell slots. Stannis Baratheon doesn't have spell slots. Paskennarion doesn't have spell slots. Rincewind doesn't have spell slots. Glinda the Good Witch of the North doesn't have spell slots. Why should your D&D character have spell slots?
Trick question: your D&D character should not have spell slots. Your D&D character should have a mana pool. Mana is easy to understand because it matches people's intuitions about how fantasy magic works. Everyone "knows" that when you cast powerful spells you temporarily exhaust your magical mojo, and when you go to sleep or take a break from spellcasting, your mana slowly replenishes. So if you want to write rules that people can understand, go with what people already know.
There are plenty of ways to make mana pools work, but as one example, let's say your character gets 1 mana point for each point of Charisma, plus 1 mana point for each tier of Spellcasting that she has access to. Full casters like Wizards get a new tier of Spellcasting after every 1 or 2 levels; partial casters like Paladins or Rangers only get a new tier every 3 or 4 levels. So, a level 5 Wizard might have access to Tier 3 casting and a Charisma of 12, which would mean they have a mana pool of 3 + 12 = 15 mana points.
Spells are ranked by tiers, with "cantrips" treated as Tier 1 spells, and the most devastating ultra-bombs like Time Stop and Resurrection and Wish topping out somewhere around Tier 12. Each time you want to cast a spell, you've got to pay 1 mana for each of the tiers involved, adding the value of each tier together. So, a tiny cantrip like Acid Splash is a Tier 1 spell that costs 1 mana. Sleep would be a relatively weak Tier 2 spell that costs 1 + 2 = 3 mana. Fireball would be a moderately powerful Tier 4 spell that costs 1 + 2 + 3 + 4 = 10 mana. Raise Dead is a awesome Tier 6 spell that costs 1 + 2 + 3 + 4 + 5 + 6 = 21 mana. You get the idea.
If you don't have enough mana to pay for your spell, then your mana dips into the negative, and you've got to beat a difficulty check with a target equal to your current mana debt. So, let's say you want to cast Fireball (10 mana cost) but you've only got 3 mana left. Now your mana pool dips down to negative 7, which means you've got to roll your dice and get better than a 7, or else the spell fizzles or explodes or hilariously backfires. If you want to cast *another* fireball before resting, now your mana is at negative 17, which means you need to roll your dice and get better than a 17 to avoid serious consequences. Beating those checks quickly becomes impossible, so you've got to stop casting *eventually* -- but you can press your luck if you want, dig deep for a little bit of extra mojo, instead of always standing around with your hands in your pockets after 3 pm because you're "out of spell slots."
You recover mana at the rate of...1 mana point per hour. This is, coincidentally, the same rate at which you recover hit points. None of this "short rest" or "long rest" garbage that has everyone gaming the system to figure out how to spring back to 100% capacity while sleeping on the moldy floor of a cursed dungeon. The longer and more often you rest, the more you'll recover. You can rest for as long or as short of a period as you like, just like real life. If you take too many naps in the wrong places, you risk being fired by your boss and/or a Red Dragon, but that's just business as usual.
4. Use 3d6 instead of 1d20.
I know, I know, the d20 is iconic; a literal symbol of all that D&D stands for. They're fun to fidget with, and fun to roll, and they work well for Initiative, which comes up nearly every session, and which ought to be mostly random. For other skills and checks, though, d20s are mechanically terrible, because they have such a huge, flat probability curve that a +1 bonus doesn't have any noticeable effect. Like, suppose you have an 55% chance of accomplishing a task instead of a 50% chance of accomplishing the task. Would you even notice the difference? Statistics says no. If you flip a fair coin 100 times, it usually won't come up exactly 50 heads and exactly 50 tails -- it's more likely that it'll be slightly biased in one direction or the other. In fact, the 'average' amount of bias is that you'll get either 55 heads or 55 tails. This means that your +1 bonus gives you literally the same amount of variation you'd expect to see by chance when you make the same roll 100 times.
How often do you actually go and make the exact same roll 100 times in D&D? Let's say it's a really basic roll, like attacking with your sword against a perfectly ordinary goblin. You make the roll 5 times per combat, you fight in 3 combats every 4-hour session, and it takes you 3 sessions to gain a level, and you gain 3 levels before you get any new abilities or equipment that affect your combat roll That's still only 5 * 3 * 3 * 3 = 135 rolls...barely enough to notice a +1 difference, even when you're doing nothing but hacking relentlessly away at the goblin hordes for 36 straight hours of play. In more realistic circumstances, where you fight a variety of enemies using a variety of tactics, you probably would never make the same roll more than 10 times in a campaign, which means you wouldn't even notice a +4 difference in skill. Sure, it *feels* cool to have a fat bonus, but it's all in your head. Mathematically speaking, whether you succeed or fail on any given roll, or even on any given ten rolls, has much, much more do to with luck than with your exact skill level.
That doesn't have to be the case. The problem with the d20 is that it's flat -- you have the same probability of rolling any number. If you switch to 3d6, though, then certain numbers are much more likely than other numbers. It's a lot easier to roll a 10 on 3d6 than it is to roll a 3 or an 18. This means that even small bonuses can have a big effect. Let's say you're trying to whack that goblin again, and you would normally hit on a roll of 11 or better. Using either 3d6 or 1d20, you've got a 50% chance to hit. Now I hand you a magic sword that gives you +2 to all your attack rolls. Using the 1d20, that only boosts your chance to hit from 50% up to 60%. Not much difference. Using 3d6, though, the same +2 sword boosts your chance to hit from 50% all the way up to 74%. That's a big change that you're likely to notice. With a +5 sword, the d20 would still only give you a 75% chance to hit...but the 3d6 would give you an impressive 95% chance to hit.
The probability curves for 3d6 still deliver on D&D 5e's promise that no hero will ever be invulnerable against a swarm of hundreds of goblins, even without dipping into edge cases like critical hits or natural 20s. A goblin *could* roll triple 6's and score 7.5 points higher than the expected attack roll, just like a goblin *could* roll a 20 and score 9.5 points higher than the expected attack roll. The difference is that the crazy, implausible events are *actually rare* when you use 3d6. If you use 1d20, you get supposedly miraculous events all the damn time, to the point where attributes and proficiencies start to lose most of their meaning. If a Level 15 Barbarian, a Level 1 Wizard, and an irritated NPC goat all have a decent chance of kicking you in the shins, then why bother keeping track of stats for different kinds of characters? You could just skip the character sheets and roll totally random dice every time there's a combat. Or, you could use 3d6, and have die results that actually correlate with your character's abilities.
5. Give players a clear choice of tactics in combat.
In theory, D&D players can do all kinds of different things in combat, limited only by their imagination and their core attribute stats like Dexterity and Strength. In practice, almost nobody ever figures out how to do anything other than swing a sword, cast a fireball, drink a healing potion, or run away. The rules for what you can do in combat are buried in a long, rambling encyclopedia of every possible action, and they don't make it clear why you would want to give up your attack to futz around with shoving or grappling or opening doors. There's also a horribly unclear set of rules about surprise attacks, readied actions, and passing your place in the turn order. Do you have to go when it's your turn, or can you wait till later? Can you take two turns in a row if you're sneaky enough? When? Why? Nobody really knows.
So, let's fix this. In combat, you go only on your turn. You never go on anyone else's turn. If enemies are surprised, that means they get a temporary penalty to their stats, not that they lose their turn. Everybody gets one turn, and there's a fixed turn order in each combat, and it almost always goes in a perfect clockwise circle around the table. Is that realistic? No, but neither is interrupting combat for 2 minutes on every single frigging round to pause and try to figure out whose turn it is and what they want to do next.
The turn goes in a circle. When it's your turn, you do any TWO of the following things: attack, cast a spell, move yourself around, move an opponent around, activate an object, talk, or hide. If you want, you can pick the same thing twice, but you normally can't attack twice in the same turn unless you're holding two different weapons. Why two things? It gives you enough flexibility to be interesting, but not so much that you freeze up trying to think about it. If you're not ready with your list of two actions when the DM points at you, you lose one of those two actions. It's harsh but fair. It has to be if you want to keep combat moving at anything even remotely pretending to be a real-time pace.
The players also need a *reason* to care how they're all arranged. Like, if you can hit anyone from anywhere and expect to give and receive about the same amount of damage, then it makes no sense to waste your actions moving around, or, worse, moving your opponents around. Why would anyone want to do that?
I've heard people say that it makes sense to focus on 'glass cannons' -- on enemies who deal a lot of damage each round but who don't have many hit points. And, sure, if you have a choice of who to target, you may as well target someone who's squishy but dangerous. But it's very rare that you would actually want to spend a whole round of combat trying to change your target: most combats are decided within the first 3 rounds; at that point, by then, one of the teams has half its members dead or dying or or paralyzed or unconscious, and victory is just a matter of going through the motions. By the rules of D&D 5e, pretty much the strongest situational advantage you can gain in combat is..."advantage," which lets you roll 2d20 and keep the higher roll. You can't stack up advantages; if you have advantage twice, the second one is just wasted. On average, advantage gives you about +5 to your dice roll. Even if you could spend your first turn to guarantee yourself advantage on your next two turns, that's still only +10, or about half an extra attack. You'll almost always be better off attacking on every single turn -- better to get three attacks (and maybe miss one of them) than to limit yourself to two attacks (and still maybe miss one of them).
Same thing with rescuing teammates: sure, sometimes your own glass cannon is standing right next to your healer, or your healer doesn't have anything better to do, and it makes sense to reach over and slap your poor beleaguered ally with a Healing Touch. On average, though, you can deal more damage each turn by attacking then you can restore by healing -- and if you manage to kill or incapacitate an enemy, that works as a kind of forward-looking healing, because that enemy won't be around to keep lowering your hit points. If your goal is to keep your allies on their feet -- and it should be -- then killing enemies is almost always more useful than casting healing spells or doling out potions.
So we've got to fix that: we've got to make it so that when 4 angry goblins surround your squishy wizard, and you're standing at the other end of the cave crossing swords with one big orc, you have a reason to disengage from the orc, run over to your wizard, and protect her.
There are lots of ways you could do this, but my suggestion is to make advantage modular: for each fact about the situation that gives you an advantage, roll 1 extra d6 (remember, we're using 3d6 now, not 1d20). So, if one goblin is attacking one wizard, the goblin rolls 3d6. If two goblins are attacking one wizard, it's much harder for the wizard to dodge, so the goblins each roll 4d6 and keep the best 3 dice. If three goblins are attacking a wizard, dodging gets even harder, so the goblins each roll 5d6 and keep the best 3 dice. If three goblins are surrounding a wizard and the wizard is hopping around with her legs tied together by rope, dodging gets even harder, so each goblin rolls 6d6 and keeps the best 3 dice. If the number get ridiculous, the DM can declare that the goblins get an automatic hit, or even an automatic critical hit.
You can stack disadvantage this way, too. If your big beefy barbarian is trying to attack a manatee, but the manatee is ten feet underwater and it's hard to swing a greataxe through the water, maybe the barbarian rolls 4d6 and keeps the worst 3 dice. If the barbarian is also being grappled by a hostile squid at the time, maybe he rolls 5d6 and keeps the worst 3 dice, and so on.
So that's how you give players a clear choice: tell them they can attack, cast a spell, move their character around, move an opponent around, activate an object, talk, or hide -- and then use their decisions to stack up multiple advantages and disadvantages, including advantages based on flanking. If you do it right, the game moves faster, players feel like they understand what's going on, and you get real dramatic tension when an ally is in danger.
1. Tie ability scores directly to game mechanics.
2. Always let players roll their own dice.
3. Replace spell slots with a mana pool.
4. Use 3d6 instead of 1d20.
5. Give players a clear choice of tactics in combat.
1. Tie character stats directly to game mechanics.
D&D has arguably the best set of core character stats ever developed: Strength, Dexterity, Constitution, Intelligence, Wisdom, and Charisma really do an excellent job of describing a fantasy character's abilities. They fit together well, they don't overlap too much, and they give you an interesting set of tools for thinking about why characters are different from each other.
The problem is that instead of just *using* these character stats, D&D asks you to put them through a series of Rube Goldberg exercises before the stats take on any meaning in terms of the game mechanics. If you want to know how many points of damage you can absorb, you've got to take your Constitution stat, subtract 10, divide by 2, round down, multiply by your current character level, and then add that value to the total shown on a number of dice of varying sizes that's one less than your current level. What? No. Bull****. This is a game, not math class.
A better way to set this up would be to...say that you can absorb one point of damage for each point of Constitution. Got a CON of 14? Congratulations, you're super-buff and you have 14 hit points. Got a CON of 8? Man, that sucks, you only get 8 hit points. If you really want characters to become invincible as they level up, you can add a flat bonus per level -- maybe your hit point maximum is your CON stat plus your current level. That's it. Keep it simple. All that math isn't adding anything to the gameplay value for 95% of the players, and the handful of geniuses who actually like this kind of thing would rather see the complexity applied to something more interesting, like spells.
Speaking of which...your Intelligence can be the number of spells you're capable of memorizing when you prepare your daily spells. Got an INT of 15? You're a smarty-pants, so you can learn to cast 15 different spells before you start forgetting your old cantrips. Your Wisdom can be the total mana cost of spells that you can simultaneously concentrate on maintaining in any given hour. Got a WIS of 13? Great, you can maintain up to 13 mana points' worth of buffs for your allies without slipping up and letting the oldest spell fade away. Your Charisma can be the target that opponents need to beat in order to break free of your magical charms. Got a CHR of 17? Guess those Orcs will need to roll a 17 or better to shake off your extremely persuasive Sleep spell.
Dexterity can represent the difficulty of hitting you in combat when you're not wearing armor. Got a DEX of 14? Cool; you're reasonably shifty, so your opponents need to roll a 14 or better to hit you. If you're wearing leather armor that gives you a +2 to dodging, they need a 16 or better to hit you. This stuff should not be hard. If I never hear the phrase "Armor Class" again it'll be too soon.
Strength can represent the number of different things you can carry, or the number of kilograms you can carry, depending on whether your friends are capable of thinking in the metric system. Got a STR of 15? Cool, that's pretty strong; you can carry 15 things or 15 kilograms (40 pounds) of weight on your back and still be able to thrust and parry and dodge for an hour and then hike 10 hours over a hilly trail.
2. Always let players roll their own dice.
For some reason, D&D 5e tells players to keep track of the differences between "skill checks," "saving throws," "attack rolls," and "magical attack rolls," like these are somehow four separate things. Bull****. As the Angry GM teaches, there is a player, there is something she wants to do, and there is somebody or something that makes it possible for her to fail. That's all. The active player -- the person trying to do something -- gets to roll dice. The DM sets a difficulty target based on how hard the thing is, and if the player rolls equal to or better than the difficulty target, she succeeds. Otherwise she fails.
While this is happening, nobody else gets to roll any dice. They can roll dice later, when it's their turn. That's how taking turns works.
Also, saying that your character is "dodging out of the way of a poison arrow" or "not getting flame-broiled by a Red Dragon" or "resisting the seductive call of the magical Sirens" does not count as "trying to do something." None of these are excuses for wrecking the turn order and taking an extra turn. The person trying to do something in these three examples is, respectively, the orc archer, the Red Dragon, and the mermaid Sirens, respectively. They're probably all controlled by the DM. Let the DM roll dice for them. You can roll dice when it's your turn again. If you really can't wait, consider investing in a fidget spinner.
3. Replace "spell slots" with a mana pool.
Nobody other than D&D players has any idea what a spell slot is or how it's supposed to work. They don't match up with anybody's intuitions, and even after someone explains the mechanics of the rules to you, they still don't make any sense. What does it feel like to be a Wizard who has three Level 2 slots left and one Level 4 slot but no remaining Level 3 slots? I have no frigging idea. Obviously magic isn't real, but this kind of magic isn't even a thing that shows up in books or comics or movies about fantasy worlds. Thor doesn't have spell slots. Gandalf doesn't have spell slots. Stannis Baratheon doesn't have spell slots. Paskennarion doesn't have spell slots. Rincewind doesn't have spell slots. Glinda the Good Witch of the North doesn't have spell slots. Why should your D&D character have spell slots?
Trick question: your D&D character should not have spell slots. Your D&D character should have a mana pool. Mana is easy to understand because it matches people's intuitions about how fantasy magic works. Everyone "knows" that when you cast powerful spells you temporarily exhaust your magical mojo, and when you go to sleep or take a break from spellcasting, your mana slowly replenishes. So if you want to write rules that people can understand, go with what people already know.
There are plenty of ways to make mana pools work, but as one example, let's say your character gets 1 mana point for each point of Charisma, plus 1 mana point for each tier of Spellcasting that she has access to. Full casters like Wizards get a new tier of Spellcasting after every 1 or 2 levels; partial casters like Paladins or Rangers only get a new tier every 3 or 4 levels. So, a level 5 Wizard might have access to Tier 3 casting and a Charisma of 12, which would mean they have a mana pool of 3 + 12 = 15 mana points.
Spells are ranked by tiers, with "cantrips" treated as Tier 1 spells, and the most devastating ultra-bombs like Time Stop and Resurrection and Wish topping out somewhere around Tier 12. Each time you want to cast a spell, you've got to pay 1 mana for each of the tiers involved, adding the value of each tier together. So, a tiny cantrip like Acid Splash is a Tier 1 spell that costs 1 mana. Sleep would be a relatively weak Tier 2 spell that costs 1 + 2 = 3 mana. Fireball would be a moderately powerful Tier 4 spell that costs 1 + 2 + 3 + 4 = 10 mana. Raise Dead is a awesome Tier 6 spell that costs 1 + 2 + 3 + 4 + 5 + 6 = 21 mana. You get the idea.
If you don't have enough mana to pay for your spell, then your mana dips into the negative, and you've got to beat a difficulty check with a target equal to your current mana debt. So, let's say you want to cast Fireball (10 mana cost) but you've only got 3 mana left. Now your mana pool dips down to negative 7, which means you've got to roll your dice and get better than a 7, or else the spell fizzles or explodes or hilariously backfires. If you want to cast *another* fireball before resting, now your mana is at negative 17, which means you need to roll your dice and get better than a 17 to avoid serious consequences. Beating those checks quickly becomes impossible, so you've got to stop casting *eventually* -- but you can press your luck if you want, dig deep for a little bit of extra mojo, instead of always standing around with your hands in your pockets after 3 pm because you're "out of spell slots."
You recover mana at the rate of...1 mana point per hour. This is, coincidentally, the same rate at which you recover hit points. None of this "short rest" or "long rest" garbage that has everyone gaming the system to figure out how to spring back to 100% capacity while sleeping on the moldy floor of a cursed dungeon. The longer and more often you rest, the more you'll recover. You can rest for as long or as short of a period as you like, just like real life. If you take too many naps in the wrong places, you risk being fired by your boss and/or a Red Dragon, but that's just business as usual.
4. Use 3d6 instead of 1d20.
I know, I know, the d20 is iconic; a literal symbol of all that D&D stands for. They're fun to fidget with, and fun to roll, and they work well for Initiative, which comes up nearly every session, and which ought to be mostly random. For other skills and checks, though, d20s are mechanically terrible, because they have such a huge, flat probability curve that a +1 bonus doesn't have any noticeable effect. Like, suppose you have an 55% chance of accomplishing a task instead of a 50% chance of accomplishing the task. Would you even notice the difference? Statistics says no. If you flip a fair coin 100 times, it usually won't come up exactly 50 heads and exactly 50 tails -- it's more likely that it'll be slightly biased in one direction or the other. In fact, the 'average' amount of bias is that you'll get either 55 heads or 55 tails. This means that your +1 bonus gives you literally the same amount of variation you'd expect to see by chance when you make the same roll 100 times.
How often do you actually go and make the exact same roll 100 times in D&D? Let's say it's a really basic roll, like attacking with your sword against a perfectly ordinary goblin. You make the roll 5 times per combat, you fight in 3 combats every 4-hour session, and it takes you 3 sessions to gain a level, and you gain 3 levels before you get any new abilities or equipment that affect your combat roll That's still only 5 * 3 * 3 * 3 = 135 rolls...barely enough to notice a +1 difference, even when you're doing nothing but hacking relentlessly away at the goblin hordes for 36 straight hours of play. In more realistic circumstances, where you fight a variety of enemies using a variety of tactics, you probably would never make the same roll more than 10 times in a campaign, which means you wouldn't even notice a +4 difference in skill. Sure, it *feels* cool to have a fat bonus, but it's all in your head. Mathematically speaking, whether you succeed or fail on any given roll, or even on any given ten rolls, has much, much more do to with luck than with your exact skill level.
That doesn't have to be the case. The problem with the d20 is that it's flat -- you have the same probability of rolling any number. If you switch to 3d6, though, then certain numbers are much more likely than other numbers. It's a lot easier to roll a 10 on 3d6 than it is to roll a 3 or an 18. This means that even small bonuses can have a big effect. Let's say you're trying to whack that goblin again, and you would normally hit on a roll of 11 or better. Using either 3d6 or 1d20, you've got a 50% chance to hit. Now I hand you a magic sword that gives you +2 to all your attack rolls. Using the 1d20, that only boosts your chance to hit from 50% up to 60%. Not much difference. Using 3d6, though, the same +2 sword boosts your chance to hit from 50% all the way up to 74%. That's a big change that you're likely to notice. With a +5 sword, the d20 would still only give you a 75% chance to hit...but the 3d6 would give you an impressive 95% chance to hit.
The probability curves for 3d6 still deliver on D&D 5e's promise that no hero will ever be invulnerable against a swarm of hundreds of goblins, even without dipping into edge cases like critical hits or natural 20s. A goblin *could* roll triple 6's and score 7.5 points higher than the expected attack roll, just like a goblin *could* roll a 20 and score 9.5 points higher than the expected attack roll. The difference is that the crazy, implausible events are *actually rare* when you use 3d6. If you use 1d20, you get supposedly miraculous events all the damn time, to the point where attributes and proficiencies start to lose most of their meaning. If a Level 15 Barbarian, a Level 1 Wizard, and an irritated NPC goat all have a decent chance of kicking you in the shins, then why bother keeping track of stats for different kinds of characters? You could just skip the character sheets and roll totally random dice every time there's a combat. Or, you could use 3d6, and have die results that actually correlate with your character's abilities.
5. Give players a clear choice of tactics in combat.
In theory, D&D players can do all kinds of different things in combat, limited only by their imagination and their core attribute stats like Dexterity and Strength. In practice, almost nobody ever figures out how to do anything other than swing a sword, cast a fireball, drink a healing potion, or run away. The rules for what you can do in combat are buried in a long, rambling encyclopedia of every possible action, and they don't make it clear why you would want to give up your attack to futz around with shoving or grappling or opening doors. There's also a horribly unclear set of rules about surprise attacks, readied actions, and passing your place in the turn order. Do you have to go when it's your turn, or can you wait till later? Can you take two turns in a row if you're sneaky enough? When? Why? Nobody really knows.
So, let's fix this. In combat, you go only on your turn. You never go on anyone else's turn. If enemies are surprised, that means they get a temporary penalty to their stats, not that they lose their turn. Everybody gets one turn, and there's a fixed turn order in each combat, and it almost always goes in a perfect clockwise circle around the table. Is that realistic? No, but neither is interrupting combat for 2 minutes on every single frigging round to pause and try to figure out whose turn it is and what they want to do next.
The turn goes in a circle. When it's your turn, you do any TWO of the following things: attack, cast a spell, move yourself around, move an opponent around, activate an object, talk, or hide. If you want, you can pick the same thing twice, but you normally can't attack twice in the same turn unless you're holding two different weapons. Why two things? It gives you enough flexibility to be interesting, but not so much that you freeze up trying to think about it. If you're not ready with your list of two actions when the DM points at you, you lose one of those two actions. It's harsh but fair. It has to be if you want to keep combat moving at anything even remotely pretending to be a real-time pace.
The players also need a *reason* to care how they're all arranged. Like, if you can hit anyone from anywhere and expect to give and receive about the same amount of damage, then it makes no sense to waste your actions moving around, or, worse, moving your opponents around. Why would anyone want to do that?
I've heard people say that it makes sense to focus on 'glass cannons' -- on enemies who deal a lot of damage each round but who don't have many hit points. And, sure, if you have a choice of who to target, you may as well target someone who's squishy but dangerous. But it's very rare that you would actually want to spend a whole round of combat trying to change your target: most combats are decided within the first 3 rounds; at that point, by then, one of the teams has half its members dead or dying or or paralyzed or unconscious, and victory is just a matter of going through the motions. By the rules of D&D 5e, pretty much the strongest situational advantage you can gain in combat is..."advantage," which lets you roll 2d20 and keep the higher roll. You can't stack up advantages; if you have advantage twice, the second one is just wasted. On average, advantage gives you about +5 to your dice roll. Even if you could spend your first turn to guarantee yourself advantage on your next two turns, that's still only +10, or about half an extra attack. You'll almost always be better off attacking on every single turn -- better to get three attacks (and maybe miss one of them) than to limit yourself to two attacks (and still maybe miss one of them).
Same thing with rescuing teammates: sure, sometimes your own glass cannon is standing right next to your healer, or your healer doesn't have anything better to do, and it makes sense to reach over and slap your poor beleaguered ally with a Healing Touch. On average, though, you can deal more damage each turn by attacking then you can restore by healing -- and if you manage to kill or incapacitate an enemy, that works as a kind of forward-looking healing, because that enemy won't be around to keep lowering your hit points. If your goal is to keep your allies on their feet -- and it should be -- then killing enemies is almost always more useful than casting healing spells or doling out potions.
So we've got to fix that: we've got to make it so that when 4 angry goblins surround your squishy wizard, and you're standing at the other end of the cave crossing swords with one big orc, you have a reason to disengage from the orc, run over to your wizard, and protect her.
There are lots of ways you could do this, but my suggestion is to make advantage modular: for each fact about the situation that gives you an advantage, roll 1 extra d6 (remember, we're using 3d6 now, not 1d20). So, if one goblin is attacking one wizard, the goblin rolls 3d6. If two goblins are attacking one wizard, it's much harder for the wizard to dodge, so the goblins each roll 4d6 and keep the best 3 dice. If three goblins are attacking a wizard, dodging gets even harder, so the goblins each roll 5d6 and keep the best 3 dice. If three goblins are surrounding a wizard and the wizard is hopping around with her legs tied together by rope, dodging gets even harder, so each goblin rolls 6d6 and keeps the best 3 dice. If the number get ridiculous, the DM can declare that the goblins get an automatic hit, or even an automatic critical hit.
You can stack disadvantage this way, too. If your big beefy barbarian is trying to attack a manatee, but the manatee is ten feet underwater and it's hard to swing a greataxe through the water, maybe the barbarian rolls 4d6 and keeps the worst 3 dice. If the barbarian is also being grappled by a hostile squid at the time, maybe he rolls 5d6 and keeps the worst 3 dice, and so on.
So that's how you give players a clear choice: tell them they can attack, cast a spell, move their character around, move an opponent around, activate an object, talk, or hide -- and then use their decisions to stack up multiple advantages and disadvantages, including advantages based on flanking. If you do it right, the game moves faster, players feel like they understand what's going on, and you get real dramatic tension when an ally is in danger.