My players and myself feel the "wear down" mechanic of hit points in 5e considerably subtract from the fun of the game and the strategy.

I've been considering introducing a exploding dice mechanic for damage in 5e by which any dice can ace at any time. I feel this would introduce a level of randomness which will make the players have to think twice about strategic combat. All of the sudden that low level goblin can be a real danger, maybe not to the extend it is in Savage Worlds but at least something the players should stop and consider.

The mechanics would work exactly like the exploding dice of Savage Worlds but limited to Damage (maybe healing too):

  • Criticals will disappear, either you hit or you don't
  • After hit you roll damage normally but if your roll aces (i.e. a 6 in a D6, an 8 in a D8) then you throw another dice and add the results. This is done per die (i.e. in a 2D4 either or both dice can ace)
  • Dice can explode indefinitely leading to possible massive damage. As in Savage Worlds, theoretically, a single blow can kill anyone.

Has anyone experimented with including exploding dice in DnD 5e? Did you experiment (or envision) any unbalance or difficulties?

I'm interested on actual experiences or analysis on games. Note some analysis already exist:

But this focuses on statistics rather than analysing whether it would break the game or things to be careful about when implementing it.

Exploding dice average damage values:

Die Normal Exploding Ratio
d4 2.5 3.3 1.33
d6 3.5 4.2 1.20
d8 4.5 5.1 1.14
d10 5.5 6.1 1.11
d12 6.5 7.1 1.09
d20 10.5 11.1 1.05
  • 1
    \$\begingroup\$ Are you only looking for play experience, or are you interested in numerical analysis too? \$\endgroup\$ Dec 18, 2022 at 14:06
  • 4
    \$\begingroup\$ "My players and myself feel the "wear down" mechanic of hit points in 5e considerably substract from the fun of the game and the strategy." If you ask this as a separate question, there are a few answers to this (from various different schools of thought). \$\endgroup\$
    – user2754
    Dec 18, 2022 at 17:05
  • \$\begingroup\$ @user2754 I can't see any way to formulate that question without it being "opinion based" :) \$\endgroup\$ Dec 19, 2022 at 11:33
  • 3
    \$\begingroup\$ Exploding dice do have some weird corner cases on probability. For example the average total of an exploding d6 is higher than that of an exploding d4. But if you need a 6 or better, you're actually more likely to succeed on an exploding d4 than on an exploding d6! (1/4 * 3/4 chance to roll a 4 on a d4 and then a 2 or better is greater than the 1/6 chance to roll a 6 on a d6) \$\endgroup\$
    – Ben
    Dec 20, 2022 at 1:28
  • 2
    \$\begingroup\$ Side note: if you implement this, you might need to adjust some other rules, like e.g. the rule for automatic critical hits against unconscious or paralyzed targets. \$\endgroup\$ Dec 20, 2022 at 15:37

3 Answers 3


This is going to become grittier

To sum up: changing this will lead to a roughly 10% higher mortality rate in tier one play. It will become less at higher levels, but still increase.

As you can see from the table you included in the question, this type of critical implementation will lead to higher damage output -- it's nearly the same for exploding d20 (as double exploding there is very unlikely), but compared to the 1 in 20 critical hit chance on a nat 20, the percentage of added damage is higher for all the dice, the higher the smaller the dice. On a dagger dealing d4, you will get 33% more damage instead of the 5% you get from the d20 crit system.

How high is the chance for a KO punch?

However, what is more important to your question, this system does have a chance to reach some rather high single hit damage levels -- as you say, theoretically unlimited high, but in exchange also being unlimited unlikely. In practical terms you can use these functions on anydice to see how likely it is to get an extreme outcome on a single hit, setting the output table to "at least":

output [explode d6] named "explode d6"
output [explode d8] named "explode d8"
output [explode d10] named "explode d10"
output [explode d12] named "explode d12"

And below is the output for a d8, for example

enter image description here

As you can see, the probability for exploding 3 times, for at least 25 points of damage, is about 0.2% here, and the higher values are not shown.

Much like with normal criticals, this is especially deadly at lower levels, as you do not need a lot of consecutive explosions to take the characters out. Even with some light Constitution bonus, typical characters maybe have around 10 hp on the first level, 16-17 on the second level, 24-25 on the third level, and 32-33 on the fourth level.

That means a d8 hit would have a 0.2% chance to outright drop a character of level 1-3. If you move up to a d12, you have a 0.23% chance to deal 33 or more, dropping a character up to level 4.

How many punches will you take?

The issue is that over the course of their career, the characters get hit many, many times. Just roughing it out, with an average combat maybe taking 3-5 rounds, attacks hitting about half the time, and the typcial day having maybe 3 combat encounters, you get a ballpark of 6 attack hits per character and day -- lets triple that, because there may be multiple weaker opponents that have more attacks in total, and stronger opponents often have multi-attack routines, and round up for good measure. We do not need to be super exact to get an idea. This could mean about 20 attacks per day of adventuring.

According to the rules in 5e, it would take you about 33 days to go from level 1 to 20, if you remove rest days and such. To get to about level 12, where many campaigns tend to wind down will take you about 24 days. You spend about 6 days of combat in tier one, good for about 120 times being hit.

You have a 20% chance to get hit by a triple-exploding d8 during that time, an 11% chance to get hit by a quadruple-exploding d6 that likewise deals more than 24 damage, and looking at orcs, who hit with greataxes, a 24% chance to get hit for at least 33 damage, and killing you all the way up to level 4.

  • 1
    \$\begingroup\$ I agree with this take. The result will be that on average the damage will go up some relatively small percentage. But you are also almost guaranteed to die to some random hit at some point. That's fine, if that's what the questioner is going for! \$\endgroup\$ Dec 19, 2022 at 14:01
  • 1
    \$\begingroup\$ @JordiVermeulen: In 5e, you only insta-die if damage takes you to -max HP, which is a lot less likely than just going to zero. (If you're already low and the damage explodes, that could be lethal, especially at lower levels with smaller HP pools). Barring that, you have 3 death saves for an ally to cast a spell or ram a healing potion down your throat. So there's a large cost in action-economy. In tier 2 or higher, often someone in the party will have revivify, which is expensive (300gp). Fully dying and getting revivified is probably a big deal in an RP sense for many PCs, too. \$\endgroup\$ Dec 19, 2022 at 23:29
  • \$\begingroup\$ @PeterCordes Suffering a critical hit is 2 failures. Any attack from within five feet has Advantage and is a critical hit if successful. One multi-attack, by a creature dealing 1 damage per hit, can kill an incapacitated character, no death saves allowed. \$\endgroup\$
    – ValhallaGH
    Dec 20, 2022 at 15:24
  • \$\begingroup\$ @PeterCordes You are right on being dropped not immediately equalling death. Were I to critizise my own answer, this would be one of the points, the other that maybe 20 hits per day is on the high side of things, and something like 12 or 15 could be closer, and this number affects the overall chance of taking a big hit over time. But on the other hand, the answer does not consider at all that you easily can get hit but such an attack when you are already low on hp. That makes it much easier to drop you directly from middling hp, and if you are very low, can kill you outright. \$\endgroup\$ Dec 20, 2022 at 15:37
  • 1
    \$\begingroup\$ @AndrewSavinykh Yes, that is how AnyDice does it. I think because the tail is theoretically endless, it cuts off at „24 and up“. Obviously, you‘d never get a 24, as it explodes. \$\endgroup\$ Dec 20, 2022 at 22:23

Will This Break D&D?


The effect will be similar to the old d20 Star Wars system, with its critical hit mechanics*. The collected experience of thousands of players, across years of play, was that the primacy of player characters, scaling of damage, and frequency of attacks against the player characters meant that a PC had a ~8% chance to survive until level 20 - there was a ~90% chance of being killed by a random critical hit before reaching that level.
Major foes may be distressed and defeated, but the cumulative effect will be the death of the player characters.

Using your described rolling method, I've personally rolled 47 damage with 2d4+0. I've witnessed 3d6-1 roll 1802 damage. I've got dozens of samples of 2dX rolling 50+. This house rule will kill player characters about as quickly as not increasing HP after level 4.

Aside: Removing critical hits has a ton of other interactions, especially with class features tied to that core design mechanic. That's a second major change that is going to cause problems for players.

*Characters had two kinds of health - Vitality (which increased each level, like 5e HP) and Wounds (equal CON score, rarely changed). Mooks and minions did not get Vitality. Normal damage went to Vitality, which recovered in minutes, and only became Wounds if the character had no more Vitality; critical hits went directly to Wounds, and a character without Wounds was dying or dead.

Alternative Solutions

Put more alternative solutions into your encounters. These interactions are likely to be cinematic, interesting, memorable, and can be highly effective if the GM wants them to be.

  • The foe tries to break the characters' wills, pulling each into a frozen moment for a quick conversation, and a well delivered counter argument can halve the foe's remaining HP.
  • The dramatic set dressing (lava, infinite abyss, storm-tossed sea, etc.) is quite real and foes can be tossed into it for tremendous damage.
  • Key weaknesses are present in the encounter and can be damaged to swiftly destroy the foe. Chalices of magical blood, soul-filled crystals, ongoing magical ritual circles, and other genre-appropriate sources of power and vulnerability. Bonus points for highlighting the connection by describing changes to the power source when the foe is damaged by attacks and spells.

Good luck!

  • 7
    \$\begingroup\$ @JorgeCórdoba If you want party kills then this house rule will do that. I've never seen a party kill campaign that was successful (i.e. fun for all playing) but maybe your bunch of weirdos are exactly that kind of weird. Good luck. \$\endgroup\$
    – ValhallaGH
    Dec 18, 2022 at 15:55
  • 11
    \$\begingroup\$ "I've witness 3d6-1 roll 1802 damage" you mean someone rolled 300 sixes in a row? If so, you wintessed someone cheating... Or maybe I'm misunderstanding how the rolling works. \$\endgroup\$ Dec 19, 2022 at 0:23
  • 3
    \$\begingroup\$ @Glasses2C_Sharp By rolling damage in the same manner as the OP suggests. Rolling two d4, I rolled multiple 4's and the final total was 47 (I've not rolled above 35 since). For the 3d6-1 roll, the player rolled three 6's one hundred consecutive times, and then rolled three 1's; after subtracting the -1 modifier, the total was 1,802. It took about five minutes of die rolling, entirely for fun (24 damage killed the target). \$\endgroup\$
    – ValhallaGH
    Dec 19, 2022 at 23:25
  • 5
    \$\begingroup\$ @ValhallaGH I have bad news for you, your other player is a dice cheat, and perhaps the worst dice cheat of all time. Rolling 3d6 and getting 18 just three times in a row is as unlikely as winning the mega millions lottery. The chance of doing it six times is equivalent to getting struck by lightning twice in a single year. I hope you can imagine how comically unlikely it is to do it 100 times without weighted dice. \$\endgroup\$
    – Daniel B
    Dec 20, 2022 at 4:29
  • 3
    \$\begingroup\$ @PeterCordes Yes, I saw physical dice, using my physical eyes, absorbing actual light reflecting off those physical dice. I didn't watch all of the rolls, because it stopped being relevant after a certain point and watching one person roll dice for 5 minutes is quite boring, but I saw many of the rolls (all sixes) and the final roll (all ones). \$\endgroup\$
    – ValhallaGH
    Dec 20, 2022 at 20:11

It likely won't make much difference.

The occasional d6 will roll a ~9 instead of a 6. The more occasional d12 will roll a ~18 instead of a 12. Player characters often go 'down' in 5e D&D, and at higher levels even die (to be brought back nigh instantly). The expected damage increase of max die number exploding is fairly similar to rerolling damage on a 1 - this is such a marginal change that many optimizers outright ignore it in their math. The major change I expect from this rule is that at lower levels (1-2, maybe 3) low-HD characters will be at greater chance of instant death (2x hp done in a single swing) from weapon attacks.

For attacks like those rogues or paladins use, where you're adding a bunch of dice to the roll, you may see more damage. But even a max level divine smite doing 6d8 has merely 'good' odds of getting a single 8, adding another ~4 damage to the overall roll.

Players may become excited when someone rolls a number of max numbers in a row. A string of 6's such that a 'weak' attack becomes a much stronger one will be exciting, if much rarer than critical hits and probably the result of a poorly manufactured die/online roller (I recently played a game on foundry where the browser vtt rolled 6 18s in a row, and then something like 30 4's on d6, with only a single other number between the 4's).

But overall, the occasional 'exploding damage roll' knocking down a player character is not meaningfully different from weight of attacks or breath weapons knocking down a player character, and Healing Word will still pick them right back up. It's very unlikely to cause any additional tpks, or meaningfully increase the combat prospects of your average goblin who is already on shaky ground trying to survive in a world of fireballs and hit the magically-armoured tower-shielded paladin in the first place.

It will slightly increase damage, especially damage using smaller die sizes (magic missile!), meaning that greatswords will definitively be better than greataxes in this system, and fireball will get a bump, as will breath weapons and big chunky monster melee hits. So damage will go up (again, not much on average) which is a change many DMs already talk about making in lowering monster hp and increasing monster damage (gets around the 'padded sumo' effect of low damage parties + low damage monsters making combat drag on for 5+ rounds of largely static hitting).

I have not used exploding damage dice in 5e D&D. I have played a number of games that did use them, typically without 5e's hp-pool and death-survival mechanics - someone exploding a combat roll in Ars Magica tends to rip off your arm if not outright kill you, and in shadowrun if someone edges an attack you tend to be either totally immune to it (sam vs pistol) or it fills your boxes and blows you into a fine red smear without much trouble (face vs assault rifle).

To reiterate, this is not meaningfully different to adding more monsters to encounters or upping monster damage while lowering monster hp, both common measures taken by DMs to increase 'lethality' in the game. It likely won't meaningfully change the way the game is played or the danger of goblins with sharp sticks.

  • 8
    \$\begingroup\$ I think this analysis focuses too much on the averages, rather than the "spikyness" of damage. A 6d8 will, on average, have one die explode 3/4 of the time. However, rolling multiple max rolls still happens often enough. A 4d4 might roll 4 4 4 1 and even if the extra dice all roll 1s that's still a good chunk of damage, proportional to the roll. And chances are it will roll even higher. So it's these spikes that are dangerous, rather than the average. The lower the die, the more it tends to explode. Which can dramatically change how planning is done, especially at lower levels. \$\endgroup\$
    – VLAZ
    Dec 19, 2022 at 12:48
  • 2
    \$\begingroup\$ The averages are important to consider for fireball, smite, or sneak attack. You can expect to get about one or two aces on a fireball pretty consistently, so that's the same bonus as upcasting fireball by one or two levels. For free. The average damage isn't different merely because you're rolling more dice, but it becomes less... spiky. \$\endgroup\$ Dec 19, 2022 at 20:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .