I ran a variant of DnD3.5 a few years a go with this house rule Player Only Rolls. The basic of the idea was that the referee never rolls at all. The players roll to attack vers the creatures AC. They also roll to defended like a save vs the creatures static attack DC. This rule applies to everything, so I changed all NPC rolls to static defense values that the PC roll against.

What are the advantages, disadvantages or problems with these types of rules?

  • 3
    \$\begingroup\$ Related: How can the GM go diceless in classic D&D? \$\endgroup\$ Commented Nov 2, 2010 at 15:39
  • \$\begingroup\$ Actually this looks like it is the answer to that question ;) \$\endgroup\$ Commented Nov 2, 2010 at 15:53
  • 2
    \$\begingroup\$ @David That Q is more about how, while this is about why. They're nicely complementary. :) \$\endgroup\$ Commented Nov 3, 2010 at 1:46
  • 1
    \$\begingroup\$ In my not so humble opinion, this is not system agnostic but D&D. \$\endgroup\$ Commented Feb 11, 2014 at 9:04
  • 1
    \$\begingroup\$ Some games have this baked into the rules - eg, Numenera \$\endgroup\$
    – Adeptus
    Commented May 6, 2015 at 1:45

9 Answers 9


Typically the DM has to do a lot more rolling than the players. So, this can be helpful to some DMs. Plus, even if the players don’t know what the rolls are for, it can make some players feel more involved. And for some rolling dice is simply fun and more rolling means more fun.

You might think that this prevents the DM from hiding the results of rolls, but that isn’t necessarily the case. The DM can ask for rolls without telling the players why, and he can shift the meaning of the results to obscure it further. (Have a handy table for each die type that scrambles the results. e.g. d4: 1=3, 2=1, 3=4, 4=2.) Or the DM can simply still roll for those kinds of rolls.

This can have further effects if you end up changing the mechanic. e.g. When a PC casts a spell, does the player roll a single “reverse save” for all monsters affected or do they roll individual “reverse saves” for each monster? The latter is the same as if the DM had rolled individual saves for each monster, but the former is different. Whether that’s good or bad, however, is subjective.

I suppose the only problem I see is that it takes some effort to carefully think through each reversed mechanic and make sure that it is equivalent or that you are OK with the difference. Plus a bit of effort to translate things on the fly.

  • 3
    \$\begingroup\$ -1 Coming up with elaborate new rules for hiding information from the players is a colossal waste of time. \$\endgroup\$
    – Alex P
    Commented Feb 10, 2014 at 13:35
  • 1
    \$\begingroup\$ I agree, at least in part, Alex P, but I was just answering the question. \$\endgroup\$ Commented Feb 13, 2014 at 14:20

I'd be concerned about two classic DM techniques that seem to be prohibited by this approach:

  1. Hidden knowledge - If the players know the result of every die roll, they prematurely learn about armor class, monster strength, or DC requirement.

  2. Cloaked passive checks - Another important technique is DM rolling unexpectedly for passive checks, or more importantly, fake-out checks (again to hide when something is or isn't happening.) I sometimes use this to keep the player's attention on the game. Though I can see how this could be accomplished with players-always-roll, it would be more cumbersome and overt.


I firmly believe that having the players roll the dice as much as possible maximises their investment in the game outcome (where the game involves dice, of course); the looks on my players' faces when they get a natural 20 in the middle of a nasty combat, or fumble against a kobold, are evidence of a degree of investment and enjoyment that I don't want to deprive them of. That said, I have a number of techniques to keep my ability to fudge:

  • Although they roll their own to-hits, they don't know their own THAC0 or the opponent's armour class. Whilst they get some idea over the course of any given combat what they need to hit it, they're used to something that worked earlier failing now, or vice-versa. Because they themselves are aware of short-term magics that can raise or lower their own to-hit, they don't find it odd that opponents might avail themselves similarly. The same is true of damage.

  • For saving throws, I can fudge the outcome rather than the roll. I don't set them up with "save vs poison or die", then have to explain away why Thongor The Studly is still waking around after a 2; I say, "Oh, you've been bitten. I need a saving throw.". If they roll a 2 and for my own reasons I don't want them to die, maybe they go green and fall into a coma for two days, or their left arm (where they were bitten) is paralysed and they can no longer use a shield, or some other bad but non-fatal outcome.

  • For hit point rolls when levelling-up, I present them with a simple "Would you like to roll it, or shall I?" choice. If they roll it, the roll stands; if they have me do it, I may fudge up a low roll - although it's never understood that I will, and from time to time if I roll a 1 that's what the character gets, provided it won't cripple them unduly.


Never played with such rules but this is what came to my mind when looking at the rules.

The good is the players are in total control. The players feel like they have direct control over their destiny. They're not worried about the DM rolling really well all the time and hitting their AC. They're worried about themselves rolling well enough to defend. Additionally they don't have to worry about the DM making things suddenly more difficult—or the opposite—feel like they're being coddled 'cause the DM made stuff easier out of pity. (This can annoy some players.)

The bad is the players are in total control. You can't fudge the dice, if need be, to help or challenge them more. I know to some fudging dice rolls is a bit of a touchy subject, but there are some days where your monsters need a little bit of a boost to actually make a fight challenging or you need to lower a DC as the session went on 'cause you overestimated how good your party's skills are. Additionally, some days the dice are against you and you need to boost, and some days they're against the players and they need a magical boost.


This is the way I usually play most games; as other have mentioned it keeps the players interest high and gives them more of a sense of being actively involved in their fates. For me rolling against static numbers is fine, I don't see any value in players being unsure whether a roll that was good enough last round is still good enough... unless something has changed since last round, like the lights going out or the opponent switching to defense, it's actually quite helpful to speed things along. I don't even usually bother to keep things like the target number secret; if their opponent is wearing chain, they know what the AC ought to be, and if the opponent is better than that for some reason (like high DEX or something) that also ought to become apparent after a few moments.

One thing I do think you have to watch out for is knowledge rolls; I usually do those myself and keep them secret. If it's just a passive check to notice something you can fake them out by sometimes asking them when it's not needed, but if they initiate a search I'd rather the players not know whether they found nothing because there's nothing to find or they found nothing because they rolled poorly. Letting them roll for things like searches and detect lies can give the players too much information the character shouldn't know if they roll really well or really poorly; good players can refuse to act on that info, but why make things hard on them? Keeping the roll secret lets them decide whether they want to go on searching or asking questions without having to second guess themselves.


(Since this is tagged , here's a general answer with a bit of secondary focus on .)

Implementation Details

  1. You may have to reprocess a lot of rules.

    "Reversing" a die roll mathematically isn't that complicated, but it does involve more thinking than I'd really want to do all the time on the fly. You'll slow down the game unless you go through and figure out the new rolls ahead of time and write those values down. Depending on the game, this is could be a ten-minute task or something that requires annotating an entire rulebook.

    Old D&D, for example, doesn't really use a consistent single approach to rolls. So you have a bunch of individual mechanics to sift through and change.

    If you don't put in the work (whether that's annotating the rulebook or just making a little handy conversion chart), handling time will suffer.

  2. You're essentially turning one mechanic into two.

    If your game normally has a mechanic for attacking, you now effectively have separate "attack" and "defend" mechanics, depending on who's doing it. Et cetera.

    Overall, this means you're increasing the complexity of the rules. There's always some possibility that players who already struggle with "What do I roll now? What do I add?" may find the added extras too much to process.

  3. Not all rolls are "reversible" in a way that makes a lot of thematic sense.

    For instance, in D&D, you can replace the monster's attacks with rolling to block, dodge, and parry. Cool. But then if you fail, you're taking damage. You could make the players roll this, but they're likely to feel like they're rolling for the monster rather than rolling for their own character, at that point.

  4. "Dice manipulation" mechanics become more important.

    A lot of games, including some variants of D&D3.5 (e.g. Eberron), have "action points"/"fate points"/&c. — various mechanics representing "heroic effort" or a "second chance." Giving the players more things to roll means more opportunities to do special stuff with those dice.

    In general, I think being able to use your action points for defense rolls is a good thing, but I could see it encouraging analysis paralysis and thereby slowing down the game.

  5. What about PC-vs-PC checks?

    Even in a game where the player characters will always be allies, there are occasions where one player will want to take an action against another — a party argument the group decides to settle with dice, for instance. So you'll need alternative rules for handling situations where PCs are on both sides of the conflict. That's not really that big of a deal, but it adds to the "I have multiple mechanics for everything now" issue.

  6. "Secret" rolls are less secret.

    Pretty self-explanatory, really: if the PCs are truly rolling all the dice, then the GM can't make covert rolls. This isn't really an issue since you can ask for rolls without stating what they are, if needed.

    Though, in my experience, telling players what's at stake when they make a roll tends to actually heighten tension. I think you really can't go wrong offering at least a vague description of whatever danger's afoot, even if the PCs fail their rolls.

  7. It might be slightly harder for the GM to cheat.

    Who cares, really? Most games that expect the GM to cheat give her plenty of ways to do so. I've never seen "roll in the open" stop a determined illusionist.


  • Will it make the game simpler? Nope.

    You've actually introduced more mechanics, thanks to the mirror-image different between "PC attacks monster" and "monster attacks PC."

    At best, you might lighten the GM's load a bit, in GM-bottlenecked games. But the GM is still pretty involved in the overall process, so at best she's sparing herself a bit of math.

  • Will it make the game faster? Possibly! It depends.

    It's really about where the bottleneck is (see above). You haven't simplified the mechanics, just shifted who's responsible for some of the rolls.

    One thing to be aware of is that a lot of players of crunchy turn-based games kinda learn to tune out when it's not their turn. Having to "wake them up" to get a defense roll for every enemy attack is slower than just saying, "Okay, Steve, the ogre hit you for 18 damage, moving on..." On the other hand, having players roll more may discourage the "tuning out" behavior in the first place — it's something the group needs to manage and actively strive for, though.

  • Will the players be more engaged? It really depends.

    In my experience, most players respond positively to mechanics that make them feel like they have more influence over their character's fate. That said, when's the last time you really felt like an empowered player when the GM said, "Okay, now make a saving throw vs. sleep?" I think adding actual decisions — even relatively simple ones — is more likely to build engagement than simple having the players roll dice more often.


I agree with alot of what has been said by David and Randall and thought I would add my 2 cents on top of that. There are times when you don't want the players to A) know you have rolled and/or B) know what the roll was. Being both a player and an occasional DM, I don't believe I would want to play by such a house rule.

I am not a fan of everything being static like the question mentioned. Take, for example, an opposed grapple check over multiple rounds. The player rolls a 15, the monster rolls a 14. Player wins this round. Next round, player rolls a 15. With static numbers, he expects to win again. What if the monster were to roll a 16? It adds uncertainity andt adds some variety that wouldn't be there with static numbers. And I think both are good things to have. It would also add a bit to the DM's prep time in converting the stats.

  • \$\begingroup\$ Erm, the question is more about moving rolls usually made by the GM to the players, rather than replacing GM rolls with static numbers. In the variant of 3.5 mentioned by the OP, that's exactly what happens: Dice normally rolled by the GM on behalf of monsters are rolled by the players, instead. Which isn't to say your answer isn't valuable, as you're right, if you replace rolls with static values, a game can become dangerously predictable. It's just that the question didn't mention a change to "everything being static" in the way you describe; vanilla 3.5e does not use opposed rolls. \$\endgroup\$
    – GMJoe
    Commented Feb 11, 2014 at 1:28

I ran a Babylon 5 (Mongoose d20 version) game and a DnD3.5 adventure using the Players Only Roll from above with the Mid3d20 variants. I thought it works quite well over all.

My thoughts:


  • Not sure all the players like the idea of saving vs being damaged.
  • Sometime the logic of players rolling their defense just felt wrong.
  • It requires some relearning as a referee as there where times I reached for the dice to roll an attack.
  • At first I spent more time working out the changes to the creature in prep time, but after a time. It became Defense roll add 10, Attack roll add 12. That is, say a PC want to sneak past guard, then target is Spot + 10 vs PC Stealth roll, an Assassin want to sneak attack a PC is Stealth + 12 vs PC Spot roll to notice.
  • Players tend not to mention Crit and Fumble when they are not in there favor as often so you need to look out for the 1 and 2 as well as the 19 and 20.

Good Points

  • I did find that I have more mental time in combat as I only had to say yes a hit or no a miss for both player and creature attacks.
  • Less maths for the referee
  • Player get to roll for defense gives some players more immersion in the combat as they get to attack more often, that is, not just on there turn. So attention to what as going on was higher.


  • As I used the Mid3d20 then opposed rolls did not require both antagonists to roll to get a bell like curve. Some might find flat rolls are wrong if players only roll one die.

I tried incorporating this variant as a GM in a D&D 3.5 game a few years ago.

I, as the GM, loved it, for two reasons:

  1. I could focus my time on managing the encounter. As a GM, I usually have plenty to do at any given round of combat anyway. Rolling and adding together results takes time of its own, but most importantly, there was less "mental context switching" going on. I didn't need to look away from my notes and to the dice and back to my notes five times per round, it's all there in the notes and the players handle the math-gruntwork.
  2. The players were more engaged. A major impact of this variant is that the players can expect having to physically pick up a die and roll at any given time, regardless of whether or not it was their turn.

The players weren't quite as pleased. There was only really one reason: Confusion.

They suddenly found it hard to remember when to roll what, how the modifiers and static numbers worked, and generally what they should put down on their sheets. And it didn't get better in a few sessions. It was as if they constantly had to work through a translation layer between what it says on their sheet/in the books and what they had to actually do. We ultimately dropped it.


You must log in to answer this question.

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