On Light

I played a short game of D&D this weekend with some good friends (OD&D with OED interpretations). Played very well with 2 first-time players, my beginner-level girlfriend, and an almost 30-year veteran at the table (who gave me a great compliment about his suprise at how well the game worked sans clerics).

One thing that popped up is how far a torch illuminates, which isn't specified in OD&D. I started researching this, comparing across different rulesets, and the results were interesting. Partly this will be a critique about how games evolve towards greater abstraction over time, from realistic beginnings to nonsensical endings. It seems that the inertia, the infatuation with the game system itself takes over and late-version designers wind up working in an echo chamber. (And it's not just D&D: I've seen the exact same thing happen at games I worked on at a few video game companies. Perhaps it's true for other media as well, like books, TV, and movies.)

Question: How far does a torch let you see in reality? Consider this snippet from a Scientific American Supplement: "Torches consist of a bundle of loosely twisted threads which has been immersed in a mixture formed of two parts, by weight, of beeswax, eight of resin, and one of tallow. In warm, dry weather, these torches when lighted last for two hours when at rest, and for an hour and a quarter on a march. A good light is obtained by spacing them 20 or 30 yards apart." This indicates a bare minimum radius of visible illumination of 30 feet (half of 20 yards), maybe 45 feet (half 30 yards); possibly even 60 or 90 feet (20 or 30 yards itself) depending on how liberal the above usage of "good" is taken.

Question: How far does a torch let you see in D&D? In OD&D, the issue is seemingly not addressed; without directly comparing them to torches, the light spell is given a 3" radius, and the continual light spell a 12" radius (ostensibly 30 and 120 feet). In the AD&D 1E PHB a torch is given a 40-foot radius (p. 102, quite compatible with the research above), and the light spell is described this way: "The light thus caused is equal to torch light in brightness, but its sphere is limited to 4” in diameter." (Note that the second clause highlights the fact that while brightness is torch-like, the range of the magic spell is distinctly and intentionally shorter: just a 20-foot radius.) The continual light spell is reduced to a 6" radius, yet "its brightness is very great, being nearly as illuminating as full daylight".

Let's skip ahead to 3E D&D. Clearly some designer wanted to synchronize all of these effects and make them identical, a pretty reasonable motivation. If a light spell has been compared to a torch, why not make it equivalent to a torch in all ways, for brevity's sake? Well, the problem arises when this late-era designer doesn't do any research, and takes as his basis (looking solely from inside the rules) the effect of the magic light spell, and revises the effect of the mundane torch to match it. Thus in 3E you have both normal torches and the various light spells illuminating only a 20-foot radius.

Now, not only is the 20-foot radius torch unrealistic (whereas it formerly was), it's also extremely awkward from a gameplay perspective. The torch bearer only lights up 4 spaces (3E) away; routinely you'll have the front-line party member in darkness, or the front-most enemy in direct melee unsightable, or the extent of most rooms indeterminable during routine exploration, if you adjudicate this literally. (Now in 3.5E both light sources were given a new rules category of "shadowy illumination from 20 to 40 feet", but don't even get me started about trying to adjudicate that.)

The truth is that I'd recently been looking at the 3E SRD spells listing with its 20-foot radius torch, and so made a similar ruling in my game this weekend, and did get a look of disbelief from at least one of my players at the awkwardly short range of the party's light. And, I see now, he was right (in both realism and gameplay), my being led astray by late-era D&D rules-mechanic navel-gazing. At this point I have half a mind to say that torches give "good enough" light up to 60 feet away, illuminating most rooms in their entirety, and just using a whole 12" ruler (at 1"=5 feet) if we ever need to check it in play.

A rule-of-thumb I discovered over 10 years ago at one of my game programming jobs, and refreshed at times later on (even while building miniature models not long ago): If stumped by a particular design problem, ask yourself "What solution is used in the real-life situation?" In my experience, the answer is usually immediately applicable as a solution in your game rules. I'd guess that only a fetish for over-abstraction in a game would lead one away from this principle. I'll say again that we don't want realism-for-realism sake (see DMG p. 9), but for pre-existing gameplay problems, it often provides the most elegant fix.

There's other stuff about the interaction of light in published D&D that's bugged me over time (like the effect of the darkness spell, SKR's absurd-but-successful rant "infravision and why it should be destroyed" in 3E, etc.) That may have to wait for another posting.


  1. This issue came up for my own gaming group recently when we all went on a camping trip together. (It wasn't planned as a gaming group outing, it was just coincidence that the people who could go was the same group I game with.) We were comparing modern gear against some gear I brought that I use in colonial reenactments, one of which was a tin single-candle lantern.

    I suppose the reflective surface of the tin increases the illumination to some degree, but even then the light was so dim that when used to walk around the dark paths of the camp site, I had to hold it down near my feet. By this means I could at most see about three feet ahead of me, enough to be comfortable about where to place my next step and no further.

    It was a pretty cool moment for my players who have attempted to explore some dungeon interiors by candle light. Though I held up in game that the candle light barely illuminated anything, I think they thought I was cheating them until that moment.

    I'd love to do some other real world experiments. I bet you could get a juggling torch and carry that around a dark place to see how far you can really see by it. I'm not sure where you might find an oil lantern.

  2. Paul, that's awesome! I must admit I wouldn't have expected that with the candle, either.

  3. My last experience with 3E torchlight radius involved some goblins somehow knowing that our vision extended exactly 20' and using it to make tiny steps just outside of our vision range.

    I'm torn about whether thinking it was sort of cool since yes there is an advantage to having better visibility and this is how that emerged from the ruleset, or wondering how the goblins knew exactly how far we could see and so throwing this on the pile of the bad effects that a gameable tactical system (the combat minigame) has on D&D.

    Anyway, the way 3E makes up for its crappy torch range is with the plethora of magical light options available from very early on.

    As for SKR, his post would have been more convincing if he'd produced a picture of the "cloud of heated air" around someone (the couch picture is clearly a heated-up couch probably of someone who's been there awhile) which he uses in so many of his kvetches. In fact the face picture shows there's no cloud at all. Another good chunk of his reasoning is of the "there's no rules for the effects of rain therefore the ground is never muddy" variety which can be simply ignored. He even tries to have it both ways by saying it's stupid that a 60' infravision range won't let you see a red dragon at 70', but also that because of the magical air halo an infravision viewer would have reduced vision abilities; wouldn't one explain the other?

    All the rest (gee, illusion spells create weird situations?) can be swept away as "it's magic, innit?".

    But he's making his point inadvertently--infravision is so tough to adjudicate that SKR can't get it right. (Not that I really think he's trying.) So maybe the shaky simulation of a bad DM is worse than the mindless adherence to easy rules, after all.

    I don't love infravision and wasn't sad to see it go, but I respect a straightforward "simpler is better" more than that garbage. Where's the article on how unrealistic it is to calculate the volume of a fireball?

  4. The quality of light, eye sight, and what is being illuminated is what would really matter the most. I haven't done all that much research, except experience with camp fires and such.

    Light a candle in complete darkness and what you get varies by the wick and other factors.

    With a cheap but large emergency candle, I got poor light I wouldn't read by. But the light expanded to at least 15 feet or so, which is where it got to be indistinct. Mind you, that was for looking at walls which reflect well, not foot paths as in a previous post.

    I'd love to see some real work measurements. What I would guess is decent torches and lanterns would give off bright enough light to read only close up; but decent light to 20 or 30 feet in most enclosed areas where it reflects more; and dim light beyond.

    "Beyond is important". In a warren of mostly flat and white or grey stones, you could probably see the shape of "walls" to 40 or 60 feet. Not traps and doors and such -- just "There's a wall and opening over there."

    Add multiple torches and you get more light.

    What D&D designers would have had to do is add rules for number of torches and so on and they kept it simple.

    I myself find some things too stifling; the further characters can see the more they can map at any time, so expanding the radius of light reduces time spent measuring and mapping. So, players can get on with exploring.

  5. 3e's "shadowy illumination" means you can see everyone who isn't specifically hiding in the shadows, so it is illuminated.

    They should have perhaps defined a torch as still giving a 40' radius of illumination (as it does). Hide could then have said you can hide anywhere beyond half the radius of illumination, rather than trying to give every light source the "shadowy" tag with two radii.

    Anyway, I'd put a torch at about the equivilent of a 60W incandescent bulb, which is about 70 candlepower, for 8 times the radius of a single candle, or maybe 25 feet to barely read at. Close enough.

  6. How far can you see in torchlight in D&D? Far enough to make out detail when you're kinda close, and shadowy shapes and movement when you're kinda far. When you encounter a monster and determine distance, you can see vague detail at that distance, whatever it is, including the general shape of the monster ("you see ten orcs").

    Is it really that important to be more precise? In a room you can make out the shape of the room; in a corridor you can see to the encounter distance. If the referee decides some bits are too far away to be seen, then they are.


  7. At this point I have half a mind to say that torches give "good enough" light up to 60 feet away, illuminating most rooms in their entirety, and just using a whole 12" ruler (at 1"=5 feet) if we ever need to check it in play.

    The 12" rule also falls nicely with a movement rate of 12. So, you know where you're going, but you are not sure what you will see (there) ...
    Works for me

    1. Yeah, I think that's good enough. Adjudicating one end of a party not seeing the other, not seeing a full move ahead, etc., was really frustrating.

  8. This post is very late to the party, but i think will be helpful. I believe that the implicit rule for torches in the 1974 ruleset is 30' normal illumination with a dim and shadowy illumination out to 80'. I go into details in my blog post:
    I also cover torch illumination in all other iterations of D&D as well.

    1. Thanks for that reply and comment, it's helpful! I broadly agree, and it's about what I have in my OED house rules now, too (with 30'/60' rule-of-thumb, and same for infravision, too).

      There was someone on the ODD74 boards who ran a very nice experiment in their darkened house with similar results a few years back, too. If one has an account can see that here: