
What, it's the end of summer this week? Better go out with a bang...
In July, I had a sizable series of posts on those iconic and special D&D missile spells,
fireball and
lightning bolt. One of the things I highlighted was how they have special targeting rules ("declare distance of shot") unlike any other spell -- a legacy of their appearance in
Chainmail as analogs to catapults and cannons, in a completely separate category from what are otherwise called "spells" there.
See some of those posts:
here,
here,
here,
here,
here,
here, and
here. (Leaving out the posts on giant stone-throwing which likewise grew out of the same original artillery rules...). And, in one of those articles, I
asked the poll question,
"What's the best way to target fireballs?":

So that's pretty interesting because I don't think I've ever asked a poll question here that got such a wide range of replies, almost evenly spread over all the available responses. That said, one of the top ones was "Declare target & check 'to hit'... else apply variation"; that seems "right" to a lot of people, even though it appears nowhere in classic OD&D, AD&D, B/X, etc. -- although it was given that way in the more recent 3E. Also tied for the top was "Something else entirely", so I guess
fireballs are like opinions... everyone's got their own! Note that pretty much dead last was the option for "Pinpoint any target in range as desired".
But with such a divergence in opinion, and a pretty small sample size, for the first time ever I went and cross-posted the same question over at the Dragonsfoot forums, in the hopes of casting a somewhat larger net. Had some
good discussion here, and generated these responses:

Did that help me? Well, it's not yet double the replies I got on my blog, and the reactions are sort of upside-down from the poll here. Now, "Pinpoint any target in range as desired" came out first (with a narrow majority), but "Declare target & range of shot (legacy rule)" also made a strong showing. So... yet more difference in opinion.
One point of analysis is that while this blog is more focused on OD&D and its close descendants, the Dragonsfoot forums are of course dedicated to 1E AD&D, and don't make too much distinction between other stuff they lump together as "Classic". (Not to say that many individual members are not highly knowledgeable about every edition.) And importantly, looking at
Chainmail/OD&D you'll be staring at a very explicit "player calls distance on tabletop" rule, while in 1E AD&D this gets submarined into "caster declares distance of shot", which is very easy to interpret as being in-character-flavor and no longer player-skill-mechanic. So the large difference in opinion between visitors, even between OD&D and 1st-Edition sites, is not so surprising.
So what to do for
Original Edition Delta? Well, this important interpretive issue was one of the last things that I had to "get right" for my upcoming miniature wargame rules,
Book of War. And as I thought about it deeply, using the great responses from these polls as fodder, I came to the conclusion that
fireballs and
lighting bolts simply had to maintain some kind of increased variance, intrinsically baked into their mechanic, for balance and other purposes.
I mean -- We all know how crazy-powerful these key 3rd-level missile spells are. In classic D&D, they actually spit out significantly
more damage than higher-level spells like 4th-level
ice storm, 6th-level
death spell, or even 9th-level
power word kill, for a caster of the same level (if you crunch the numbers). So what's the balancing factor? The thing that was in the core rules from the inception was the special targeting uncertainty.
But on playtesting I found that the direct legacy "declare your distance" was simply far too difficult to succeed at. And more generally, it would be wildly dependent on player skill, such that someone who's spent time practicing their estimation (and thinking to pre-measure the table dimensions, say) would be at a monumental advantage over a newbie player, which was simply unacceptable for the kind of game I wanted to produce (namely: immediately accessible to non-hardcore gamers). Furthermore, that kind of wide-open player-skill mechanic was not something I could estimate or analyze when cost-balancing the different units and powers.
So having spent all that time dissecting and thinking about the issue, here's the rule as you'll see it in
Book of War:
Roll one die for accuracy: 1-2, 1" short; 3-4, on target; 5-6, 1" long.
Thus, we maintain some kind of variation, which is fundamental to the powerful
fireball and
lightning bolt spells. It's not outright player-skill, but it is a callback to the original
Chainmail cannon rules [CM p. 14, top picture], for example, where all this stuff springs from. Shooting into an enormous formation of troops is easy/automatic. Picking off an individual hero or monster is much harder to accomplish. I've already had some excellent and nail-biting wizard-vs-wizard long-range shootouts with my friend
BigFella using this rule. (Side note: This rule is, for its brevity, not unique for the amount of backstage justification that went into producing it. The whole book is pretty much like that.)
I'll leave with one keenly important insight. For this mechanic to be justifiable,
fireballs must be inherently more variable than archery. Like, based on our normal-curve ballistics insights (
here), if you want to shoot an arrow and get within at least 10 or 20 feet of a target, then that's close to automatic at any range, for any trained bowman. Not so with a
fireball, even for highly-trained and high-level wizards! So it's not just a matter of "presume the wizard has learned how to shoot a
fireball properly"; what we're concluding here is that a
fireball is specifically a weapon with such a high rate of variance that it's impossible to be as accurate as anyone is with a bow. They're like the earliest forms of wrought-iron cannons (in so many ways, again as per
Chainmail), which were enormously inaccurate compared to other weapons (look up "obturation"). I guess this brand of magic is simply undependable and not-fully-controllable in that way (but it's still generally okay, since it's an explosive area-effect blast, after all).
At some point I'd like to come back with exact statistics at the man-to-man level for just how inaccurate a
fireball must be (back-reasoning from the simple mass-combat rule). But for now consider this: If landing an arrow or a flask of oil in a particular space is a to-hit against AC10, then doing the same with a more-volatile, harder-to-control
fireball shot must be significantly more difficult: maybe AC4 or harder?
(Top image by santus, under
CC2.)