This is basically repeated in the entries for kobolds and orcs. It's always seemed to me that the −1 attack modifier is really too small to make any real difference. In fact: These days I'm of the mind set that I don't want to deal with any situational ±1 modifiers at all; the payoff from such things is almost surely not worth the mental effort spent tracking them. Certainly we can agree that a 1 in 20 difference does not resemble the complete abhorrence of daylight that you see in say, Tolkien's goblins or northern orcs.
I've known that this modifier had its origins in Chainmail, and assumed that the difference in scale to the combat mechanics was to blame, and that when viewed in those rules it would work in a rational and significant fashion. Let's actually look at that today, which I'd never done before. The interesting thing is that you have to join up three separate tables before you see the actual result in those rules. First, the text unit description (p. 29):
Note that the original had both a −1 modifier to morale and a −1 to "any die rolled"; this was included by reference in the entry for Orcs. Note also that this attack modifier was copied forward to any later version of D&D that we might care about. The −1 to morale was copied into Original D&D (which ostensibly makes use of Chaimail's inscrutable morale rules), but does not appear in Holmes or Moldvay Basic or AD&D, as seen above.This is slightly strange because Moldvay has excellent and highly playable morale rules in which the −1 modifier would be quite well-balanced, and technically there are percentage-based morale rules in the AD&D DMG (even if no one ever used them).
Second, let's look at the Chainmail Fantasy Reference Table near the end of the book (p. 43), where we can see exactly how well goblins and orcs normally fight:
As we can see in the 2nd-to-last column, each of these types attack as "Heavy Foot" (the middle of the 3 infantry types, equivalent to men in chain mail armor). But what does that mean, exactly? Now we flip back a few pages to the Chainmail Combat Tables (p. 40):
The "-" are organizational dashes (not subtractions), so see what the mechanic is here: Players attacking with these types of units will be rolling some handfuls of d6's, the number based on a ratio of attackers to defending types. In every case except one (light foot defending) kills are only possible when some dice show a perfect "6". Therefore: When orcs/goblins fight in daylight, and take a −1 modifier on their attack dice, then the highest they can roll is 5 and it is totally impossible for them to score hits on any opponent except light foot!
That seems like an amazing result when I noticed it tonight. In some wordings, you could sort of try to interpret it as "losing one of the dice they get to roll", but in truth the original language for Chainmail goblins, as shown above ("subtract 1... from any die rolled") can't be read that way -- it's not one subtracted from "dice", it's "any die", that is, every individual die clearly must get the same subtraction from its rolled result.
Amazing! As much as I always wanted a more significant penalty, that seems really stunning, making orcs and goblins entirely ineffective and impotent against almost any enemy when struck by light; how tremendously different that same -1 is in the d20 alternative system, where it is almost negligible. Recall that technically OD&D uses the Chainmail d6-based combat tables above by default -- scaling changes to the modifiers should have been necessary -- surely the difference is so huge it was an oversight and not an intentional adjustment.
Before we go, let's also check in on the little brown book that is Gygax's Swords & Spells fantasy miniature rules for use with (original) Dungeons & Dragons: here, all combat modifiers are given in terms of percentages (p. 24):
Notice that in full daylight, kobolds, orcs, and goblins are assessed a −30% melee output penalty (3rd row from the bottom). That's equivalent to subtracting 6 pips on a 20-sided die -- or close to 2 pips on every 6-sided die in a system like Chainmail, Warhammer, or Book of War! Even in "near full daylight" they take a −10% penalty, equal to subtracting 2 pips on a d20 (bottommost row). This is in addition to the −10% morale penalty that is specified elsewhere (p. 20). We might conclude that when Gygax was looking at the issue with intention, he assessed a severely heavy-duty penalty. Only when he was doing a mindless copy-and-paste job do we get the niggling −1 on a d20 modifier -- which is, unfortunately, how it appears in all the core rulebooks for OD&D, B/X D&D, AD&D, etc.
Looking at this closely, I'm actually motivated to revise my house rules and even Book of War miniature rules for orcs & goblins. Consider this: Let's say that the orcs & goblins D&D d20-based to-hit penalty "should have been" on the order of −4 or so in daylight. Then in a d6-scale like Book of War this becomes a substantive −1 penalty to hit, as in Chainmail; note that in my system they would then be helpless against opponents in plate armor (like heavy cavalry; "6" needed to hit by default), but would still have some chance to hit targets in leather or chain. So this would give greater interest to these types, and rationalize with what we see in Gygax's Chainmail and Swords & Spells games.
And it would also help rationalize the troop costs that he gives in OD&D Vol-3 p. 23 ("Orc support and upkeep is only half that of a man"); when I run these stats through our simulator program (link), preferred costs of just about half do in fact fall out of the system quite nicely (whereas previously there was almost no difference in effectiveness between orcs & men). If we do this re-interpretation, then Orcs & Goblins become a high-risk, high-reward gamble on the battlefield; you really want to avoid the weather being sunny (or at least make sure to bring a Wizard who can control the weather in your favor). What's your preference for that?