1e's Time in the Dungeon

The introduction to the AD&D 1e DMG mentioned that it's reasonable to omit random encounter wandering monster rolls if you understand the purpose of random encounters and it doesn't apply.  It doesn't tell you what that function is, though.  So I went looking for it and sadly didn't find it.

Instead, I found the table of how long various actions in dungeon exploration take.  It's split across a page boundary (pages 96 and 97) so I have reproduced it in text here rather than as a screenshot:

• DOOR - search for traps: 1 round
• DOOR - listening for noise: 1 round
• ROOM - mapping, and casually examining a 20'x20' area: 1 turn
• ROOM - thoroughly searching after initial examination: 1 turn
• SECRET DOOR - checking for by simple tapping of floor or wall, by 10'x10' area: 1 round
• SECRET DOOR - thorough examination for means to open, by 10'x10' area: 1 turn

I was very surprised to see the time-costs of some of these actions listed in rounds (which in 1e are minutes, 10 per turn, not 6-10 seconds) rather than turns.  I went back and checked and B/X (well, OSE) simply doesn't have times listed for any of the interactions with doors, and doesn't distinguish between types of searching - it just says that searching a 10'x10' area takes a turn.  I think I had been running listening at doors, searching them for traps, and searching for secret doors as taking a turn each.  I'm not sure how I feel about breaking the atomicity of the exploration turn and allowing it to be subdivided further into rounds.  I also find it very amusing that on page 97 shortly after this table of suggested times to perform these actions, Gygax complains about players who search everything and listen at every door.

Assume that your players are continually wasting time (thus making the so-called adventure drag out into a boring session of dice rolling and delay) if they are checking endlessly for traps and listening at every door. If this persists despite the obvious displeasure you express, the requirement that helmets be doffed and mail coifs removed to listen at a door, and then be carefully replaced, the warnings about ear seekers, and frequent checking for wandering monsters (q.v.), then you will have to take more direct part in things. Mocking their over-cautious behavior as near cowardice, rolling huge handfuls of dice and then telling them the results are negative, and statements to the effect that: “You detect nothing, and nothing has detected YOU so far —“, might suffice. If the problem should continue, then rooms full with silent monsters will turn the tide, but that is the stuff of later adventures.

But...  my brother in dice, you set the time cost to perform these actions so low that players would be stupid not to search for traps and listen at every door.

Maybe it's stupid to require a turn to listen at a door - but it works well and it's an actual choice!  And if it takes a turn to listen at a door, search it for traps, or try to pick a lock, then maybe you really don't need the no-retry clauses.

This does force me to consider that maybe I should give the party "free" attempts at finding secret doors in passing, assuming that they're tapping as they go because it's quick, and then make a successful turn-length search primarily for finding the mechanism of action / trigger.

There is also an interesting bit in here:

A gnome, for instance, must remain relatively quiet and concentrate for a turn to detect facts about an underground setting. Likewise, a dwarf must work at it. An elf doesn’t detect secret doors 162/3% of the time by merely passing them unless he or she is actually concentrating on the act. A character with a sword must have it out and be thinking about its power in order for the weapon to communicate anything to him or her. To sum it all up, DON’T GIVE PLAYERS A FREE LUNCH! Tell them what they “see”, allow them to draw their own conclusions and initiate whatever activity they desire.

Emphasis mine.  I think grouping in the use of magic sword detection abilities with these inherent racial abilities is telling about the expected frequency and ease-of-use of sentient swords.  There are no swords with detection abilities on table III.G. Swords on page 124 - swords with detection abilities only arise from the Sword Primary Abilities table for "unusual" (sentient) swords on page 167.  In conclusion, further evidence that the sentient sword rules are significant and have a purpose.

AD&D 1e's Time Sinks

I've been working my way through the 1e DMG and translating/summarizing the DMing advice out of Gygaxian into plain english for my own use.  I'm not that far through it yet but one thing that is constantly sticking out at me is that there are many time sinks.  For example:

• Every time you go digging through a pile of rotting crap or spend time holed up in unsanitary conditions you might catch a disease, which may cost you 2-4 weeks of downtime (and possibly ability score damage, death, blindness...).  You can also catch parasites from drinking bad water, with similar effects.
• As an aside, the paladin's disease immunity sounds amazing if the disease rules are played straight
• I'm probably going to end up writing another post about this at some point.  I think there's a good disease system waiting to be derived from this system - one not as crushingly awful as LotFP's corruptions and not as "all-or-nothing" as B/X's save-or-die diseases.
  
• When a paladin hits 4th level, he gets a vision of a horse.  Not a horse, just a vision.  And then he gets to go on a ~2 week quest to get his special mount.
• Thieves and assassins can go spying and infiltrate organizations.  Simple missions take up to a week, complex missions take multiple weeks.
• High-level assassins can study poisons, which takes many many weeks (and costs lots and lots of gold).  Once they've done their course they can create pretty arbitrary poisons, which takes a week or two.
• If you miss your initial window for Cure Disease after contracting lycanthropy, one of your options to remove it is to spend a month or so in an abbey drinking holy water infused with wolfsbane out of a silver chalice (for a considerable donation to the monks hosting you, of course).

That's all in the first 23 pages.  It's not like there's a big list of "campaign activities that take time", it's all mixed in with everything else, and if I didn't have time on my mind I probably wouldn't see it.  But it's there.  Lots of ways for time to get spent.

Maybe it would be fun to compile that list of all the ways to spend downtime (or to have downtime inflicted on you) in 1e.

Seasonal Overland Travel Times

I stumbled upon Rick Stump's blog recently.  There's a lot of good stuff here!  There was a line from this post that got me thinking:

This means that if there is a good road directly between me and, oh, the market 8 miles away if I leave at sunrise (call it 6:30) and want to return home by sunset (call it 8:45) [I am using the sunrise/set times for Seaward for this time of year] then I wouldn't get to the market until about 9:30 and need to leave by 5:45.

As I've noted elsewhere [2], I'm always looking for more ways to inject seasonality into the wilderness game.  And varying travel per day with available daylight makes a lot of sense!  Taking fall and spring with ~12 hours of daylight per day as your baseline, we'd see negligible seasonal change in available travel time per day in the tropics, but even as far south as Miami, the shortest day of the year is 10 and a half hours, while the longest is around 13 and a half.  So that's a difference of 12.5%, an eighth.  Up here the shortest day of the year is around 8 and a half hours, while the longest is more like 15 and a half.  So just on the basis of having more or less daylight, something like a 30% reduction in overland travel rate in the winter and a 30% increase in the summer would be reasonable at these latitudes.  Dealing with modifying travel speed by an eighth might not be worth the hassle, but 30% certainly seems big enough to warrant consideration!  For latitudes somewhere in between Miami and Seattle, something like a 20 or 25% modifier might be reasonable and fairly easy on the math.

And then if you're already figuring seasonal overland travel speed modifiers for daylight, you could also just abstract weather into that modifier, especially if operating on scales of weeks.

One could also make the slower (winter) speed the default, and then scale everything else up accordingly, since multiplying is easier than dividing.

(Granted - this ignores historical details like the afternoon halt during summer marches.  Maybe the difference between getting 12 hours of rest in the dark and getting eight at night in the summer plus four in the middle of the day doesn't work out to that big a difference in practice.  Maybe it makes sense as just a winter penalty, where your time to break and make camp cuts into your marching daylight time.  And then there are those rascals with infravision mucking things up as usual...)

Week as Wilderness Turn

Last post, I noted that the distances and times involved in wilderness travel in The Hobbit were so great that a game in that style might want to use 24-mile hexes as its smallest unit, and a week as its "turn" of wilderness travel.  What does it look like if you take B/X's or ACKS' rules for wilderness travel and rescale them this way?  What would you probably have to change and what could you leave alone?

If you're taking a week as your base unit of time, then you no longer have to worry about taking a rest day per week, and can just bake it in.  So you're scaling up your movement or actions per "turn" by a factor of 6, since you get six working days.  Using 24 mile hexes vs 6 mile, you're scaling up your unit of distance by a factor of 4.  So movement will be ~1.5x more hexes per turn.

Exploration movement speed : 24 mile hexes / week

• 30' : 1.5
• 60' : 3
• 90' : 4.5
• 120' : 6
• 180' : 9
• 240' : 12

So that's rather inconvenient, having to reckon in half-hexes.  I suppose we could do something like this, taking woods or hills as the default and multiplying everything by 2/3, so we get 1 24-mile hex per week per 30' of speed.  Which would be pretty clean, until you're on plains or road.

Nautical and aerial distances covered get quite large when you look at them on per-week timescales.  The lowly rowed canoe makes 4 hexes per week, while sailing ships with good winds might make nearly 8 hexes per day, or north of 50 hexes per week (assuming no weekly day of rest under sail).  Aerial travel gets pretty nuts when you multiply it all by 1.5 as well.  These modes of transport are fast enough that you probably don't want to track them hex-by-hex on a map on weekly timescales. You use sailing ships to move between maps.  You fly Eagle Airways for a couple of hours, not for a solid week (and still might get to move a hex or two).  Getting a personal, permanent flying mount is a phase-change event where you have outgrown thinking much about wilderness travel.

Since forced march is only one day of extra speed, followed by a day of rest and no speed, if we're dealing purely in weeks you might be able to just remove the option.  On the other hand, I like leaving this sort of option available to players; trading off now vs later is always interesting.  Maybe the right way to handle forced march on this scale is through something like strategic initiative in mass combat, where when you have a wilderness encounter, you can choose to forced-march to maybe gain a terrain advantage, bonus to escape roll, or surprise bonus at the expense of fatigue if brought to battle.

The right way to handle getting lost / failed navigation rolls is probably that they reduce your movement for the turn.  Maybe halve it; you spend a couple days wandering around within hexes that you were traversing, but going 24 miles out of your way is hard.  Could do hunting the same way; spend the whole week hunting, don't move at all, and get two rolls, or spend half the week hunting, get one roll, and half movement (and then if you also get lost that week, you end up still in the hex you started in).

Rations might actually get simpler, at a stone per man per "turn".  And fresh ration decay could be simplified too; you could just have all uneaten fresh rations go bad every turn.  Then hunting and foraging can work entirely within a single week; if you find a week's worth of fresh rations, that's a stone of iron rations that you can skip eating and carry over into next week.  So then you only need to track one number: the stone of iron rations you're carrying, which might also be thought of as your buffer against foraging failures.

As far as combat resources go...  for a Hobbity feel, you really want some refuges in the wilderness where you can recover hit points.  I still think recovering spells there too (not every night in the wilderness) makes sense, provided some reworking of mid- and high-level spells.  But I could see going the other way with it too, where you're pretty much always going to have full spells for any encounter.  This lends itself to very large encounters where you need lots of spellpower to bail the party out.

Wilderness encounter frequency definitely gets weird.  You probably don't want to have to roll a pile of d6s every "turn", and having multiple encounters per unit time is awkward.  I could see having one encounter per week, with the difference between terrain types being "roll n encounters and pick the scariest / biggest one".  Or just a quantity multiplier like dungeon level, where if you're in mountains you get 3x as many goblins as if you were in plains.  But this also doesn't quite square with the frequency of wilderness encounters in The Hobbit, where they can go a couple weeks without an encounter.  I could see having some <100% chance per week of an encounter, but when there's an encounter, it's always a lair - a kingdom of elves, a whole cave system of goblins, a big honkin' pack of wargs, a gang of trolls with accompanying cave full of magic swords of elven make.  On the time-and-space scales you're dealing with here, you might encounter a warband from a lair initially, but within a week they'll report back (or be noticed missing) and you'll be dealing with the whole village shortly.  Maybe winning surprise lets you only deal with one warband initially.

Domains get...  maybe a little messy.  Clearing a 24-mile hex is a lot of work.  You could do something like wilderness lords, where every wilderness hex already has a "lord" of a sort, and if you can knock him and all his monsters over that's good enough, the rest migrate or fall into line.  This is particularly plausible in a setting where everything talks, but might feel a bit strained after the third or fourth time, and maintaining relationships between all the "lords" of neighboring hexes is a lot of work.  Another approach might be clearing to capacity; if you want to build a village, you have to displace a number of HD of monsters comparable to a village of goblins.  This is what you might expect in a wilderness that is at capacity, saturated.  But this isn't the wilderness we see in The Hobbit, which is as post-apocalyptically empty of monsters as it is of men.

Maybe that's an answer - "clearing" a wilderness hex just means dealing with any already-known lairs in it, and then the real game is dealing with wilderness random encounters, which could have wandered in from nearby hexes, or could be from unknown lairs in the hex.  Taking land is easy; holding it is the hard part.  This creates a sort of "the dungeon is too big to be cleared" feeling, and is also consistent with the incomplete clearing of eg Mirkwood by the elves.  It pushes domain rulers into the same sort of reactive posture of incomplete control in game that they held in the fiction.

Switching to 24-mile hexes changes visibility somewhat, in that the edge of the hex is over the horizon.  You could spend a week exploring within a single hex, easily, and unless the next hex over is elevated, you might not know what terrain type it is until you enter it.  Finding a dungeon within a hex might take a while, but that's consistent with the fiction too, where they can't find the back door into the mountain.  They have trouble finding Rivendell too.

This mode of play doesn't seem terribly well-suited to 1:1 timescales, if most weeks you don't have an encounter.  I could see doing two turns per week of real time though, to make sure there's time to resolve wilderness encounters without too many more piling up.

Tooling for 1:1 Timescale Games

I've had boring, practical questions about how to run 1:1 timescale games rolling around in the back of my head for I guess a couple months now.

• World-design considerations.
• Many of the games of that era seemed centered around a single city and/or megadungeon (Blackmoor, Castle Greyhawk, and the City-State of the Invincible Overlord, for example), and Jeff's Trollopulous seems to have followed a somewhat-similar pattern.  How important is having a central locus like this?
• Is the right way to lay out a region for this style of play more radial, with wilderness increasing out away from The City, rather than a linear civilization->borderlands->wilderness gradient from a large, civilized heartland?
• What's the right scale, so that travel is long enough to be a meaningful allocation of time-resources, but not so long that the game grinds to a halt as everyone is traveling?
• Session structure.
• BDubs' game seems to have a core community of players who play adventuring PCs and have a regular session, often dungeoneering, every week, plus some set of patron players doing their own thing in the background.  This seems like it doesn't scale up arbitrarily (but what does, really?).
• Jeffro's adventurer-level players seem to play more like patrons, doing their own thing, maybe jumping an injured patron and taking his stuff, rather than partying up and having regular sessions where they go into dungeons.  
• And then there's Gygax's example, where (purportedly) people would show up at his house unannounced and he'd run a session.
• Under what circumstances do which of these different approaches to sessions (or lack thereof) work well?
  
• Keeping the information organized.  What's the right (or at least adequate) way to keep the information about who is doing what when where (and when they'll be finished) organized?  And how do you keep (at least some of) that information secret, to let the interesting behavior emerge from patron interactions under fog of war?
  
• Index cards?
• Spreadsheets?
  
• Shared google calendar?
• Paper calendar that only the DM has access to?
  
• Gantt charts (lol)?
• Database?
• How do you make such a style of play amenable / accessible to more casual players rather than just "elite auditor mindset" players?  (I'm bummed that I can't find either BDubs' or Jeffro's post where that phrase was invoked)
  

I think I have an angle of attack on one of the information organization question, at least.

The answer is databases.

This post brought to you by my assignment at work this week, auditing code that interacts heavily with databases

Why?  Because you can automate around them.

The barest, crudest sketch goes like this: have a table of ongoing operations with start dates, end dates, and descriptions.  When someone starts something, add it to the table.  Have a table of players, including their discord IDs, and a table containing (operation ID, player) pairs of "players who should be notified when this operation finishes".  Have a discord bot that runs every night at midnight, checks to see if any operations have finished on day rollover, marks them complete, and notifies players (and probably DM) by PM.

Ha, maybe run the bot at 5PM instead, right as people are getting off of work - a character's work for the day is done at that time too.  And any "orders" gotten in before 9AM the next day are completed by 5PM the next day.

But this doesn't really let you answer the question of "what is any particular character doing at the moment?"; there's no notion of character there.  So maybe we also want a table of characters (including an "owning player" column), and then a table with (operation ID, character ID) pairs indicating "this character is currently unavailable for play because they're busy with this long-running thing".

Of course, once you have a table for characters, it's very tempting to throw their stats and XP in there too...  and a table for all the unidentified magic items they're hauling around, with both the in-world description you've given and the actual identity of the item.  And how many charges that wand that they figured out the command-word for but never really properly identified has left.

It would be kinda nice to put wilderness expedition logistics information in there too and update it on nightly rollover...  but that gets a bit thorny because then we need a notion of a party, which presumably can have a set of characters (as well as a count of NPCs like mercenaries and shared equipment like wagons), and which might be busy with an ongoing operation (like overland travel).  So then we run into this: do we have a table of party/operation pairs, for parties that are busy?  Do we then query party membership to determine which characters are busy, or also still have a character/operation table and have to check both it and party membership?  Or do we just have a party/operation table, and have parties of only one character sometimes?  That might be kind of nice, for the case where you pick up friends along the way (or lose friends along the way and a party goes down to just one member).

Do we want some notion of location?  Of grid coordinates?  (I really hate programming against hex grids)  Parties having a location which is updated overnight based on their overland speed?  (But then I would need to encode their paths too, which sounds painful...  maybe this is a good argument for Arnesonian "12 miles a day regardless of terrain")  But it would be neat that if there are multiple parties in the same location in the wilderness, they might have a chance meeting.  Automatic random encounter determination, putting a pause on the travel operation until resolved?

If we're already considering notifying players in response to stuff in the world, we could extend that for information flow and intelligence.  Maybe having a spy in a location adds you to the notification list for start and completion of operations in that location.  Maybe operations should have secret vs common-knowledge descriptions, with spies having some chance to uncover the secret descriptions and town gossips yielding the common-knowledge ones.  Spies might dispatch messengers in response to some operations beginning or ending, creating operations in their turn for the messengers to travel to their destination.  Some notion of queuing news at a place (to be received on your return) might be necessary.  Automatic weekly collation of common-knowledge events in (say) a central city into a "News in Kezmarok" style digest sent to all players could be neat, and a good way for players to advertise "looking for group" or patrons to advertise "looking for temporary, deniable help".

Finally, NPC actions.  If you have all this tooling for keeping track of what PCs are up to, it seems natural to extend it to also keeping track of what NPCs are up to.  Clocks on steroids; give "orcish warlord gathers horde" a concrete end-date and set it up to notify just you, the DM, when it finishes.  And if players interfere, move the end date around accordingly.  Likewise, if you want to send an NPC party on a wilderness/trade expedition, make a series of operations, one to travel there, one to buy and sell goods for a week or two, and one to travel back.  NPC party raids a dungeon?  Make a single die roll to see how they did, make an operation where they take a week of bedrest and recruiting, and then put them out of your mind.  They're in motion, you don't need to worry about them for a week.  But your players might hear about them automatically.

This starts to look a lot more like a living world.   And the table of operations starts to look like a log of all the stuff that happened outside of combat, which would be a pretty neat artifact of a campaign (much like player-produced maps are a neat artifact). 

It also seems like it would be pretty easy to keep it game-system-agnostic.  Character stats seem like the most likely place to break this, and I'm not even sure I really want them.

Current 1:1 Timescale Campaigns

So apparently I haven't been the only person taking an interest in Gygax's timekeeping lately.  I stumbled into a cluster of blogs describing three campaigns currently being run with 1:1 timekeeping:

• Bdubs' ACKS campaign
• Bdubs discusses observations on the effects of different timekeeping strategies in play here
  
• Jeffro's AD&D campaign (which has been running for about a year now)
• Waterdeep OSR's 5e campaign (this post also discusses the how to effectively delegate the activity of high-power NPC patrons to players)
  

This seems like a very natural/consistent/worthy next thrust for the OSR.  Having recovered the early rules, and made them readily available, rediscovering and repopularizing the culture of play of the early days for which those rules were intended.

It kind of tempts me to run Traveller in this style...  the most obvious difficulty is figuring out the resources available to patrons, and Striker's rules for planetary GDP would provide a very workable starting point.

"A Meaningful Campaign"

I was thinking in the shower about that oft-quoted bit of Gygaxiana:

YOU CAN NOT HAVE A MEANINGFUL CAMPAIGN IF STRICT TIME RECORDS ARE NOT KEPT.

(all-caps his).

His detractors use it to make fun of him - look at this guy, taking strict timekeeping too seriously and getting mad enough to put it in all caps.  His adherents take it as dogma.

But I got to thinking about Boot Hill's perspective on campaigns, and started to wonder - maybe he meant a structure with much more autonomy for individual players than we typically think of in campaigns.  He's not talking about an "adventure path", obviously, since the quote is from the AD&D 1e DMG and Dragonlance was yet a twinkle in Weis and Hickman's eyes.  Less obviously, I think he's not even talking about the way OSR campaigns are usually run, where the party has a lot of autonomy but players are more or less bound together by it.  One example of this position from Dungeon of Signs - "Old style D&D is not a story about any one PC, it's a story about the adventuring party as a whole, or ultimately about a fictional world as a whole."

I got curious about the context around the Gygax quote, so I dug up my previously-unopened pdf of the 1e DMG and lo and behold, a paragraph after "YOU CANNOT HAVE...", we get:

For the sake of example, let us assume that you begin your campaign on Day 1 of the Year 1000. There are four player characters who begin initially, and they have adventures which last a total of 50 days — 6 days of actual adventuring and 44 days of resting and other activity. At this point in time two new players join the game, one of the original group decides to go to seek the advice of an oracle after hiring an elven henchman, and the remaining three “old boys” decide they will not go with the newcomers. So on Day 51 player A’s character is off on a journey, those of B, C, and D are resting on their laurels, and E and F enter the dungeon. The latter pair spend the better part of the day surviving, but do well enough to rest a couple of game days and return for another try on Day 54 — where they stumble upon the worst monster on the first level, surprise it, and manage to slay it and come out with a handsome treasure. You pack it in for the night. Four actual days later (and it is best to use 1 actual day = 1 game day when no play is happening), on Day 55, player characters B, C, and D enter the dungeon and find that the area they selected has already been cleaned out by player characters E and F. Had they come the day after the previous game session, game Day 52, and done the same thing, they would have found the monster and possibly gotten the goodies!

Some penalty must accrue to the non-active, but on the other hand, the over-active can not be given the world on a silver platter. Despite time differences, the activities of the newcomers to the campaign should be allowed to stand, as Destiny has decreed that the monster in question could not fall to the characters B, C, and D. Therefore, the creature was obviously elsewhere (not dead) when they visited its lair on Day 52, but it had returned on Day 56. Being aware of time differences between groups of player characters will enable you to prevent the BIG problems. You will know when the adventuring of one such group has gone far enough ahead in game time to call a halt. This is particularly true with regard to town/dungeon adventures.

 (emphasis mine)

Can you imagine the drama that would ensue if two new players joined your campaign (group?) and your existing players decided to not go adventuring with them?  I suspect that even in the vast majority of OSR campaigns today, this would be seen as a gross violation of social norms.  Even the West Marches assumes a cooperative community of PCs, a sort of "whoever's at the table that day goes adventuring together".  And yet here Gygax writes as if it were perfectly normal, as if association within the game-world was never taken for granted (and this example wasn't even one of those 20+ person campaigns mentioned in OD&D Book 1 - this is six players!).  This is what I mean by "no party" - there are player characters in the world and they interact with each other as they will.  Sometimes they work together to achieve their objectives, but that's a choice.  A party exists only for a single adventure.  Outside of an adventure, you can do as you please.  You want to go haring off after some oracle (for about a month, in the paragraph following the quoted one)?  By all means!  You want to spend a month just resting and healing?  Sure!  But you might get scooped on treasure.

It makes a lot of sense that if you have that norm, of players going off and doing their own things, and splitting into multiple parties, then yeah, you are going to want timekeeping!  As he says, to prevent "BIG" continuity problems.

The time lines of various player characters will diverge, meet, and diverge again over the course of game years. This makes for interesting campaigns and helps form the history of the milieu. Groups of players tend to segregate themselves for a time, some never returning to the ken of the rest, most eventually coming back to reform into different bands. As characters acquire henchmen, the better players will express a desire to operate some of theirs independently while they, or their liege lord, are away. This is a perfectly acceptable device, for it tends to even out characters and the game. Henchmen tend to become associates — or rivals — this way, although a few will remain as colorless servitors.

You may ask why time is so important if it causes such difficulties with record-keeping, dictates who can or can not go adventuring during a game session, and disperses player characters to the four winds by its strictures. Well, as initially pointed out, it is a necessary penalty imposed upon characters for certain activities [mostly magic item creation]. Beyond that, it also gives players yet another interesting set of choices and consequences. The latter tends to bring more true-to-life quality to the game, as some characters will use precious time to the utmost advantage, some will treat it lightly, and some will be constantly wasting it to their complete detriment. Time is yet another facet which helps to separate the superior players from the lesser ones. If time-keeping is a must from a penalty standpoint, it is also an interesting addition from the standpoint of running a campaign.

Emphasis again mine.  A couple of interesting things in these bits.  One is that the "diverge, meet, diverge again" is very much in the model of, say, Thieves' World, where there are persistent characters who make temporary alliances.  I have heard fiction in this form referred to as picaresque, but looking up the actual properties of picaresque I'm not so sure (I do like almost all of those in my D&D though).  The temporary nature of "bands" is again emphasized.  "interesting set of choices and consequences" is a very gamey thing to say, as is "separate the superior players from the lesser ones".  Gygax here is talking about D&D as a game that you can be better or worse at, and I think would not only say that it is acceptable for some players to fall behind as a result of bad choices and bad luck, but desirable.

So, returning to "meaningful campaign".  It seems that "campaign" here is meant much like the Boot Hill style of campaigning, a big set of interacting characters and a series of skirmish actions with a degree of continuity.  It seems that "meaningful" is meant in some combination of "maintaining internal consistency" and "preserving the agency of each individual player, that their choices about how to spend their resources [time] have consequences."

I'm really curious to see how this notion of great individual agency squares against the practice of having a caller.  I get the feeling I'm going to end up doing a lot more reading from the 1e DMG in the near future.  I found this a more thoughtful position on time, and better-articulated, than I expected.