I will describe two different dungeoneering doctrines and then have a few comments after them. Doctrine, here, means a set of tactics one has available as tools. (The maps are player maps by diog.)
The slow crawl
The campaign is Rajamaat 2 and the rules chassis is Old school essentials with advanced fantasy rules. The group has been playing for a while. I am mostly a newcomer to the group, though I have played in a few sessions with them before. We are just going down into some dungeon associated to the god Set, related to poisons, evil, murder, snakes and other such fun things. We proceed carefully, investigate a suspicious painting to no benefit, then find suspicious statues and pressure plates in front of them. Putting up pressure with a ten foot pole does not do anything, so we get a pile of stones from outside, put one of them onto the pressure plate, apply pressure with the pole, and repeate until the plate a triggers: a deadly jet of flame! We figure out some secret doors, are scared of a statue of Set with a sphere of lava lamp -like greenish swirling, get some treasures from there by tying a rope to a table leg and dragging it outside the room before messing with the stuff in case the statue or the sphere is triggered, and so on.
No random encounters thus far. Progress is slow, but steady, though someone does take a risk and get cursed by the statue. Do not steal lava lamps from telepathic statues of evil gods, looks like.
This is the slow and careful dungeon crawling routine. We expect (and find) traps and secret doors, and try to trick our way past possible dangers while getting the presumaed rewards, that is, mostly treasures.
The fast raid
The game is Coup de (main in) Greyhawk and the rules are a work in progress originally loosely based on BECMI D&D, but no evolved to be very much their own thing. Many in the group have played a lot together; I and another active player less with them, but we have shared history in discussing roleplaying, plus then there are more random players coming in once a while. The core group is experienced, with pretty much everyone having been both a player and a referee in wargamey play. My player experience is on the low side; what I write here is new to me, but presumably well known to many of the other players.
The first example of a fast raid is taking over the barbican of Castle Greyhawk. With luck and scouting we managed to get us under the barbican without the enemy (a bandit group) noticing, entered the place, swiftly moved up and assaulted their barracks before they had even figured out we were there. We were victorious, but were afraid the leadership would be too much for us, given our wounded state. The leadership had been arguing on the rooftop of the barbican, which was our luck (the referee rolled openly for if they get their act together up there; they did not). We barred the doors, thinking to starve the enemies before taking them on. Unfortunately, being high level D&D characters (my best guess is fourth and second level), they descended, fell or jumped down and vanished into the night. Nevertheless, we managed to secure a fortified position that has been later used for further expeditions into the castle.
The second example is in Dolmenwood (spoilers ahead). A cult in Illmire captures a hireling of ours, which causes us to take action. We have figured out there is something afoot and the closed and barred church under reparation, as they say, is the obvious place for a cult hideout. Scouting confirms this. We do a ruse to get all their guardians in one place, assassinate them (the alignment of many of our characters is heading towards evil) and enter the church via a dungeon entrance in a crypt. Some confusion with walking dead and we are in, opening a door swiftly, everyone moves in to eliminate all opponents and securing all doorways, and then we move to the next room. The entire place is cleared with no complications, and some prisoners, but not our hireling, are rescued.
The cult also operates the local inn, so most of us get in position to watch it while our thief finds the hireling, they enter the hall, and everyone who tries to run or fight is captured. In the meanwhile a couple of us explain what is going on and get the townfolk to be on our side in case the enemy would try to raise a rabble against us, but everything goes fine.
Unfortunately the cult leader, a former priest, is in neither of these places and manages indeed to escape into a further hideout, which we still have not cracked.
In both of these examples the key was to move faster than the enemy information flow; we were inside their decision making (or OODA) loop and thus could get surprise attacks and not face a concerted force. In both cases the key figures managed to escape; we did not know how great resistance we were about to face and thus did not dare set people apart to stop them, and in the heroic world of D&D it is pretty hard to stop a high level (such as 4+) character from doing whatever they want, at least if you do not have overwhelming numbers or surprise.
Fast and slow
I think that fast and slow dungeoneering are two doctrines. They both consist of a numer of tactics – hard and soft room entry; far, near or no scouting; various problem solving and fighting tactics. Strategy is choosing which to use in a given situation. By listening to rumours about OSR one never hears even the faintest hint of the raid doctrine. You find it in no rulebook. (Or maybe you do, but I sure have not.)
«Don't split the party» is a meme; often splitting the party is an excellent idea. For low level dungeons a large party, totalling 10+ people (player characters, henchfolk, hirelings), seems to be a winning recipe and makes splitting the group even more natural.
A question I am interested in investigating is what other such illegible strategic matters there are in wargamey OSR play. What other invisible choices do we make without ever seeing them?