I had a shipboard battle in a game I DMed. Shoving enemies overboard turned out to be the dominant strategy, even better than manning the heavy arbalests mounted on the deck.
Riffing off this for the most important lesson from my last campaign:
Around level 7 or so, I decided that my nautical (Saltmarsh) campaign was due for a good old fashioned pirate attack. I set it up real good: see the pirates from a distance, slow approach, pirates have artillery (fireball-lobbing wizards) and submarine support (sahuagin and dire sharks towing the ship) and command-and-control (a couple of clerics). I imagined a fight starting at a distance with the artillery, closing to be a hand-to-hand combat on deck, all while the PCs tried to protect the low-level NPC crew of their own ship.
Nope.
The PCs used water-walking, wildshaping, water breathing, and other magics to abandon their own ship and send it away. They attacked entirely underwater, rendering most of the pirates irrelevant for most of the fight. The whole thing had little to do with the magnificent set-piece battle I’d intended.
The moral of the story was, “DM makes plans. Players laugh.”
Right now I’m working on my first encounter, a tavern brawl against human bandits. I’ve had to nerf it a couple of times 'cause D&D Beyond kept telling me I was going to kill all my players.
Character creation is the best part anyway!
It uses the CR method of determining difficulty, which is… flawed.
How so? Genuinely curious. One of my worries is properly balancing encounters.
If you’re allowing feats or have been at all generous with Magic Items or stats, their CR method is probably underestimating the power of your PCs.
I’m a big fan of what is now termed a “funnel”; each player takes 3-4 minimally sketched out characters through a lethal scenario, and they pick from whatever survives. It gives some instant backstory and makes it clear that character survival is not a given. And while this seems to have been associated with the OSR movement and Dungeon Crawl Classics, it was the standard way of starting a Gamma World campaign (albeit usually with just two characters per player; with one becoming a sacrificial ‘meat shield’ and the character they would end up playing).
I say to give your players an unwinnable brawl and see how they think (or luck) their way out of the situation.
Stranger
We did a funnel in the DCC game I played. The DM had such an aversion to killing characters that almost no one died.
I had a fun time with the game, but I don’t like DCC as a system.
When a CR calculator tells you that an encounter is “deadly”, what that means is that there’s a possibility, if absolutely everything goes wrong for the players, that one character might die. You need to go to, like, two or three times deadly to seriously challenge a group.
Encounter balancing is really more of an art and is going to vary widely from table to table, group to group. If you have all novice players along for the ride, things will be different than with experienced players or even if a couple new players start looking up “Ultimate God Wizard Combat Destroyer!!!” builds online. A party with front-loaded classes like a Moon Druid with bear form will handle combat differently than others. How tactics/group oriented they are, what equipment they have, all sorts of things will make a difference.
As DM, you have the luxury of adjusting hit points on the fly, deciding on enemy tactics (attack the 18AC paladin or the 12AC sorcerer?), making the enemy flee when they take 50% losses or meet some other objective (“The bandits decide that grabbing the unattended sacks is good enough and run”), having allies join midfight if they’re getting slaughtered and so forth.
This doesn’t mean you should never let your players roll over an encounter or take losses. Sometimes, if you under-balanced an encounter and the players are demolishing an ambush, let them have and enjoy the win. And you shouldn’t be doing things like deciding “This guy is supposed to be a cool boss fight so I’m going to ignore the Hold Person spell he just rolled a 3 to save against and say he saved anyway”. But you definitely do have tools at your disposal to use with discretion as you get a handle on what makes an adequate challenge for your actual party at the table versus what the CR guide or D&D Beyond thinks.
My suggestion: waves of attackers, with variable strategic acumen.
Start the combat with maybe a third of the bandits that you planned on, with the rest of them out back, or in a room upstairs, or just drunk off their asses. Round two, a few more bandits roll into the brawl.
As the combat progresses, you can have more enter if the PCs need more challenge, or hold them back if things are going poorly. if the PC are winning, the bandits flank and target casters and otherwise act intelligently. If the PC are losing, the bandits are drunken idiots.
By round four or so, some major change should happen. Maybe the bar has caught on fire. Maybe the bandits have taken a captive. Maybe they hear the guard coming, and the guard is likely to arrest everyone equally. Maybe the combat moves outside.
Mix it up, so it is different from any low level combat they have been in before!
Or put the party in the middle of two adversarial groups, and make them pick which side to support, or take shelter and see how things pan out.
Stranger
Yes- especially allowing better than the standard array or point buy. In other words, until a PC gets to lvl4 , when they get a bump, no stat should be higher than 17- including racial bonuses. If you have a 20 before level 4, the DM has been very generous.
Yep. I do several of them.
OTOH- I roll out in the open.