Before I start going room-by-room there's a few general things to take care of. First if you recall my Improved Map I'm trying to re-arrange the dungeon to make it more linear and less sprawling. This means that the Region following Region B is actually going to be Region F. Since I'm still on Region B there's no real changes that need to happen just yet. However to keep in line with this plan there's a few things to keep in mind.
First, ignore the "key" plotline in Region B. In fact, even if you're keeping the layout of Region B as-is you should ignore it. It's dumb. First and foremost it brings up the question of why there are these amazing, unpickable, permanently trapped locked doors here...but not anywhere else in the dungeon..where they might actually make a difference.
First, it makes no sense for Region F to be locked away at all...it was originally a prison for minor demons and it was not a major bastion of celestial power. The text in Region B implies the celestials were actually trying to seal up Region B...but then why did they leave the key on the wrong side of the locked doors? Also...what's the point of only sealing the passages between B and F while leaving all other passages open? It's entirely possible to stumble on Region F (by going north from Region A into Region E then east to F) before reaching B.
It also doesn't make much sense from a balance perspective. The suggested path seems to go A-B-C-F....but C is for levels 7-9 and so is F. And you can't get the key from C without going through almost the whole region anyway. So by the time you get the key and retrace your steps to open up Region F you've already hit level 8 or 9...so F won't be much of a challenge. In fact, it's more likely that you'll never run into the key at all (it's very well hidden) and will instead probably hit Region G then maybe F (at which point you'll be at least 10th level...and F will presumably be a cakewalk).
So my suggestion is simply ignore this "key" business and the unpickable, inexplicable doors. If you're going with my new dungeon layout then make sure that all passages north (in both A and B) are dead ends. You can have them sealed with rubble...flooded with slime...whatever. My personal suggestion is to relocate rooms that are interesting/useful that the PCs missed the first time around. Say you've got a group exploring region B and they head north. Rather than a passageway to another region they find the smithy from Region A with the incomplete magic sword...or a trap or chamber you found particularly interesting.
One last thing...the region includes a sidebar called Goblin Tactics...or 21 ways to die. Don't Listen To It. Here's the first reason, a quote from the description:
"If played out properly a single goblin encounter can take up half of the evening. By having wave after wave of reinforcements pushing into the room the PCs are likely to tire quickly. Or at least, have their patience tested..."
How does this sound like a good or enjoyable way to spend your time? This is the proper way to run an encounter? Really? This is terrible DMing advice and they proceed to give even worse.
First, they recommend you cheat, which pisses me off to no end. Now I'm not saying I've never fudged rolls and I certainly am not one to look down on GMs who do it. What I do find contemptible is for adventure designers to suggest it. This was a 100 dollar book when it came out and the purpose of spending 100 dollars on an adventure is to buy a set of plots and encounters that have already been designed by professionals. And here those professionals are telling you that they did not do a good enough job designing this section to properly challenge PCs without fudging dice rolls.
Second, they suggest adding toughness to all goblins...basically also cheating but harder to catch. Still it just shows that they did not do the job on their own.
Third...they suggest adding +2 or +4 natural armor? What? Wasn't this section supposed to be tactics? you could have filled this with interesting tricks, traps and strategies that a large group of weak enemies could use to defeat a stronger foe. So far we've gotten cheat, cheat and cheat again without any creative thought. These are the first suggestions?
Four and Five (aid another and grappling) are actual tactics and they aren't bad suggestions...but they're also incredibly obvious ones that any decent GM should already know about.
Six isn't really a suggestion...it basically says "there should be traps and ambushes and stuff"...wasn't the point of you making this book that you design the traps and ambushes and stuff?
Seven...again this isn't really a suggestion. Limit your PCs options? What does that mean...the only example they give is "spells like fireball should be limited at this point in the campaign so fighting in tight ranks won't really hurt goblins." Is this saying that since PCs don't have fireball goblins don't have to worry about it? How is that a suggestion...it's like saying "don't worry...your PCs can't walk through walls so you don't have to worry about them going through a wall." Or is it actually suggesting that you cripple your PCs abilities to make things harder...again if things are too easy it's a failure on the part of the designers.
Eight is crap...if every encounter needs a hobgoblin why didn't you do it! Again, this is not tactics! It's just adding more enemies.
Nine...this isn't a terrible suggestion at least.
Ten....I want to hurt someone. I quote "Add encounter conditions we didn't think of".
Eleven. Improvise....great. Just great. That's wonderful advice.
Twelve. Basically give the goblins lots of magical equipment and poison. There's certainly worse suggestions (especially considering how low the dungeon is on magic) but it's still not tactical thinking.
Thirteen seems to suggest giving every goblin you encounter Divine Strength and Shield of Faith...that's pretty insane. A more mild suggestion is adding a 1st level cleric to every encounter. Again this isn't tactics it's just adding more enemies.
Fourteen. Be smart. Wow, what genius advice. The only example given is not to let PCs lure goblins into traps. Wow.
Fifteen. This one suggestions that you control the pacing and lock and bar some rooms...These are all exactly the things a game designer is meant to do when they write an adventure. Meaning the thing that we paid 100 bucks not to have to do.
Sixteen. Give the goblins spell resistance 12. I.e. more cheating. except stupider.
Seventeen: This one suggestions goblins should lure PCs into confined spaces where their large weapons are useless. First there are no rules that prevent the uses of large weapons in confined spaces. In fact, this is a "tactic" that works against the goblins. Confined spaces means that they can't use their superior numbers, flanking or teamwork while the PCs can take them on one or two at a time.
Eighteen. Add critters. Again, just adding more monsters does not qualify a "tactic".
Nineteen. Once PCs are tired and out of magic have the most powerful enemies show up. That's horrible advice.
Twenty. They suggest you get the rules for certain weapons from another book they make and basically suggest that you add extra damage to the goblin's weapons because the PCs won't get ahold of them. Considering that we already spent 100 dollars on this book you couldn't bother to print those rules in here? Really?
Twenty-one. Make the PCs work for it. I want to slap whoever wrote this one. It's not advice at all.
So, out of 21 pieces of "tactical" advice 5 basically tell you to cheat, 3 say to add more monsters. Most are nonsensical and only a handful could be called "tactics". So in the interest of improving the dungeon experience here are my 4 "Goblin Tactics" suggestions.
1) Strength In Numbers: The goblins will outnumber the PCs by a significant order of magnitude. They should always try and fight in large, open areas where they can easily split up and regroup to move around defending PCs (and more easily reach ranged-fighters and spellcasters). They should always have ranged weapons (javelins or slings) so goblins who are unable to close in combat can fire on PCs (especially those who keep to the rear). Aid another is important when fighting heavily armored PCs but it's usually better to use a flurry of attacks against a lightly armored PC. Flanking is always essential. Grappling is an extremely effective tactic when used against spellcasters, archers and similar non-mighty PCs. Don't bother trying it against the raging half-orc barbarian (average 3rd level raging 1/2 orc probably has +9 to grapple. The average goblin has a -2. That means it takes 6 goblins, with aid another, to give an even chance of grappling and the "leader" of the grapple is probably going to get killed by an AoO anyway).
2) Control Engagement Range: This depends a lot on adapting to PC tactics. When facing a group of melee combatants goblins should try and keep their distance. Reach weapons (Glaives probably work best) are helpful with ranged attackers behind the frontlines. A short retreat covered by caltrops or using tanglefoot bags also works great. If the PCs have strong ranged support (spellcasters or mundane) then the goblins should split up into multiple groups, forcing melee defenders to choose which group to engage and allowing more goblins to get within range of ranged fighters. It also prevents being devastated by AoE attacks.
3) Mundane Tricks and Equipment are Useful: At 3rd-5th level the PCs can still be significantly disadvantaged or injured by mundane equipment. Caltrops are excellent tools for ambushes or controlling PC movement. Nets make excellent control weapons (even if the goblins are non-proficient they only require touch attacks) and can easily disable extremely powerful enemies (the average 1/2 orc barbarian above can only break a net on a roll of 19 or better). In fact be careful with nets because they can be devastating. Alchemical items like smokesticks, acid and alchemists fire are effective ways of dealing with opponents in heavy armor. Keep in mind that goblins have darkvision with most PCs don't. Have goblins with slings and crossbows attack PC light sources, or even using things like tanglefoot bags to cover magical light sources in sticky gunk. Heck, if the goblins know they're facing humans or other surface-worlders then they might prepare flasks of paint to try and cover magical light-sources. If alchemical items are too expensive to justify for a bunch of goblins a simple trap of flaming oil is a good substitute. burning rags and oil is a good way to create barriers of fire and fill rooms with smoke. Weak, cheap poison (centipede venom for instance) is another good way to increase the challenge of an encounter but it can be a pain to keep track of round-by-round.
4) Ignore the Goblin Stats: While thinking up these tactics I looked over the stats for the goblins and they're are truly terrible. Perhaps the writers wouldn't have felt the need to suggest that you cheat to make encounters challenging if they actually made good choices when statting out the goblins. The goblin stats are full of mistakes and poor choices.
First problem: the weapon stats are wrong. The goblins are equipped with shortspears but they're listed as d6 damage rather than d4 (the proper damage for small size). Second problem, they're using shortspears. These are fighters and they can easily use martial weapons. A longsword or battleaxe is certainly a much better option than a shortspear. Bizarrely they also provide the goblins with bucklers...but give them a completely random, unjustified +1 bonus to AC...why not just give them a large shield to begin with? The goblin's armor isn't specified but it seems to be leather. It should certainly be at least Studded Leather, bringing their AC to 17.
Their HD are wrong as well (they have 1d8 hd when they should have d10). I'd suggest just increasing all hit points by 1 to even out the difference.
Their feats are also a mess. The goblins have Power Attack and Improved Sunder (it claims that "special training" allows them to take these feats with a Strength of only 12) and it says that goblin cavalry have Mounted Combat in place of Alertness (which they don't have). The thing is that Power Attack and Improved Sunder are both terrible feat choices. Their hit bonus is low enough that they can't afford the penalty to hit and Improved Sunder is complete crap. They don't inflict enough damage with shortspears (or any other weapon for that matter) to effectively penetrate a weapon's hardness and damage it effectively. These feats are actively bad. Instead give them Weapon Focus and Exotic Weapon Proficiency (Bastard Sword). Mounted goblins have Weapon Focus (lance) and Mounted Combat.
Their stats should look like this:
Goblin Ftr 1; CR: 1; Small Humanoid; HD: 1d10+1 (6 hp); Init: +1; Spd: 30 ft; AC: 17 (+1 size, +2 shield, +1 dex, +3 armor); Attack: +4 melee (1d8+1, B. Sword) or +3 Ranged (1d4+1, throwing axe); SQ: Darkvision 60 ft; AL: NE; Svs: Fort +3, Ref +1, Will +0; Str-12, Dex-13, Con-12, Int-10, Wis-10, Cha-6; Skills: Hide +3, Listen +1, Move Silently +3, Ride +3, Spot +1; Feats: Weapon Focus (B. Sword), Exotic Weapon Proficiency (B. Sword);
Non-elite goblins (i.e. those with 1 Warrior level rather than 1 Fighter level) simply use small-sized longswords (1d6 damage) and have Alertness.
Also make sure you ignore the tactics mentioned in the goblin encounters. The standard tactics seem to be forming into a 4 goblin-wide block formation (even in encounters where the number of goblins is not divisible by 4). The thing is this might be an effective military tactic for large armies in an open field and presumably the writers were trying to make the goblins more 'militaristic' but this is a terribly strategy for a relatively small group of weak individuals taking on a smaller group of stronger individuals. Surrounding or moving around the enemy is much more important so the goblins need to split up into smaller individual teams to move around their opponents, gang up on weak-looking enemies or force them into chokepoints.