Community

I find a sweet spot of 60-70% worked best for me.

A normal monster hits a normally defended PCs 60% of the time.
An accurate monster hits a heavily armored PCS 60% of the time.
A inaccurate monster hits a soft PC 60% of the time

Basically the monster's attack bonus should be around the AC value of Medium armor- 12.

I personally think there shouldn't be a gap of more than 6 points between the lowest AC and highest AC party member. That way monster accuracy can be set to the mid point of +3 or +4 so that around 70% to 80% hits.

Its ridiculous that monsters can't hit anyone except the Wizard on a roll of 15.

@mellored

I happen to disagree with you there, at least for moderate values of luck factor.  Sure at the far extremes luck becomes important enough to trump tactics and render them meaningless, but I think that value is far far away from the 35-45% hit ratios I'd like to see.  

Actually, the luck factor is not the same between 90% hit chance and 90% miss chance (assuming damage and HP are modified to result in the same average durability).  Because the 90% hit chance will pretty much always kill you on the "expected" round, while the 90% miss chance could kill you pretty much any time.  That said, if you were right, than why would you complain about me wanting to swap 65% for 35%, since apparently, they're the same luck factor and by extension tactics factor?

@lawolf

I would rather play the game where you have a 40% chance of hitting and do 6.25 points of damage, personally.  Reductio ad absurdum only actually works as an argumentative tactic if your opponent recognizes no limiting principle at which to halt.  I fully recognize the dangers of swing, but I think the sweet spot at which the marginal cost and benefit equalize is a little bit higher than you do.  I don't want the corner case, and my argument does not require me to want it, so quit trying to defeat an argument nobody is making.

@ Doug

I agree with english language that it's a poorly designed tank that requires a bottleneck to function, and a poorly designed dungeon that results in frequent encounter-length bottlenecks.  That 5e tanks and playtest dungeons are poorly designed does not defeat my position, nor that a more careful reading of the playtest packet than I've bothered with can break the AC curve.  I'm not advocating for monsters that can only hit on a 20.  I'm advocating for monsters that hit on a 10-16 depending on how AC optimized the target is (ie +2 average, not far off from where it is now, at least once kobolds and goblins stop being idiots).  If I'm wrong about where the curve is, then by all means make them more accurate.  I'm not advocating for particular bonuses, I'm advocating for hit ratios.  While you're right that, given more accurate monsters, a +4 from total defense would not be as overpowered, it does not follow that one is the cause and the other is the effect.  Each affects the other simultaneously, and correcting either imbalance would solve the problem.  I would prefer to solve the problem by reducing the +4 than by making monsters substantially more accurate, not only because I prefer the systemic rhythm of lower values but because inevitably the ability will be used by the high end of the AC curve more than the low end and I would prefer to limit that range as much as possible.  I will grant that this is a preference.  I grant you yours, please grant me mine.