problem is it fails to state a lot of deadly encounters are deadly.I agree that the actual value for your example should be ~1.5 instead of 2. My take, however, is that Shea's rule 2. is doing its job if it successfully predicts that an encounter will be deadly (given we are also complying with his rule 1). So if we put a demi-lich against four 9th level PCs because we think it makes sense in our campaign, his rule 2. successfully predicts that the encounter will be deadly.
The problem is that XP percents and CR percents are not the same. Deadly is 2x medium in XP, but it is more like x1.5 in CR.
I made the same error many times, and I think Mike has as well.
See, 2 CR 7s are already deadly, but his system only triggers at 3 of them. 2 CR 6s are close to deadly, but in his system they are miles away.A more likely encounter - also deadly - would be three or four CR 6 creatures. Which is also about right for XP award, although both the single and the multiple monster encounters will fall within the XP/day expectation.
No, adding CR type balancing doesn't need the encounter size multiplier.His rule ignores the encounter multipliers table and I think points out a kind of flaw with it. It should have made 1x the value at 3-6 monsters - the standard encounter - and offset from there. I think the values it uses are somehow connected with the inherently "easy" setting of 5e.
The encounter size multiplier is needed because XP isn't linear in CR, but closer to quadratic. Challenge is pretty close to linear in CR, and adding CR is a really easy way to get within a step or two of the real challenge.
This isn't the exact math, but imagine XP = CR^2 * 100.
Take two monsters CR 5. They are worth 2500 XP each, and add up to 5000 XP. We then encounter size multiplier of x2 for 10,000 XP.
If we naively add up CR we get 10. The XP of a CR 10 monster is 10,000. We didn't need thr encounter size multiplier.
Now the CR <-> XP relation isn't XP=CR^2 * 100, but more like CR^1.5 * 200 or something (I have it calculated somewhere, but am on my phone - 1.3? 1.5? 1.7? I forgot). And CR isn't quite linear -- CR 1 to CR 5 is about 4x power, not 5x, and CR 1/8 to CR 1 is 6x not 8x -- but it is close.