Spell design does follow the 2014 DMG table rather closely. The notable exceptions are that 1st level spells tend to do slightly more than the table suggests, so as are not outclassed by cantrips; and that at every spell level around two spells do more than the table because WOTC felt like it. Call the second the Fireball problem.
Comparing effects to spell damage isn't how they are valued (this is a second-level spell!!), the status effects are valued as flat virtual damage as if a single-target attack. An attack that does some damage and applies a status is valued as a single attack that does more psuedo damage and no status.
So from a DPR perspective; the medusa does 23 to 26 damage with a +5 to-hit. This puts its actual damage under the average for the level, which is 35 damage with a +7 to-hit. This is, outside the stone gaze, average stats of a CR 3 or 4 monster
Meanwhile, the AC and HP values put it more in line with a CR 7 or 8 creature than the listed CR 6. This is ignoring the presumed auto-disadvantage attack rolls against her have.
So the medusa sure looks like the CR is probably correct based on how one values the stone gaze. CR 3 and CR 8 averaged out equals CR 5.5, plus the stone gaze tipping it over the edge to CR 6. This is probably WOTC's reasoning here
If you value the stone gaze as effective AC (which you probably can comfortably do) and value the pseudo-damage as a higher value than WOTC does, by those criteria the Medusa could comfortably be bumped in CR a few levels. Probably CR 8 or so.
Hope this helps
That did point me towards the answer! (Thanks!) In this case, I don't think it's taking into account the petrifying effect
at all.
The Medusa's hp and AC give it a defensive CR of 4. There is no debate there, thats how it works. The attack routine also gives it an offensive CR of 4. So, if we ignore the special ability, it's just a CR 4 creature.
So, what if, instead of messing with the petrification, they are assuming all the PCs are averting their gaze to avoid getting petrified? In that case, the PCs are effectively blinded against the medusa the entire fight. That means they have disadvantage on their attacks and it has advantage on attacks against them. And in monster design, that converts to a +4 effective AC and a +4 effective bonus to attack. Since each +2 raises the CR of its respective part (offense or defense) by 1, that means that both offensive and defensive CRs (and thus CR in general) increase by 2. Which bring that CR 4 to the listed 6. Either that or they are just treating it as 21 damage for action denial and didn't count it twice for multiple targets (One would think they might have caught their error for the 2024 version if the latter were the case though.)
As far as your direct points, I think you are probably right that they are going with single target damage from a spell converted to the condition. If that is what they are doing though, it means they are not valuing condition causing spells as much as damage causing spells, because otherwise they would be doing it the way I described before (which is how they do actual damage and damage spells). That in and of itself is a noteworthy concept.
I'm also leaning more strongly into the idea that they really are only treating the strongest condition as 2nd level action denial. I mean, what conditions (other than dead) actually have a stronger impact on the battle than taking a character out of the fight for a round? The only one I can think of is getting PCs to attack their allies. That would likely count as a 4th-level effect, and it's anyone's guess what virtual damage they would assign to it.
And I'm wondering now about the cantrip and 1st level values of 5 and 11 offered in the OP. If we are taking magic missile and rounding it up from 10.5 to 11, why are we going with 5 for a cantrip? I would assume fire bolt rounded up from 5.5 to 6. Or maybe those values are derived from the DMG tables. But, as Jeremy said, thats not actually where the condition-equivalent numbers are coming from. Those come from the strong iconic damage spells. Otherwise 2nd level would be 16.5 from the table, rather than the 21 from scorching ray that it is.
And it seems from that that chromatic orb might be a better one to pick for first level, giving us a virtual damage value of 13.5, rather than 11. Anyone's guess which way to round. In general, design tends to round off, as opposed to the in-play rule of rounding down.
Probably late to the game here, or I would have been asking the OP about his numbers when this thread started.