I agree that Warcraft Logs cannot guarantee 100% accuracy when trying to identify specs - as specs aren’t known during combat, Warcraft Logs needs to somehow determine specs from events.
The trouble with using Master of Elements as the primary indicator of spec - Fire versus Frost - is that Master of Elements only procs if a critical strike occurs.
My mage is well-geared - but without World Buffs - I have only 13% critical strike chance for frost spells.
Numerous encounters are fairly short - and let’s say, for BWL, that I’m forced to cast frostbolt - due to boss immunity to fire.
Looking at our BWL logs, this means that I can only cast, say 18 frostbolts, for short encounters.
The math here is:
(0.87**18)*100 = 8%
Which is to say, using MoE to gauge whether or not a mage is fire-spec can be off by 8% in cases where a fire spec mage has lost world buffs but is specced into MoE in a short encounter where frost spells are cast.
Moreover, as a previous poster pointed out, some fire-spec mages will not take MoE - opting for POM-Pyro. The MoE approach to mage spec identification will miss all such mages - probably the POM-Pyro spec is somewhat popular for mages wishing to spec fire for PvP but don’t want to spend gold to respec for PvE.
Bottom line - focusing on MoE events - which depend on critical strikes - will miscategorize roughly 8% of mages in short encounters where world buffs are lost and the mage doesn’t apply wizard oil; also, POM-Pyro spec mages will be miscategorized regardless of crits.
Such mages won’t be in competition for top parses (in most cases) - so perhaps this is acceptable.
On the other hand, mage spec detection can be dramatically and easily improved:
- Consider MoE procs over the entire raid, instead of a single encounter;
- Consider certain spell casts as indicative of fire-spec since some fire specs won’t take MoE; e.g., Pyroblast and Blastwave, if cast, indicate a fire mage spec