Start with the possibility
For soloing, I ask whether each boss in every raid can be defeated by a character locked to that expansion's level on the highest difficulty.
WoW theorycrafting / tested raid strategies
My theorycrafting starts with one question: can this raid problem be solved in a way nobody has proved yet? I question the normal answer, strip the problem back and verify it again from A to Z.
I record and log the work so the answer can be reviewed instead of remembered.

No ego / always learning
A new player, a veteran or somebody solving a completely different problem can all see something I did not. I go into conversations ready to learn, because protecting an image of being right would only slow me down.
I often work alone because I think best with space, patience and no pressure to perform for a room. Working solo lets me be honest about bad ideas, but it never means I stop listening.
The method
The question changes between soloing and group play, but the work still moves from possibility to evidence.
For soloing, I ask whether each boss in every raid can be defeated by a character locked to that expansion's level on the highest difficulty.
When a normal group strategy fails for Legacy Thirty, I spend hours across days or weeks breaking the encounter down and looking for a route that works for us.
I record and log the work so I can review the data and footage instead of relying on how a pull felt.
When it is useful, I compare older strategies or approaches from WoW's Classic remakes to see whether they expose a different answer.
Example from group play / Halion 25 Heroic
Halion is one example of how I theorycraft for a group. At 50%, most raids split between the Material and Twilight realms and balance Corporeality near the middle. I devised the opposite plan for Legacy Thirty.
Keep both copies moving together and hold Corporeality near the safer middle.
We let Halion reach 400% more damage taken in the dark realm, accepting that he would also deal 200% more damage.
Once the multiplier was ready, we funnelled our defensives into the tank, used Bloodlust and burned Halion before the danger caught us. That plan secured the world-first level-30 kill on 25-player Heroic.
Example from solo play / Mythic Tindral
The flaming trees were the real wall. Instead of treating enrage as an automatic failure, I rebuilt the entire route around surviving it.

While I stayed on a flying mount he could not teleport me across, so I never had to enter the next platform's flaming trees. Staying behind removed the mechanic I could not outlast, and forced the boss into enrage.
Against a lower-level target, each block removes a larger percentage of the hit. That made surviving the enrage more manageable than dealing with the flaming trees.
The enrage locked Tindral into a permanent Supernova cast, so the rest was waiting the trees out, reaching him and out-healing damage that would otherwise kill me.
Example of changing the question / Mythic Fyrakk
Phase two asks for something one player cannot give: heal the Spirits of the Kaldorei and clear the line of adds, inside five seconds. I spent a year building plans to meet that demand.

Every plan I built tried to answer the window on its own terms: more healing, faster clears, better cooldowns.
A Hunter levelled to 90 did enough damage to burn Fyrakk down in the first thirty seconds of phase two, before he flies off. The five seconds never arrived.
I had been asking how to survive the window. The kill came from asking whether I needed to reach it at all.
Watch the full Fyrakk soloThe bigger idea
Theorycrafting is how I express creativity and intelligence in a game. I take the same board everyone else is looking at, drop the assumptions around it and find a path that did not exist before.