So, I did a speedrun at 53 minutes. It would have been sub-50 except that I ran into the stasis trap before getting the onyx key… I get ahead of myself sometimes.
I was on pace for 48ish, but lost 4+ minutes to the stasis trap and then revisiting Kylearan’s tower. Interestingly, the Wizard beat Mangar without dying once…
Overall efficiency was better, but there are some small gains (90 seconds?) to be made. A lot of the actions in the sewers and Mangar levels were unpolished and unnecessary. It’s just a function of having not done any work on this speedrun in some time.
This specific run exploited the interesting gold/item duplication bug in the shop. This lets the adventurer spend about 90 seconds traveling to and from the shop to have essentially infinite gold. Gold economy is no longer a constraint which means even more time (and gold) can be used in leveling as rapidly as possible.
I am near the edge of the limits of this overall strategy (Magi 1–7 to Conjurer 1–13 to Wizard 1 to 30). If I can come under 46 minutes using this path, I’ll be impressed.
The next level of breakthroughs are a function of XPM. There are a few ways to get more XPM (25-70kpm+) that may be accessible through the shop duplication bug.
Currently, Ogres offer a flat baseline of about 17,000 experience points per minute. This is one of the best and most consistent sources for XP and represents a minimal base-line in terms of time spent and XP gained. Walking the Catacombs without APAR is slow as compared to Ogres in terms of XPM for all but the last 2 levels of Magi/Conj. APAR speeds the Catacombs up to 1 level per 75 seconds or so. And finally, SPSP is almost certainly the fastest possible XP in the game for a Wizard level 11 or higher that does not have a draining weapon. (The effort required to get a draining weapon is not materially different from the effort required to beat the game.)
So, XPM in the 25-70kpm+ range—which is more than Ogres but far less than SPSP—represents an opportunity for substantial time gains. Some of the scenarios are interesting but rather complex and difficult to execute. For example, take an army of hobbit bards, level 2, and arm each of them with a firehorn, and then hope for good first-strike RNG. Be sure the hobbit army dies, leaving only 1 entity to gain all the xp, repeat ad nauseum.
Ultimately, the ability to successfully flee, the ability to survive high level breath attacks, and the ability to successfully cast a spell at a high-level target, will trade off time spent leveling. A level 13 conjurer could in theory beat the game in just a few minutes. But the probability of the RNG allowing that to happen is very small (1 in 100,000,000 runs?).