By using the data from Dungeon Clear Times it is possible to create heuristics for the overall check efficiency of different dungeons and dungeon routes. By combining this with the data from Dungeon Progression Biases we can also derive weighted heuristics that take into account the likelihood of finding progression in each dungeon.

Note that (currently) only "approximately representative" routes are considered for each dungeon -- as opposed to a more thorough analysis, which would evaluate the time taken across all (or many) different possible key and item layouts for that dungeon.

Where applicable, the item counts are adjusted for each dungeon route to account for this -- for example, routing through Skull Woods front to back yields 1.9 items on average instead of the full 2 when accounting for the chance of the big key being in the back and locking an item in the big chest.

The above chart shows a comparison of time efficiency for various dungeons and partial dungeon clears, in terms of the number of seconds spent per "progression unit" (lower is better), where a "unit" of progression is the equivalent progression weight of a single overworld check. See Dungeon Progression Biases for a full explanation of why progression is unequally weighted between different dungeons.

This second chart shows the same efficiency metric, but with the dungeon itself separated from the amount of time taken to travel to the dungeon. Some dungeons are commonly routed together with other checks in the nearby vicinity, so their travel time can effectively be amortized.

For some dungeons it is common to obtain all of the checks in that dungeons except for the dungeon's boss -- either because additional items are required to reach and defeat the boss, or because the boss item simply takes a disproportionate amount of time to obtain compared to the rest of the dungeon. This chart shows the relative efficiency difference when clearing those dungeons with or without the boss included as part of the clear.

As always, the data and metrics above are open to interpretation, and are not necessarily fully representative since not all key layouts and routing deviations are (yet) accounted for. Some suggested takeaways are as follows:

• Palace of Darkness is incredibly efficient when skipping Helmasaur, at only 38-51 seconds/unit including travel. This lends credence to the common adage of "always believe in pendant POD".
• Clearing the entirety of Skull Woods is still a generally mediocre value proposition, but the non-fire rod locked portion can be done in around ~41% of the time for ~66% of the items, making the front of Skull Woods among the most efficient of the routes considered. Combined with the large progression bias in SW, this lends credence to the (also) common adage of "never skip Skull Woods".
  • However, it is worth carefully considering whether the travel time to Skull Woods can be amortized (e.g. paired with a trip to VOO or TT), and whether additional time will be spent walking back out of Skull Woods, which can lower the net efficiency of the dungeon.
• Because Ice Palace takes 84-110 seconds longer to route without being able to do the Icebreaker glitch, its relative efficiency decreases from relatively efficient (50-59 seconds/unit) to quite unefficient (67-82 seconds/unit) without somaria on hand. Skipping Iced T is a valid (though risky) strategy, but only increases efficiency by a small amount.
• Despite generally being the slowest dungeon in the game (with or without travel time), Turtle Rock is very efficient thanks to its high progression density (the equivalent of 9.31 overworld checks).
• Despite its high progression bias, Misery Mire remains one of the most inefficient dungeons in the game. Skipping Vitreous brings it more in line with some of the other slow dungeons, but other options are generally preferable.
• A full clear of Desert Palace is surprisingly inefficient thanks to the boss clear taking 1.5-2.22 minutes for only 0.71 units of progression. Clearing only the front of Desert Palace is usually much more efficient.
• Despite taking under 2 minutes (including travel time) with good equipment on hand, clearing Tower of Hera via Herapot without doing the basement is not that efficient at 60-83 seconds/unit.
• Clearing only the front of Thieves Town only results in 1.56 items on average (equivalent to 2.07 progression units), but is still very efficient if otherwise in the area. However, this must be weighed against the cost of potentially having to re-enter the dungeon in a later trip.
• Hyrule Castle is always an efficient play, even considering the travel time. Though it is common to separate the dungeon into two trips -- taking the graveyard dropdown into the back of the sewers -- this doesn't actually save significant time over doing the dungeon in one pass as the metrics here are based on.