Mountaintop Seed Finder
The Mountaintop rises in the northwest quadrant during its Shifting Earth event. The dataset holds 50 records for it.
A quarter of Mountaintop identifications finish without marking a building, and 91.6% use two landmarks or fewer. Township is the strongest mark on this map by a distance: when someone marks one, it is load-bearing 96.2% of the time, which matches how rarely it appears here.
- Seeds in this map
- 50
- Landmarks to identify (median)
- 1
- Identifications logged
- 18,397
- Dataset game version
- 1.03.2
Township first, then the DLC buildings
Only 16 of the 50 records on this map contain a Township. That is the lowest rate of any map in the dataset, and it makes a Township sighting unusually valuable here, cutting the pool to under a third in a single mark. The inverse is weaker but still worth the look: confirming there is no Township leaves 34 candidates.
The 50 records split into 40 base game and 10 DLC, and Blacksmith Village and Marsh only exist in the DLC ones, where both appear in all ten. Between a Township check and a Blacksmith Village check, two marks will usually take you to a very small set. Great Church, Main Encampment, Fort and Ruins each appear in 45 of the 50 records, so none of them opens anything.
How often each landmark appears on Mountaintop
The share of this map's 50 records that contain at least one of each landmark. A landmark near 100% cannot narrow anything by presence alone, so the rows at the top of this table are the marks worth making first.
| Landmark | Records containing it | Share |
|---|---|---|
| Blacksmith Village | 10 of 50 | 20% |
| Marsh | 10 of 50 | 20% |
| Scale-Bearing Merchant | 15 of 50 | 30% |
| Township | 16 of 50 | 32% |
| Fort | 45 of 50 | 90% |
| Great Church | 45 of 50 | 90% |
| Main Encampment | 45 of 50 | 90% |
| Ruins | 45 of 50 | 90% |
| Church | 50 of 50 | 100% |
| Sorcerer's Rise | 50 of 50 | 100% |
What identification actually takes on Mountaintop
The table above is the dataset. This one is what players did with it. Across 18,397 identifications on this map, a search used 1.2 landmarks on average, selected the Nightlord 75.2% of the time and marked a landing point 58.3% of the time. 24.5% resolved without marking a single building.
The gap between the two columns is the useful part. The most marked landmark is rarely the one carrying the search, because a landmark that appears in almost every record cannot separate them by presence alone.
| Landmark | Times marked | Decided the search |
|---|---|---|
| Township | 468 | 96.2% |
| Blacksmith Village | 679 | 94.7% |
| Marsh | 1,155 | 93.3% |
| Ruins | 3,093 | 92.4% |
| Main Encampment | 2,955 | 88.2% |
| Great Church | 4,132 | 75.2% |
| Fort | 3,666 | 61.7% |
| Sorcerer's Rise | 1,397 | 55.7% |
| Church | 4,313 | 47% |
These come from the site’s own logs: every identification between 1 May 2026 and 30 July 2026, 400,702 of them across all maps. A landmark counts if it was still marked when the seed resolved, so a mark that was placed and then changed does not appear. Decisive means removing that one mark would have left more than one candidate standing, which is why several marks in the same search can each be decisive.
Running the Mountaintop
This is the most matchup-dependent area in the game, and it pays out along two separate lines. The Favor grants frostbite resistance; the bosses hand out the anti-dragon Dormant Powers. The smaller enemies here drop gear along the same two lines.
Caligo is where both halves land at once, being an ice dragon, and the difference is most obvious in its arena, which applies heavy frost. Adel is worth the detour for a narrower reason: it is a dragon but not an ice one, so the anti-dragon half applies and the frostbite resistance does not.
Outside those matchups the area gives you very little. The buff does not apply, and what is here is mostly small enemies rather than the major bosses that justify visiting other Shifting Earths. The one large boss is a dragon, and the fight takes long enough that it is usually not worth the day unless you are already committed to the anti-dragon package.
One rule holds across every Shifting Earth: the bosses are static. Same bosses, same places, same rewards, every run. Plain Limveld does not work like that, so this is the part of a map you can plan before you land.
The Favor and the bosses that grant Dormant Powers
Roughly half the frostbite damage and half the buildup, plus a short attack window when frostbite procs near you. That is the whole of it. The anti-dragon effect people attribute to the Favor comes from the bosses instead, handed out as Dormant Powers, and all of them grant it. Most published guides merge the two into a single buff, and the difference decides whether you can take the Favor and leave or have to clear the map.
The bosses that hand it out are cheap enough to be worth the trip on their own. Demi-Human Swordmaster, Flying Dragon and Snowfield Trolls sit at 18,000 runes each and none of them has to die for you to progress, with spiritstreams to help the climb and crystals holding strong items scattered around the slopes. The Mountaintop Ice Dragon is the exception at 104,500 runes: skippable, but it swoops on Nightfarers during the climb and spreads Frostbite while it does, and it is the only non-Nightlord boss on any map rated very hard. Four Evergaols survive the snowfields, level with the Crater and three behind plain Limveld.
Opening those four costs four Stonesword Keys, and this is the worst map in the dataset to be short of them. Only 16 of the 50 records hold a Township, the lowest rate anywhere, and Township is the only place in the game that sells a key, one to each player. The identification note further up the page values a Township mark for an unrelated reason, that it is rare enough here to cut the pool to under a third. It is the same number doing two different jobs.
Nightlord distribution
- Adel5
- Balancers5
- Caligo5
- Fulghor5
- Gladius5
- Gnoster5
- Greg5
- Heolstor5
- Libra5
- Maris5
Dataset counts describe the records available to this tool, not universal in-game odds. Read the data methodology and limitations before treating a small sample as a probability estimate.
