Best brawlers for Dungeon Train in Gem Grab

What the data says about Dungeon Train

Dungeon Train is one of the Gem Grab maps with the most recorded battles in our sample: 185 high-level games over the last 14 days, with 40 different brawlers seen on it. This page ranks those brawlers by their real win rate on this specific map, not by their average performance across the game, which is the figure that usually leads to bad picks.

The sample for Dungeon Train is still short: 185 recorded battles. The percentages below are already useful as a guide, but read them as a trend rather than a settled number; as the sampler keeps adding games, the order of the top spots will stabilise.

The best picks on this map

The best brawler on Dungeon Train is PAM, with a 78.2% win rate across 14 battles recorded on this map. It is the first option to consider when this rotation comes around, and also the pick most worth banning or countering if the enemy team gets there first.

What is interesting about PAM here is not the absolute number but the gap: across the rest of the Gem Grab maps it sits at 73.8%, so Dungeon Train adds 4.4 percentage points. That jump is the sign that the map specifically favours its kit, not that it is a strong brawler in general.

The second option is AMBER, with 77.2% across 12 battles, 1 points behind the leader. When the distance between the first and second pick is that short, the choice depends more on the enemy composition and on the rest of your team than on the map itself.

EMZ closes the podium at 73.7%. Having three clear options instead of just one makes Dungeon Train a comfortable map to draft on: if the first one gets banned, you are not left without a plan.

The most picked brawler on Dungeon Train is not necessarily the best: it is Cosmo, present in 15 battles, 8.1% of every game recorded on the map. The distance between what is most picked and what actually works is usually where the edge is, because most players repeat a pick out of habit rather than out of data.

Dungeon Train is a balanced map: there are only 6.2 percentage points between the best and the worst brawler in the ranking. That means almost any reasonable composition is playable and that the game is decided more by execution and positioning than by the selection phase.

Brawlers that perform worse on Dungeon Train

There is not enough sample yet to confidently point out which brawlers perform worst on Dungeon Train. We only publish the bottom tail when the ranking has enough depth for the best and the worst to be genuinely separate groups instead of overlapping.

How Gem Grab is played

In Gem Grab the objective outranks the kills: the win goes to whoever holds ten gems through the countdown, not to whoever eliminates the most. That rewards brawlers who can control the centre and protect the carrier, and punishes those who need to look for duels away from the mine. On Dungeon Train, read the ranking thinking about who holds the middle, not about who hits hardest.

Dungeon Train rotates in and out of the official Gem Grab rotation every few days. This page does not depend on it being active right now: the ranking is computed over recent history, so it remains the best reference available the day the map shows up again.

Every percentage comes from real high-level battles sampled continuously, aggregated over a rolling 14-day window and adjusted with Bayesian smoothing. That adjustment avoids the classic false positive: a brawler with three games and three wins does not show up at 100%, but close to the average, until it builds up the sample to back it.

If you want to compare Dungeon Train with the rest of the Gem Grab maps, the links are at the bottom of the page. And if what you are after is what to pick in today's rotation, the picks page filters the maps that are active right now.