UTDX Co-op Team Composition Guide (Tested in Endless Mode)

How I tested: 12 attempts, roughly 6 hours

The twelve attempts were not planned as a structured experiment. Attempts 1 and 2 were just my squad trying Endless Mode because we had the units for it. When we failed at Wave 62 both times, I started writing down what went wrong. By attempt 5, I was taking notes on every wave transition, every reroll window, and exactly where the team DPS dropped below what the boss HP required.

I used Void Court for all 12 attempts because it is the most predictable Endless map: the loop path is consistent, the boss contact points are stable, and placement errors repeat reliably so you can identify them. I ran Storm Harbor once as a secondary test after the Wave 100 clear; it introduced split-lane stress at Wave 40 that our comp handled without changes, which told me the placement rules were transferable.

The four of us are regular UTDX players but not no-life min-maxers. Two of us have full Mythic carry builds; the other two were running a mix of Legendary and Epic units. Total play time across the twelve attempts was approximately six hours across two evenings, including lobby time and the arguments between attempts 6 and 7 about whether Water God's slow was actually working (it was: the placement was wrong).

The pattern that emerged was that the progression from Wave 62 to Wave 97 across eleven failed attempts was almost entirely about fixing support placement and reroll timing, not about unit selection. We had the right units from attempt 4 onward. The gap between attempt 4 and attempt 12 was eight attempts of operational errors: which I think is a useful thing to say plainly. The composition guide is more than a unit list. It is a list of things that can go wrong with the right units in the wrong positions.

The composition that cleared: four-player roles

Here is exactly what each player ran on attempt 12, the wave-100 clear. Unit selection is followed by the reasoning for the pairing, and the specific position note that changed between failed attempts and the final run.

The pairing logic is consistent across all four slots: one unit that deals or amplifies damage, one unit that extends how long enemies stay inside the firing zone. Water God and Ice Empress both extend exposure time: Water God through a stronger slow with wider radius, Ice Empress through a consistent chain that covers the mid-lane before Water God's slow applies. Love Goddess and Virtual Idol both provide team-wide buffs: Love Goddess through a damage aura that lifts every unit inside its radius, Virtual Idol through a tempo buff that shortens the collective attack interval.

What failed in attempts 1 through 11: and why

Attempts 1 and 2 were the easiest failure to diagnose: we brought four carry units and assumed raw DPS would handle everything. The logic was not wrong for solo play: in solo, you decide every slot and can fill control on your own schedule. In co-op, no one owns control if no one is assigned to own it. Four Mythic DPS units with zero slow coverage against the Universal Core's 3.4M HP is a race between your total DPS and the boss's path length. At Wave 62, the path length won.

Attempts 3 and 4 introduced Water God but solved the wrong problem. We knew Water God's slow extended boss contact time, so we added it. What we did not think through is that slow must apply at the point where carry units are actively firing, not at the first contact point in the lane. Placing Water God at position 1 slows the boss during the opener segment, where none of the Mythic carries have firing range yet. The slow expired before the boss reached the boss-killer firing zone. We added slow and got zero benefit from it for eight waves.

Attempts 5 through 7 were deliberately testing a lower-budget version of the comp to see whether the Wave 88 ceiling was unit quality or role structure. Player 3 and Player 4 switched from Mythic to Legendary builds: Crow Shinobi, Fallen Prince, Bambee, Bulmo. The comp pushed to Wave 88 before stalling. That result told me two things: the role structure was working (Wave 88 was a big jump from Wave 62), and the ceiling was genuine unit stat pressure rather than structural error. F2P players contribute in the control and wave-clear roles, not the boss-burst role.

Attempts 8 and 9 are the failure I am most irritated about because we had the right comp and the right positions. Both runs failed near Wave 90 not because a unit was in the wrong lane but because Players 3 and 4 both chose to reroll their units during the same wave window. Rerolling takes a unit offline temporarily: the unit is placed but not firing while the reroll animation and application resolves. Two units going offline simultaneously during a boss wave dropped the team's effective DPS below what the boss required for those five seconds. We lost because of scheduling.

Attempts 10 and 11 brought Water God one position later as planned, but I misjudged "mid-bend" on Void Court. On that map, there are two bends before the boss approach segment. Mid-bend on attempt 10 was the first bend, which was still one segment early. Attempt 11, I moved Water God to the second bend and the slow duration finally covered the First Emperor firing window for the full boss pass. We reached Wave 97 before a Wave 95 boss stack depleted the team's upgrade reserves. Close, but out.

How to assign roles in a four-player UTDX co-op run

The default failure mode in uncoordinated co-op is not bad units: it is duplicate roles. Two boss-killers and no control is a common setup when four players each independently pick their strongest unit. The conversation that needs to happen before the lobby loads is which player owns which role, not which player has the highest-tier unit.

The four roles that need to be covered in Endless Mode are: boss-killer burst (one player), long-range sustained DPS (one player), splash wave clear (one player), and support coverage: at minimum two distinct support units to cover the full lane rather than a single aura at a single point. Eight unit slots across four players, four roles split two units each. That structure is what produced the Wave 100 clear. For how these roles shift across every wave band, more than Wave 100, see the Endless Mode milestone ladder.

The conversation before attempt 12 lasted about five minutes and covered four topics: who owns the boss-killer slot (Player 1, because they had First Emperor), who covers the second carry slot (Player 2 on Alpha Devil), who covers splash and wave 50-80 cleanup (Player 3), and who covers the long-range back segment (Player 4 on King Sailor with the best range stat in the roster). The support assignments followed from the primary assignments: each player pairs a support with their primary based on which support category fits that lane segment.

Lane ownership is the second pre-session conversation. Void Court has four distinct lane segments: the opener straight, the first bend, the mid-loop boss approach, and the exit straight before the loop reconnects. Assign Player 1 to the boss approach, Player 2 to the exit straight, Player 3 to the first bend for splash coverage, and Player 4 to the opener for range extension. Supports go with their assigned player in the same segment.

What co-op changes compared to solo: and what it does not

The unit priority list for co-op is different from solo in two specific ways. First, support traits are worth more in co-op. When Water God has an aura-strength Legendary trait, that trait amplifies the slow applied to enemies passing through its radius. In solo, that amplification benefits your own two or three carry units. In four-player co-op on Void Court, the same slow covers four players' units across two lane segments simultaneously. The effective value multiplier on Water God's support traits is roughly 2x in co-op compared to solo: which means the reroll timing calculator should weight support trait upgrades more heavily for your dedicated support player.

Second, economy coordination matters in co-op in a way it does not in solo. Each player controls their own gold pool, but upgrade timing is not independent: if Player 1 upgrades their carry at Wave 25 and Player 2 upgrades their carry at Wave 40, the team's DPS ramp is smooth. If all four players upgrade simultaneously at Wave 30, the boss at Wave 32 arrives to a team that just spent its gold and has no emergency reserve. Talk about upgrade timing the same way you talk about reroll timing.

What co-op does not change is the fundamental unit role math. A carry unit that stalls at Wave 62 in solo does not suddenly clear Wave 100 in co-op just because there are three other players. The co-op advantage is role coverage and aura multiplication: it does not change the unit's base DPS ceiling or the boss HP requirements. Know your unit's actual contribution before joining a high-wave Endless lobby.

Alternatives and adjustments for what you have

The Wave 100 clear comp is not the only path. It is the specific path that worked for my four-player squad on attempt 12 with the units we had. If your group does not have First Emperor or Alpha Devil, the role each unit fills is more important than the specific name.

Boss-killer slot alternatives in priority order: Nutaru (Beast) fills the same burst window as First Emperor and handles add pressure better on Ant Island variation maps. Jinoo (Monarch) is a valid secondary option if scaling summoner adds DPS during the early boss approach. Avoid replacing both boss-killers with splash carries: the Universal Core boss does not have the add density that makes splash carries efficient against it specifically.

Control slot alternatives: Ice Empress is a direct Water God substitute with shorter slow duration but lower cost, which matters for F2P players who cannot afford the full Water God upgrade schedule. If both Water God and Ice Empress are unavailable, Psycho (100%) provides crowd-control rather than slow, which creates a different coverage pattern: effective but requires repositioning the carry units to overlap the control window rather than the slow window.

Splash alternatives for Player 3: Migumen (Finale) covers the same wave-clear role with higher burst and lower sustained output. Underworld God (Death) is a viable substitute if Migumen is unavailable, with the same wind shear mechanic applying crowd pressure rather than true splash. Do not substitute the splash slot with a second boss-killer: two boss-killers and no splash coverage loses waves 40 through 80 to unchecked lane pressure from non-boss enemies.

Pre-session checklist for all four players

The five-minute coordination call I described produced the Wave 100 clear. Here is the exact checklist I now give to anyone asking about running Endless Mode with me:

Before the lobby opens: Redeem all active UTDX codes: GiftFromMiniking!, 40kInterestedWano!, 20KInterestedWano!. These give up to 550 trait rerolls free. Do this before the session, not mid-run. Confirm your unit lineup is fully upgraded to the level you can afford: do not enter expecting to upgrade inside Endless; the gold pressure is real.

In the pre-lobby call (5 minutes): Name the boss-killer player. Name the support ownership for each support slot. Agree on lane segments: who covers opener, who covers boss approach, who covers exit. Set the reroll stagger schedule: Wave 20, 30, 45, 60 is a reasonable default. Say the hard rule aloud: no two players reroll on the same wave, and nobody rerolls during a boss wave.

During the run: Check Water God placement on the second bend of Void Court, not the first. If it feels early, it probably is: move it one segment later. Do not upgrade and reroll in the same wave window on the same player account; the gold cost and offline window together compound. Check in verbally at Wave 50 and Wave 80 about reserve state: if one player has run out of upgrade gold, redistribute the run's strategy toward their carry rather than expecting them to hold the same DPS output.