Reroll strategies in Teamfight Tactics represent a constrained optimization problem built around fixed champion pools, level-dependent shop probabilities, and timing windows where low- and mid-cost units can realistically reach three-star thresholds. Within this system, rerolling is not defined by mechanical repetition of shop refreshes, but by interaction with a finite shared resource: the global champion pool. Pool depletion, caused by both direct competition and incidental holdings from unrelated compositions, directly alters expected discovery rates and therefore determines whether a reroll line remains structurally viable. This article frames reroll play as a system governed by depletion pressure, probability gradients, and synchronization between lobby demand and level breakpoints.
System architecture of reroll and shared champion pools
This section breaks down how reroll strategies are shaped by the game’s underlying systems—specifically the way champions are shared across a global pool and how shop odds shift with player level. The following subsections explain how pool depletion directly limits upgrade potential, and how level-based probability layers define the narrow windows in which reroll remains structurally viable.
Champion pool structure and depletion mechanics
The champion pool in TFT is globally shared and segmented by cost tiers. Each individual champion exists in a fixed quantity, and once copies are acquired by any player—on board or bench—they are removed from circulation. From a system perspective, reroll strategies are therefore capacity-limited: only a bounded number of players can complete a three-star outcome for the same champion before the pool collapses into scarcity. Depletion is non-linear. Early removals have limited impact, but once a threshold is crossed—particularly when a competing player completes a three-star unit—the remaining accessible copies drop below the mathematical requirement for additional upgrades.
This structural constraint transforms reroll from a purely probabilistic search into a competitive extraction process. Each purchase alters the global state. Unlike higher-cost carry strategies that rely on late-stage probability ramps, reroll systems depend on preserving sufficient pool depth long enough to complete nine total copies before competing demand drains availability.
Level-dependent probability layers
Reroll viability is also constrained by level-specific shop distributions. Each player’s level acts as a probability filter that controls which cost tiers dominate refresh outcomes. Low-cost reroll strategies operate inside narrow level windows where one-cost and two-cost appearance rates peak. Three-cost reroll strategies are similarly bound to intermediate levels where their probability mass is highest before higher-cost units dilute shop slots.
From a systems view, reroll strategies are therefore defined by overlapping windows between two curves: the remaining pool size of the target champion and the level-based appearance probability of its cost tier. When either curve collapses—through pool depletion or level progression—the strategy loses structural feasibility regardless of remaining gold or health.
How reroll systems interact with pool depletion during play
This section explains how reroll strategies dynamically interact with shared champion pools over the course of a game, focusing on how supply pressure, player behavior, and timing reshape completion odds. The following subsections break down how extraction speed accelerates depletion, how stabilization changes competitive pressure, and how players can identify when a reroll line is no longer realistically completable.
Extraction phase and depletion acceleration
During the extraction phase, reroll players repeatedly refresh shops to convert gold into copies of specific champions. Each successful acquisition marginally decreases the global supply and therefore increases the marginal cost of the next copy. This creates a positive feedback loop: the deeper a player commits to a reroll line, the more expensive completion becomes for all other players pursuing the same unit.
Depletion accelerates sharply when a competing player approaches completion. The moment one participant reaches eight or nine copies, the remaining pool for that champion often drops below a functional search threshold. At that point, additional refreshes exhibit rapidly diminishing returns. This behavior explains why contested reroll lines frequently display bifurcated outcomes: one player stabilizes around a completed three-star, while others stall at six or seven copies despite substantial refresh volume.
Stabilization versus completion pressure
Reroll systems are not defined solely by completion of three-star units. Stabilization refers to reaching sufficient board strength to halt health loss while continuing controlled extraction. From a system perspective, stabilization delays forced spending and allows more shop cycles to be processed while the pool remains partially intact.
However, stabilization interacts directly with depletion dynamics. If multiple players stabilize on the same reroll line, total extraction velocity increases across the lobby. The global pool drains faster than any single player’s local refresh rate can compensate. As a result, delaying completion to preserve economy can paradoxically reduce the probability of eventual completion when competition is symmetric.
This creates a structural trade-off: aggressive early extraction reduces shared availability but increases individual capture probability. Conservative extraction preserves resources but exposes the strategy to depletion by parallel actors.
Detection of infeasible completion states
Within reroll systems, infeasibility can be inferred before the pool is mathematically exhausted. When observed holdings across opponents approach or exceed six to eight copies of the same target champion, the remaining pool often falls below a practical discovery threshold within the current level window. Because shop probabilities are fixed per level, additional refreshes cannot compensate for insufficient remaining supply.
This creates a soft failure state: the strategy remains technically possible but becomes statistically implausible. At this stage, continued rerolling primarily converts gold into marginal board strength rather than into meaningful progression toward completion.
Factors inside the reroll and pool axis that reshape outcomes
At a deeper systems level, reroll success is not shaped only by direct contesting, but by several hidden dynamics inside the shared pool and timing structure. The following factors explain how availability, flexibility, and lobby behavior subtly reshape completion odds and variance across reroll lines.
Incidental holdings from non-reroll compositions
Pool depletion is not driven only by direct contesting. Champions that serve as temporary trait fillers, item carriers, or transitional front-line units are frequently held by players who do not intend to complete them. These incidental holdings remove copies from circulation during the most critical reroll windows. From a systems standpoint, this background consumption produces stochastic depletion that is difficult to predict but materially alters availability curves, especially for two-cost and three-cost units commonly used as mid-game stabilizers.
As a result, reroll strategies that rely on highly flexible champions experience higher variance in completion rates even in lobbies where no explicit competitor targets the same three-star outcome.
Multi-carry reroll architectures
Some reroll structures distribute completion pressure across multiple possible carries within the same trait or damage profile. This architecture reduces dependency on a single champion pool and allows partial completion states to remain functional. From a pool perspective, multi-carry reroll lines behave as load-balancing systems: extraction can be redirected toward whichever pool remains less depleted.
This does not eliminate depletion risk but changes its geometry. Instead of one pool collapsing catastrophically, multiple pools are drawn down gradually. The system therefore becomes more resilient to sudden three-star completions by competitors, though at the cost of increased internal coordination complexity.
Timing compression in contested lobbies
When several players converge on similar reroll cost tiers, extraction timelines compress. The shared pool is accessed earlier and more intensively, while all players remain inside the same probability window. This compression shortens the functional lifespan of the reroll window itself. In such environments, completion becomes determined less by long-term refresh volume and more by early acquisition density and first-mover advantage.
This dynamic explains why reroll outcomes often appear patch-sensitive even when pool sizes remain unchanged: small shifts in early-stage viability of specific units alter how many players enter the same extraction window simultaneously.
Summary
Reroll strategies in TFT operate as competitive extraction systems constrained by fixed champion pool capacities and level-dependent probability distributions. Champion pool depletion is not a peripheral concern but the primary structural limiter of reroll viability. Each purchase modifies the global state, progressively raising the marginal cost of additional copies and accelerating failure states for parallel strategies. Stabilization delays do not exist independently of depletion dynamics; they directly interact with the extraction velocity of the lobby. Incidental holdings, multi-carry architectures, and timing compression further reshape how depletion unfolds across games. Viewed through this systems lens, reroll success is determined less by isolated shop outcomes and more by how effectively a player’s extraction timeline aligns with the remaining depth of the shared champion pools.