Notes

  • % of optimal (perc_optimal): each hit divided by the trimmed ceiling — the average of the top 5 hits against that boss tier in that season. More stable than a single best hit.
  • Z-index (z_index): the primary metric. perc_optimal is z-score normalised within each boss tier × season, then rescaled to 0–100 (mean = 50, 1 SD = 10 pts). A 60 on Mythic-1 is directly comparable to a 60 on Legendary-3 — different damage ceilings and spread are fully accounted for. Players who grind harder bosses are no longer penalised vs those who cherry-pick easier ones.
  • Composite z-score: token-weighted average z_index across all boss tiers. Token weighting rewards consistent participation.
  • Bad Data: This is made with partial season data. Finishing hits are included. Primes are excluded.


Commentary

  • I’d like to get top hit per boss/prime from GK and compare to that. Our guild’s average might be 40% on some bosses but this way we set a target to what is literally possible inside the game, rather than some relative measure within guild that changes over time as we get stronger/weaker.
  • Until then, normalized Z-score is the best way to compare. You can see that we might average 70% of highest hit into low Legendary bosses, but that should probably not be weighed more heavily than a 40% hit into mythic if the guild averages are so different between those bosses.
  • Lavistodes have the higest potential into mythic sza, but our admech consistency is much better and low-developed lavistodes performed much worse.
  • Compare final table “Aggregate Player Score” to something like average damage or total damage on Raidman and you can see important differences.
  • Neuro this season was important on primes. Not present in boss data.

Damage Distribution by Team

Ridge density plots showing the distribution of % of optimal scores by team composition. Each ridge is one meta team. Wider/taller peaks = more hits concentrated at that % of optimal value.

Legendary-1

Legendary-2

Legendary-3

Legendary-4

Legendary-5

Mythic-1

Player Performance

1. Heatmap: Player × Boss Tier

One row per player, one column per boss tier, filled by avg % of optimal. Instantly reveals specialization and weak spots. Players sorted by their self-weighted composite score.

2. Boss-Group Performance by Team

For each meta-team archetype, shows only players who used that team — ranked by their avg % of optimal on that team’s favored bosses. Compares apples to apples within each team cohort.
Season 96 Mapping Admech best in slot for Legendary 1-3, Classic Custodes for Legendary 4-5, Lavistodes for Mythic 1.
Arguably Lavistodes should be Leg4 onward, but 9/10 top hits were with classic custodes. Melegor is lone top hitter.

3. Within-Team Percentile Rank

Each player gets a percentile rank among all users of the same meta team, on that team’s favored bosses. Removes the problem of comparing Lavistodes damage ceilings to AdMech ceilings — you’re ranked only against your own cohort.

4. Consistency: Mean ± SD

Dot plot sorted by avg Z, with error bars showing ±1 standard deviation. A tight bar means a reliable player; a wide bar means high variance — great hits mixed with weak ones. Requires ≥3 hits to show SD reliably.
Remember, finishing hits are included because we don’t have a flag yet in the data.

5. Token Allocation vs Efficiency

Each point is one player × boss tier. X = tokens used, Y = avg % of optimal. Vertical line = median tokens; horizontal line = median efficiency. Quadrants: top-right = workhorse (many tokens, high efficiency); top-left = specialist (few tokens, high efficiency); bottom-right = volume player dragging score; bottom-left = low participation and low efficiency.
Can easily see players who threw tokens at Sza with low effect. Variance in 1 token column not very meaningful and some are finishing hits.

Boss Tier Reference

Best hit, who hit it, with what team, guild average, and spread per boss tier.

Player Z-Index by Boss

Z-index per player per boss tier. Sortable and filterable. Use the Boss Tier Reference above as context.

Aggregate Player Score

Token-weighted composite z-index across all boss tiers. See Notes section for methodology.

Generated 2026-03-17 · Per-hit data across all 6 boss tiers · Current roster only