Patch 16.17 · Gold+ ranked · updated 2026-10-28
What actually wins on Patch 16.17
Every other site ranks items by how often they're picked. We measure Win Probability Added — the causal win-rate impact of building an item, across 104 legendary items. Popular and effective are not the same thing.
Biggest riser
Hextech Gunblade
+2.67 pt
win-impact vs 16.13
Biggest faller
Nashor's Tooth
-2.60 pt
win-impact vs 16.13
Most overrated
Zhonya's Hourglass
-1.54%
121,272 builds
Most underrated
Protoplasm Harness
+0.95%
19,206 builds
Most overrated items
Built constantly, but among the lowest measured win-impact this patch. High pick rate is not the same as high win rate.

Zhonya's Hourglass
-1.54%
121,272 builds · 65,248 sample

Rapid Firecannon
-1.22%
29,217 builds · 8,836 sample

The Collector
-1.01%
85,967 builds · 24,096 sample

Shadowflame
-0.91%
136,690 builds · 56,647 sample

Hubris
-0.90%
46,851 builds · 22,472 sample

Luden's Echo
-0.79%
68,034 builds · 27,678 sample

Guardian Angel
-0.73%
31,249 builds · 33,890 sample

Rabadon's Deathcap
-0.68%
91,712 builds · 50,396 sample
Most underrated items
Strong measured win-impact, but built far less than they should be. The edges most players are leaving on the table.

Protoplasm Harness
+0.95%
19,206 builds · 17,657 sample

Ravenous Hydra
+0.71%
20,652 builds · 8,591 sample

Staff of Flowing Water
+0.39%
9,405 builds · 5,142 sample

Frozen Heart
+0.30%
22,000 builds · 21,055 sample

Statikk Shiv
+0.14%
25,745 builds · 8,913 sample

Endless Hunger
+0.10%
22,909 builds · 13,872 sample

Unending Despair
+0.10%
13,640 builds · 12,407 sample

Sunfire Aegis
+0.09%
15,846 builds · 10,230 sample
Biggest risers
Items whose causal win-impact climbed the most from patch 16.13 to 16.17.

Hextech Gunblade
-0.03%
+2.67 pt vs 16.13 (-2.70%→-0.03%)

Ravenous Hydra
+0.71%
+2.32 pt vs 16.13 (-1.61%→+0.71%)

Stridebreaker
-0.02%
+2.11 pt vs 16.13 (-2.12%→-0.02%)

Hubris
-0.90%
+1.52 pt vs 16.13 (-2.42%→-0.90%)

Umbral Glaive
+0.53%
+1.51 pt vs 16.13 (-0.98%→+0.53%)

Axiom Arc
-0.18%
+1.42 pt vs 16.13 (-1.60%→-0.18%)

The Collector
-1.01%
+1.27 pt vs 16.13 (-2.28%→-1.01%)

Endless Hunger
+0.10%
+1.00 pt vs 16.13 (-0.89%→+0.10%)
Biggest fallers
Items whose causal win-impact dropped the most from patch 16.13 to 16.17.

Nashor's Tooth
-0.33%
-2.60 pt vs 16.13 (+2.28%→-0.33%)

Statikk Shiv
+0.14%
-2.42 pt vs 16.13 (+2.56%→+0.14%)

Staff of Flowing Water
+0.39%
-2.18 pt vs 16.13 (+2.57%→+0.39%)

Echoes of Helia
-0.39%
-2.01 pt vs 16.13 (+1.62%→-0.39%)

Knight's Vow
+0.24%
-1.76 pt vs 16.13 (+2.00%→+0.24%)

Shurelya's Battlesong
-0.16%
-1.76 pt vs 16.13 (+1.60%→-0.16%)

Moonstone Renewer
-0.38%
-1.62 pt vs 16.13 (+1.24%→-0.38%)

Rabadon's Deathcap
-0.68%
-1.59 pt vs 16.13 (+0.91%→-0.68%)
Highest-impact items
The legendary items that add the most win probability on patch 16.17, regardless of popularity.

Protoplasm Harness
+0.95%
19,206 builds · 17,657 sample

Ravenous Hydra
+0.71%
20,652 builds · 8,591 sample

Trinity Force
+0.70%
86,310 builds · 17,905 sample

Death's Dance
+0.54%
88,471 builds · 35,383 sample

Umbral Glaive
+0.53%
37,224 builds · 12,842 sample

Voltaic Cyclosword
+0.50%
79,808 builds · 26,907 sample

Mejai's Soulstealer
+0.50%
34,354 builds · 46,563 sample

Staff of Flowing Water
+0.39%
9,405 builds · 5,142 sample

Frozen Heart
+0.30%
22,000 builds · 21,055 sample

Archangel's Staff
+0.29%
361 builds · 14,598 sample
How this is measured. Win Probability Added (WPA) is a causal estimate: a model trained on match timelines asks, for the game states where an item is bought, how much that purchase changes win probability versus the best alternative — not simply the win rate of games it appears in. Figures are averaged across every champion and role that builds the item (sample-weighted), from Gold+ ranked games on patch 16.17, minimum 3,000 sample per item. A negative number doesn't mean an item is "bad" — it means, on average, the games it's bought in would have been marginally better served by an alternative. Per-champion builds differ, so check your champion's page for the specific recommendation.