Item Slam Patch 7.41f
Sange and Yasha

Sange and Yasha: best heroes by WPA

Sange and Yasha in patch 7.41f: WPA of the purchase on 17 heroes, by role and build slot, against the other items bought in the same slot. Best confirmed result: Legion Commander core, 3rd item (+7.52 pp).

By hero and slot

WPA of the item against the other items bought in the same slot, in matches of every rank. Slots with fewer than 30 purchases are left out.

Hero Role Slot Purchases WPA 95% CI Verdict
Legion CommanderCore3rd item47 (0.6%)+7.52+3.10 to +11.89✓ Better
SlarkCore1st item62 (0.4%)+7.41+2.43 to +11.63Inconclusive
Night StalkerCore2nd item64 (0.6%)+6.59+1.38 to +11.26Inconclusive
Legion CommanderCore4th+ item85 (0.9%)+4.59+1.18 to +9.01✓ Better
Night StalkerCore3rd item90 (1.1%)+4.14+0.21 to +8.28Inconclusive
Drow RangerCore1st item236 (0.6%)+4.10+0.13 to +7.83Inconclusive
SlarkCore3rd item412 (2.8%)+3.72+1.62 to +6.31✓ Better
Night StalkerCore4th+ item133 (1.2%)+3.68+0.33 to +6.50Inconclusive
Dragon KnightCore2nd item45 (0.2%)+3.58−1.24 to +7.78Inconclusive
SvenCore1st item59 (0.4%)+3.57−2.06 to +10.72Inconclusive
SvenCore3rd item163 (1.1%)+3.54+0.28 to +6.81Inconclusive
AxeCore4th+ item34 (0.2%)+3.04+0.26 to +7.29Inconclusive
Drow RangerCore3rd item191 (0.7%)+2.91+0.10 to +5.70Inconclusive
Ogre MagiCore4th+ item40 (1.5%)+2.66−2.53 to +7.66Inconclusive
SvenCore4th+ item182 (0.4%)+2.59−0.03 to +5.42Inconclusive
SvenCore2nd item177 (1.1%)+2.54−2.61 to +7.37Inconclusive
Anti-MageCore2nd item61 (0.4%)+2.44−3.26 to +8.01Inconclusive
LifestealerCore2nd item2,777 (20.0%)+2.17+0.93 to +3.44✓ Better
Dragon KnightCore3rd item71 (0.4%)+2.01−1.83 to +6.10Inconclusive
SlarkCore2nd item636 (3.8%)+1.99−0.60 to +5.00Inconclusive
PudgeCore4th+ item61 (0.4%)+1.98−1.83 to +6.46Inconclusive
Spirit BreakerSupport2nd item41 (0.3%)+1.82−2.88 to +7.50Inconclusive
Drow RangerCore4th+ item171 (0.4%)+1.60−0.78 to +4.13Inconclusive
LifestealerCore3rd item2,905 (24.2%)+0.95−0.01 to +2.02Inconclusive
Drow RangerCore2nd item271 (0.8%)+0.78−1.70 to +3.23Inconclusive
SlarkCore4th+ item265 (1.0%)+0.71−1.54 to +2.62Inconclusive
DawnbreakerCore3rd item157 (1.4%)+0.65−2.32 to +4.11Inconclusive
Spirit BreakerSupport3rd item56 (0.8%)+0.49−4.18 to +5.68Inconclusive
Wraith KingCore3rd item31 (0.7%)+0.49−3.48 to +4.54Inconclusive
Dragon KnightCore4th+ item183 (0.6%)+0.38−2.74 to +3.56Inconclusive
Phantom LancerCore2nd item63 (0.5%)+0.16−5.11 to +4.38Inconclusive
DawnbreakerCore2nd item64 (0.5%)+0.11−3.66 to +4.29Inconclusive
JuggernautCore3rd item130 (1.0%)+0.07−2.69 to +3.22Inconclusive
DawnbreakerCore1st item49 (0.4%)−0.02−5.96 to +5.54Inconclusive
Wraith KingCore4th+ item44 (0.5%)−0.10−3.48 to +3.53Inconclusive
JuggernautCore2nd item773 (5.4%)−0.50−2.76 to +2.45Inconclusive
DawnbreakerCore4th+ item255 (1.6%)−0.57−2.99 to +1.79Inconclusive
JuggernautCore4th+ item213 (0.8%)−0.99−3.39 to +1.81Inconclusive
NecrophosCore3rd item41 (0.3%)−1.49−7.19 to +2.94Inconclusive
NecrophosCore4th+ item94 (0.4%)−1.79−6.68 to +2.32Inconclusive
LifestealerCore4th+ item267 (1.3%)−2.47−5.04 to +0.41Inconclusive
Phantom LancerCore4th+ item45 (0.2%)−3.29−6.58 to −0.55Inconclusive

How to read

What is WPA?

Win probability added: how much choosing an item changed the chance of winning, in percentage points, compared with what the hero usually buys in that same slot and situation. +1.0 means the choice won 1 percentage point more than the typical purchase. Unlike raw win rate, it accounts for the game state when the item was bought, so an item bought from ahead doesn't look good just because the game was already won.

Build slots

1st, 2nd, 3rd and 4th+ count completed items. Components that become another item in most matches (Perseverance, Dragon Lance, Kaya) do not take a slot: the item they build into does. Boots get a card of their own.

Confirmed or not

The ✓ Better or ✕ Worse badge only appears when the effect passes the false discovery control (10%) in both ways of measuring: at the start of the decision and at the moment of purchase. Without a badge, the sign of the WPA is a trend.

Where the data comes from

Ranked (All Pick) matches parsed by OpenDota in the current patch. One model estimates the chance of winning from the game state; another, the effect of each item at the time of the decision. It is association, not causation: the model sees the game, not who is playing.