infolol Editorial guide
Survivorship bias in completed-item statistics
Why a high win rate for completed core items cannot be treated as proof that buying them caused the wins.
Only games that reached completion remain in the sample
An expensive three-item core can be completed only in games that lasted long enough and produced enough gold. Early losses often end before the same sequence is possible.
A high late-build win rate therefore mixes item performance with the fact that the player was already funded and the game continued. It is not proof that the items caused the wins.
A core sequence is not a purchase script
The displayed core is a frequently observed completed-item sequence from match timelines. It does not optimize every component purchase or measure the fight immediately after each recall.
Different recall gold and enemy damage profiles can require different components even when the eventual completed item is the same.
Values to check before win rate
Start with build games and build share. A rare three-item combination may be concentrated in long or already favorable matches. A first item completed in many more games can provide a more useful default direction.
- Expect more survivorship bias as the sequence becomes longer.
- Treat rare high-win combinations as situational observations.
- Adjust defensive items to the enemy composition and current carry.
A conclusion the site does not provide
The current aggregation does not measure the causal effect of an item purchase. infolol does not claim that buying one item raises win probability by a stated amount; it reports observed frequency and outcomes as decision support.