Phantom Dancer
2,650 gold. It gives 65% attack speed and 25% crit chance. Finished in 3,102 recorded games on patch 16.14.
Patch 16.14 · item shop version 16.14.1
Ashe bot finishes Phantom Dancer more than any other champion and role on patch 16.14
965 of the 3,102 recorded games that finished it.
Phantom Dancer costs 2,650 gold and gives 65% attack speed and 25% crit chance. 34 different finished builds on patch 16.14 contain it, across 10 champion and role pairs, and 3,102 recorded games sit behind those builds. Ashe bot finishes it in more of them than any other champion and role.
The middle purchase of it lands at 18:32 on the game clock, and half of the 424 recorded purchases fall between 16:24 and 21:07.
Patch 16.14 changed Phantom Dancer, so we measured it: the win rate of games with this item after the patch against the same games before it, in wins per 100 games. The range covers zero, so this could have gone either way. The full report is further down this page and on the patch 16.14 page.
There is no win rate anywhere on this page, on purpose. The method note says why.
These numbers come from 108,616 ranked games we recorded on EUW, Korea and NA, found by following Diamond players and the people they play with. That is not every rank and it is not every region. Everything on this page is patch 16.14. Counted up to .
What Phantom Dancer gives
Riot's shop numbers, from Data Dragon 16.14.1. Each bar runs to the most of that stat any measured item on this patch gives, cost to the dearest. Nothing here is measured by us.
- cost
- 2,650 gold
- attack speed
- 65%
- crit chance
- 25%
Who finishes Phantom Dancer
9 champion and role pairs cleared 30 recorded games with it, across bot, top and mid. The bar on each chip is that pair's share of this item's own recorded games: a count, never a win rate.
Ashe bot · 965 games
Jinx bot · 882 games
Garen top · 442 games
Kai'Sa bot · 279 games
Aphelios bot · 218 games
Tryndamere top · 113 games
Caitlyn bot · 79 games
Zeri bot · 60 games
Garen mid · 40 games
1 champion and role pair finished it in under 30 recorded games each and are left out of the chips, 24 games between them. They are still counted in the item total above. The share counts finished builds. It is not a win rate and the chips are not sorted by one.
What players finish with Phantom Dancer
Items that turned up in the same finished build, counted. Riot's data here ships no recipe list, so this is what players bought together, never a build path.
Infinity Edge 2,270 games with both · 73.2% 
Hexoptics C44 2,125 games with both · 68.5% 
Mortal Reminder 766 games with both · 24.7% 
Lord Dominik's Regards 666 games with both · 21.5% 
Stridebreaker 482 games with both · 15.5% 
Kraken Slayer 319 games with both · 10.3% 
Guinsoo's Rageblade 279 games with both · 9.0% 
Dead Man's Plate 168 games with both · 5.4% 
Ravenous Hydra 113 games with both · 3.6% 
Guardian Angel 72 games with both · 2.3% 
Bloodthirster 44 games with both · 1.4% 
Stormrazor 39 games with both · 1.3% The share is how many of Phantom Dancer's own 3,102 recorded games also finished that item, so it can never pass 100%. 1 more items appeared beside it in under 30 games each and are left out. Counted, not looked up: it is what players bought together, and it is not a recipe.
When Phantom Dancer gets bought
Stamped at the moment of the buy, so how the game ended cannot move it.
- all recorded buys
- 18:32
- bought while ahead on gold
- 17:58
- bought while not ahead
- 18:51
424 recorded buys: 195 ahead on gold, 229 not ahead. Ahead means a lead of at least 1,000 gold at the buy. Bars run to 40:00 on the game clock; half of all buys fall between 16:24 and 21:07.
These are every recorded buy of Phantom Dancer, not just the first one in a game, so the middle time sits later than the time you would first build it. The two rows underneath split the same buys by whether the buyer was ahead on gold when they bought.
What patch 16.14 changed
The one measurement here that a win rate cannot fake.
Phantom Dancer
games after the patch against the same games before it
-0.3 wins per 100 games(range -1.6 to 1.0) 27,492 games Measured the range covers zero, so this could go either way 8,637 games on the thinner side of the boundary, 10 days each way. The range covers zero, so this could have gone either way. The full report is on the patch 16.14 page.
How this page is counted
Short version here. Every number above says which of these it came from.
Games come from finished builds: each recorded player lands in exactly one finished item set, so the buyers chips add up without counting anybody twice. We keep the most common finished builds per champion and role, so the total is a lower bound. How items are counted.
Why there is no win rate on this page
A finished item set is read at the end of the game. A player who bought Phantom Dancer and lost the game at twenty minutes never finished it, so they are not in these rows. Every player who is in them survived long enough to finish a 2,650 gold item. That makes the win rate of the pool high for reasons that have nothing to do with the item, and worse, we cannot say by how much.
Every other site prints that number anyway. We print the games, the timing at the buy, and the before-and-after comparison when there is one, and we say when we have nothing.
ko.lol for desktop
Your champion pool, your matchups and your patch notes, on your desktop while you play.
Free. For Windows.