I figured we could add some of our thoughts into this thread on LoL DFS. Doing a couple linups stacked with iG and Fnatic. Faker is kind of a scary pick due to possibly subbing a game.

I'm trying to put together some fantasy lineups for Worlds too. Do you know of a website that has stats on player/team results? I'm new to League (i'm a Dota player) but a LoL equivalent of http://datdota.com or https://www.dotabuff.com/esports/teams would be ideal.

What odds do you think there are of Faker being subbed?

This is close. http://oracleselixir.com/2015/09/uns...om-patch-5-18/

I highly doubt they bench Faker on game 1. I have him in 3/5 lineups.

Enough to make him -ev to draft ever.

Yeah this is pretty great, thanks! What data do you think will be most useful (ie most recent league with a decent sample size)? I'm thinking 2015 Summer Regular Season?

I've got a lot in action and did a lot of datamining in preparation for this.

Things I've learned:

The factor that most directly correlates with player performance in an individual game is total # of team kills

Winning the game is also very highly correlated with player performance. Due to the snowball-ey nature of League, teams that win the game usually had a bunch of kills that set up that win. Losing teams rarely score well.

Off of that note, the long-drawn out bloodbath games that people think happen, just don't happen all that often. I've seen a bunch of people predicting that FNC/IG will be a total ****show with 30+ kills on each side or something. It is very rare to see a game where each team gets over 20 kills, even in the LPL and LMS.

Game length is only weakly correlated with fantasy performance. If you think about how League works, if there were a ton of kills in a game in the first 20 minutes, chances are that one team racked up most of those and is snowballing to a win. It's rare for a game to be even at 20 minutes and a bloodbath. Once you hit the 20 minute mark, the teams at this level are so good at turning kills into objectives with the longer death timers that each team fight goes a long way to deciding the game. Yeah those minutes from 30-50 seem like they'd be ripe opportunities for kills and assists, but in reality, one clean pick usually means multiple towers/inhibs, or a team fight where one team wins by more than 2 kills usually means it's game over. People read way too much into Solo Queue type games and think that it's how it can work, but kills are pretty rare after 35 minutes unless it's in a team fight that ultimately decides the game.

I'd prob go with players from CLG+AHQ+SKT+KOO, with as many as possible from CLG+AHQ and Team also AHQ or CLG and most definitely stay away from IG and FNC.

I think Ziv is a really strong play day 1 and will likely have him in every lineup.

With how we're expecting the meta to be really top lane focused this worlds are we moving towards playing mostly tops at flex over adcs? Or should we still pretty much always have adc's at flex?

No a lot these champions are splitpushers and toplane is naturally involved less in teamfights anyways and thus gives in average less points.

Follow up question - the site is great - but it seems focused on NA/EU/Korea. Any idea where i'd be able to get stats for games played by, say, iG or h2k or paiN gaming?

Really appreciate the help. Just trying to throw together some quick fun lineups for a sweat in the Worlds. If anyone needs help putting together some Dota lineups i'd be more than happy to return the favour

H2K is Europe, IG is China and Pain gaming you don't pick.

Ah I found this:

Which explains the absence of that data. I'll just make do without, especially since you said to definitely stay away from IG and Fnatic. Cheers.

I think IG vs FNC is a tricky one, it's really tempting to take either side. It could be a bloodbath or either team could take it one sided, it's high variance and I'm not sure if it's worth, 50% of the time you are wrong and your lineup is out of the race.


China is a pretty aggressive Region and their games should probably yield higher points on average.

scoreboards exist for LPL games but sadly none of the full stats like on oracles

Meh. Was trying to post a lineup here via mobile. Not so easily done

Stacking ADC/Support from same team in both your regular slots and flex = +EV?

Having a Support as flex seems -ev most of the time as they on average get less points.

TOP

Huni 18.3 lock
JNG

Hai 0 lock
MID

Febiven 20.66 lock
ADC

Bang 0 lock
SUP

Wolf 0 lock
FLEX

Deft 0 lock
FLEX

Pobelter 0 lock
TEAM

Fnatic 20 lock
TOTAL FANTASY POINTS: 42.16


Let a friend make my team for me have no clue whats going on what are my chances of making money here?

Well Hai doesn't play mid but Jungle and even in Jungle he would be a horrible pick so yeah doesn't look all that good of a draft.

For the team pick, seems like you would want to pick a team that either has trouble closing out games or that is not a heavy favorite. A long drawn out close game where the winning team needs to get 5 drags and multiple barons seems better than a 1 sided stomp that finishes fast enough to maybe get the quick win bonus. Obv seems riskier and higher variance and you might not get the points for a win tho

The surrender was pretty bad for all of the Fnatic picks. They were going to slow push and probably rack up at least four more kills.

Saved me, I had zztai and kitties.

I thought IG was going to win.
I'm pretty heavy in salary with adc's and mids.
Still doubt I win anything.

Lost all the $$ today. Playing mostly AHQ and EDG players in my cash lineup and fading C9 and largely fading KOO will do that. Had a few KOO stacks that almost won some GPP's but didn't get the right combos. Back at it tomorrow.

i wouldn't normally double down and chase losses, but I'm inclined to stay confident given how bad I felt everyone else played

Yeah Ziv was a real disaster. lolme