Boards›[ARCHIVED] Fantasy LCS

Spring 2017 - Stats-based Projections - Week 5 Onwards

Pseudo Q·2/15/2017, 10:18:07 AM·3 votes·841 views

Hi,

Last week I showed my projections for Weeks 4-6.

Due to popular demand (a poll in my previous post), I'm now showing projections until the end of the split, but restructured to (hopefully) make it more digestible.

https://docs.google.com/spreadsheets/d/1vS_tdqwjerf8x9tq_6VTk_KJ3YFW_qd8NIiqFVg1Mdw/edit?usp=sharing

As before, you will find 3 tabs:

  • LCS_Stats : Which shows average deaths and kills per game, and which teams players are generally going to be more valuable based on scheduled matchup (according to the prior two factors).
  • Best_Format_Projections : Which groups players by role, and sorts them in order of who is going to be most valuable for Week 5 in the 'Best Game' format.
  • First2_Format_Projections : Which groups players by role, and sorts them in order of who is going to be most valuable for Week 5 in the 'First2' format.

In both Format_Projection tabs, you will see:

  • My projections (W# Proj.) for previous weeks have been provided alongside the actual results for each week (Week#).
  • The scheduled matchups, alongside the projections for the current week (including Riot's estimates (W# Riot)).
  • Projections for the remaining weeks of play.
  • Accumulative values for each player's projections are also supplied for Weeks 5-7 and for Weeks 5-9 (until the end of the Fantasy split).

I recommend using these stats and projections in conjunction with your own intuition for whether matches are going to turn into macro-fests (which don't usually produce many points), complete shut-outs, utter carnage, or anything else in between.

Good luck!

Extra Notes... (optional reading below)

C9 vs FLY stats

I'm still waiting for the stats to be updated for C9 vs FLY. So, although the projections aren't technically finalised yet, they should still point you in the right direction. When/if the stats for game 3 get fixed, I'll update the sheet.

Red Player Names

Red player names means they have been substituted at some point this split.

Omitted Players

Note that some players have been omitted, due to appearing in only a single week, and not completing the full series, so I couldn't generate reliable projections for them.

Accuracy Percentages

I did start calculating my accuracy with a percentage and putting the values on another tab. Then Wisdom and Hiiva (and some others) sent my percentages into the realms of meaningless with a scores like 0.71 and -0.31 respectively. When I tried to calculate a percentage accuracy, using the actual score as 100% (and all projected points were turned into a loss of percentage - regardless of whether it was higher or lower), I ended up with a -1521.13% accuracy by projecting that Wisdom would get 11.51 in the First2 format. So instead, I've simply listed my previous weeks projections against the actual results, rather than trying to be clever.

0 Comments