[Code S] Ro4 Day 2 Season 3 2014 - Page 3

On September 20 2014 09:10 Die4Ever wrote:
Hey guys, what do you think of the new feature?

http://sc2.4ever.tv/?page=tournament&tid=20

On September 20 2014 11:19 The_Templar wrote:

Took a second look. Why does it skip 2nd for zest but not Cure?

On September 20 2014 11:26 Die4Ever wrote:

WCS Predictor only finds and tallies events that either have a high enough score (how much it CHANGES their Blizzcon chances combined with how often it happens), are an upcoming match, or are a 1st place. You can see that the entries for Zest are 4th place (losing his upcoming match), and the 1st place (cause it's a 1st place!), the 2nd place is skipped because it doesn't fulfill any of the 3 criteria (score cause no change to Blizzcon chances, upcoming match, or 1st place). Cure gets them all cause the 2nd place does change his Blizzcon chances a bit.

You were close KingofdaHipHop

On September 20 2014 11:32 Die4Ever wrote:

Yes that's true it won't ever round down to 0. What makes you think otherwise? The table is showing Blizzcon chances for finishing at placements, the 2nd place slot means if you lose in the finals.

On September 20 2014 11:38 Die4Ever wrote:

Ah, I get what you're asking I think. Right now his chances are at ~0.06% because it's including the possibility of him winning the GSL, you can see that 1st place brings him up to ~0.28%. However if he gets 2nd then he can no longer get 1st so that possibility is gone, and 2nd place is just never enough for Cure to make it. So him getting 2nd place changes his Blizzcon chances because it brings them down by ~0.06% down to 0%.

On September 20 2014 11:57 Die4Ever wrote:

Oh. The 4th and 2nd place change Cure's current chances, not the successive chances, scoring of events only happens on an individual basis, since events are examined after all the simulations for the initial block are done. The blank spots are because WCS Predictor just isn't counting that event in order to save CPU and RAM. The way it works is it runs the initial block of simulations (currently 500,000 samples) and then with that initial block it finds events and scores them and then trims the list to the highest scoring ones for each player, and then it remembers those lists and adds on to the counters for them after each block of simulations. So Zest's 2nd place event was trimmed because it was able to recognize in isolation that it doesn't need to track it because it doesn't change Zest's chances.

It doesn't look at a chain of events, it just looks at each event individually.

On September 20 2014 13:28 juvenal wrote:
Wha'ts with bogus throwing his penis at cure?