Back to Recipes Homepage
recipe picture
Recipe: Rav3n_pl GAB BiS Filter v2.9
Your rating: None Average: 5 (4 votes)


Name: Rav3n_pl GAB BiS Filter v2.9
ID: 103544
Created on: Mon, 05/25/2020 - 02:30
Updated on: Mon, 12/21/2020 - 06:34

Bands! Bands in space! GAB BiS v2. Now it have options dialog!

Best For


Joined: 09/24/2012
Groups: Go Science

Added equilizer for setting own bonus weights on filtered puzzles

Joined: 09/24/2012
Groups: Go Science

Warning: equalizer (and filter weight) doesn't work if fast filter is checked

Joined: 09/24/2012
Groups: Go Science

Option Mutate to H-Bonding only or non-H-Bonding only

Joined: 09/24/2012
Groups: Go Science
How to use

The idea behind this recipe is:
"Take your bonus first".

This has proven to give top ranking at the end.

Default weight of diverse bonus is 1. By giving a higher weight, in particular just after hand designing, the recipe only retains gains if these gains are not done on the cost of bonus. Of course, any gain in bonus is retained. If you seek bonus gain, you can even optionally choose to accept reducing score (which is a very good strategy if bonus is low).

At first impression, this strategy might give less point (than default weighting) at start game. But machine is blind to good design: just after hand design, the Foldit algorithm favors huge score gains (e.g. reducing a huge clash) even if the design is broken. Bonus are a better indication of good macro design than score.

At end game, most "red" spots are resolved, the normal bonus weights (1) are perfect.

The recipe can run 12 to 36 hours and still gain bonus and/or score if the options are good chosen. My most used options are: all filters weighting 100 (no equalizer). Mutate once, no fast filters.

Want to try?
Add to Cookbook!
To download recipes to your cookbook, you need to have the game client running.



Developed by: UW Center for Game Science, UW Institute for Protein Design, Northeastern University, Vanderbilt University Meiler Lab, UC Davis
Supported by: DARPA, NSF, NIH, HHMI, Amazon, Microsoft, Adobe, RosettaCommons