Back to Recipes Homepage
recipe picture
Recipe: JET 3.5
Your rating: None Average: 4.7 (55 votes)


Name: JET 3.5
ID: 100931
Created on: Sun, 04/26/2015 - 12:43
Updated on: Fri, 12/30/2016 - 16:22

Multiple LWS and AT with puzzle exploration, by Bruno Kestemont

Best For


Joined: 09/24/2012
Groups: Go Science
Join Evolver Team

This is my very first recipe that I developed with the help of Jean-Bob who learned me to script in Lua2. I added more thinks with the time.

I wrote and shared it with L'Alliance Francophone (original French version) and later on with Beta Folder (multilingual EN-FR version). This was our "secret" end game recipe that helped me to reach rank 2 evolver at this time.

Basically, it's not much better than current banded worm pairs infinite filter or sd-multiwalk forever filter that I discovered here in Go Science. It's a concurrent of sdmwf, a complement of bmpif.

What it does:

1) Stanley exploration module
Gives some information on the puzzle and decides some internal starting options.
Nowadays with filters and bonus, this module is not giving always the right information ("remaining expected gains" is quite underestimated most of the time). But it can help. I used JET many times only to get some info in the log, like "expected target" (this has to be revised considering filter boni). With this info, I could decide to run other recipes, knowing that JET will not gain more than 2-5 pts on end game.

2) A succession of LWS and AT recipes, with more and more hard and long attempts with time. Failing recipes are skipped next time.

In a way, JET "guaranties" you to get the 2 points in one night, and "join the evolvers". JET always continues to gain some millipoints.

Used for

I use it when going to sleep or to work: it's a long lasting recipe that yields nice after several hours.
I use it at very begin in design puzzles, because it "fixes" the protein keeping most of the structure (quickfix is better for this, but its potential is finished after 1 hour - JET is still gaining after days).
Or at the very end, when competing with other teams for the top.
I use it also for the latest millipoints before to share (like getting .990 pts before to share).
If the race expires when I'm absent, I usually run JET on High wp. If there is time enough, I run it preferably on low wp.

I almost never use it on mid game because it has the tendency to stick the puzzle where it is.

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, Boehringer Ingelheim, RosettaCommons