Suggest calling Sketchbook puzzle "Speed Folding"

Case number:671071-2003997
Topic:Game: Other
Opened by:phi16
Opened on:Wednesday, August 2, 2017 - 16:47
Last modified:Thursday, July 19, 2018 - 19:07

I suggest "Speed Folding", not for its being done quickly, but for its similarity to Speed Chess in terms of a finite end to an otherwise open-ended puzzle.

I have been advocating for something similar for years. I believe there is much to be learned by taking the same sort of approach that a chess player would take to learning Speed Chess: a more global view of what is happening, and what needs to be done, while developing your "stroke" as a tennis player would say. After a few games you will be able to appreciate an economy of effort.

Some other names I thought of along the way:

Folding by the Clock
Timed Folding
Folding Countdown
Folding 3, 2, 1...
Folding 10, 9, 8...
Folding 40, 39, 38...

(Wed, 08/02/2017 - 16:47  |  8 comments)

joshmiller's picture
User offline. Last seen 6 weeks 9 hours ago. Offline
Joined: 09/08/2017
Groups: Foldit Staff

While I appreciate this suggestion, the idea of Sketchbook puzzles as I understand it isn't to try to fold as quickly as possible, but to explore several greatly different kinds of folds instead of trying to optimize one particular one. In this way, it's more appropriate to call them "sketches" than to focus on the speed of folding them. I'm open to discussion, though.

Joined: 06/24/2008
Groups: Void Crushers

Josh, I think the idea is good BUT you are using mostly puzzles we have seen. How about puzzles we havent seen.

bertro's picture
User offline. Last seen 1 year 1 week ago. Offline
Joined: 05/02/2011
Groups: Beta Folders

My guess it will come to that...

bkoep's picture
User offline. Last seen 1 hour 6 min ago. Offline
Joined: 11/15/2012
Groups: Foldit Staff

Yes, we're looking forward to trying the Sketchbook style for new puzzles! But first we want to have a better idea of how the Sketchbook style affects puzzle results.

Revisited old puzzles are useful for that, because then we can make a direct comparison between the results of the original puzzle and the results of the Sketchbook puzzle. If it looks like the results of the Sketchbook puzzles are more useful, then we'll try the Sketchbook style with other puzzles!

Joined: 09/24/2012
Groups: Go Science


Do you make the assumption that big steps could give better results than a sum of small steps?

If it is so, may I understand that, for recipes on all puzzle types (not only sketchbook), we should perhaps include a "minppi = minimum points per iteration" condition before the recipe going further ? as I've seen in older recipes.

Hanto's picture
User offline. Last seen 43 weeks 4 days ago. Offline
Joined: 05/10/2008
Groups: None

Truly it's effective points per move folding i.e. Efficiency Folding.

jflat06's picture
User offline. Last seen 3 hours 22 min ago. Offline
Joined: 09/29/2010
Groups: Window Group

Thanks for the suggestions, everyone.

We've actually had a lot of discussion internally about the best name for these, and I think all of your suggestions were names we had talked about. The rules of the puzzle type are still a bit in flux, but once we settle on specific rules, we may look into changing the name to something that fits best.

phi16's picture
User offline. Last seen 9 hours 38 min ago. Offline
Joined: 12/23/2008

I agree with Josh Miller that 'it's more appropriate to call them "sketches" than to focus on the speed of folding them.' I was trying to raise this point when I suggested "Speed Folding" by likening Folding to Speed Chess. With the ability to count 'moves', there is no need to think about "speed". Besides, back when I first suggested the idea and recognized that everyone's CPU was different, I realized that there would be unfair advantages for those with faster CPU's.

What constitutes a 'move'? Is there a shorthand that could be written for each 'move', i.e. wiggle backbone = "Y", shake = "S", etc.? If so, could a script record a person's 'moves' so that they could be published for the team or for the general public to review and modify in the hopes of finding even more efficient results? This would be akin to the evolution of proteins by improving upon past performance, keeping the best results and experimenting with new mutations. Gramps taught me the importance of good note taking. Here it would be doubly important should someone want to improve their results. Another script could be used to execute the 'moves' file.

On a separate topic, has anyone explored the possibility of parrallel processing to improve folding? It seems to me that we always will have the problem of computer simulations being only able to work in a linear fashion. So that while in nature all atoms of a protein work simultaneously towards a proper conformation, with Folding scripts we are stuck having to work on one atom at a time. The result is having a random walk through the atoms to simulate work being done over the entire protein. What if in any 'move' a file of results were kept after each atom were 'wiggled' but no actual move was made. After the results of all atoms having been wiggled were generated, then the results would be used to find new XYZ coordinates for each of the atoms. This might be painfully slow, but might give us better results.



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