20220714-9144c40722-win_x86 Inconsistent objective coloring

Case number:699969-2013405
Opened by:Bletchley Park
Opened on:Saturday, July 30, 2022 - 08:59
Last modified:Wednesday, August 3, 2022 - 11:50

The STATISFIED objective Core complex is WHITE with 0 points
The UNSTATISFIED objective Core existance monomer is GREEN (with points).
The STATISFIED objective SS design is GREEN with points.

Is the coloring determined by POINTS ? Why not determine it by statisfaction, which would, to me, make perfect sense.

This would make sense:
GREEN = statisfied
WHITE - unstatisfied .
RED = penalized (too many residues in helices for example)
and points should be irrelevant to the coloring

For colorblind mode these colors should be explained and expressed differently.

(Sat, 07/30/2022 - 08:59  |  2 comments)

bkoep's picture
User offline. Last seen 20 hours 59 min ago. Offline
Joined: 11/15/2012
Groups: Foldit Staff

Thanks for the feedback Bletchley Park! The current coloring scheme is indeed determined by points. I think your proposed scheme makes sense, but would require some revisions to how Objectives are implemented. In short, the problem is that concepts like satisfied, unsatisfied, and penalized are not actually defined for Objectives (but maybe they should be).

Currently, Objectives are not necessarily programmed with any kind of satisfaction threshold. Some Objectives do have caps (like BUNS cannot be improved beyond zero BUNS); some Objectives are artificially capped (like we usually cap DDG at -40 kcal/mol); other Objectives are not explicitly capped (like HBond Networks), but for these we can usually calculate a theoretical maximum based on the number of residues, maximum allowed bonds, etc.

We also do not have a clear distinction between unsatisfied and penalized. If an Objective is not awarding its full bonus, it is effectively a penalty. No matter if you are getting +50 points for a +100 maximum Objective, or if you are getting -50 for a +0 maximum Objective -- the effect is identical: you are 50 points shy of satisfaction.

As an aside, this last point leads us to some tricky questions about how we want to frame Objectives for players (especially new players). If a player sees a penalty like "-50 (max +0)", this communicates something like "Oh no, my design has a problem and I urgently need to fix it." Alternatively, something like "+50 (max +100)" communicates a message like "Hey, my design is pretty good, even though it could be better." The math is the same, but the message is very different.

Joined: 05/19/2009
Groups: Contenders

Your points about the framing is spot on, I struggle with this too.


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