20 replies [Last post]
Joined: 05/10/2008
Groups: None

I play both single & group. The new 'group sharing' for group play is GREAT, but when I view a shared solution, it ranks me up to that player's score. It seems unfair for a group member to jump ranks ahead of single players based on group work. In #59, ranks 2,3,4,& 5 are same-score/same-group, displacing several single players who worked alone for hours to achieve those top ranks. Group=group score; single=single score. The single players have a right to expect equal playing fields. (Also, for me, in single play I prefer to earn my own score based on my own work.)

Joined: 05/09/2008
Agreed

I feel the same way. I don't mind sharing my structure that has yielded me high scores, but then everybody from the group can go and download it and then get the same score which makes it a little unfair because they didn't do anything. I also feel bad when I look at somebody else's fold in the group and then it ranks me up to them because I know they spent a good amount of time getting to that higher score. I have completely stopped viewing the other peoples proteins. I think the way to resolve this is that you should either not get any points for viewing the structure. Or on the other hand if you can make improvements on the structure you would be able to get the points. But I do not think anybody that just views the protein should take my hard earned points, because I don't feel right doing that.

Joined: 05/15/2008
Groups: None
Agreed

I have to agree...I have the same problem. I was stuck at one point and looked at a file share of a person in my group and it jumped me up to their level. I was able to improve upon it, but I feel like they did most of the work to get me there (jumped from #26 to #4!!)

Don't like the idea of that. I would much rather just be able to look at their model for comparison without suddenly acquiring credit for all their work.

Joined: 05/10/2008
On one hand, for internal

On one hand, for internal competition purposes, not instantly getting to start where someone else left off would be more fair. Therefore, we ought to be able to view each others' solutions, but not incorporate them directly as our own.

On the other hand, if internal competition weren't a motivating factor at all, it's clear that a great deal of improvement could come more quickly if, upon deciding that someone else's work is an entirely better direction than your own, then you could start from where they are. It's kind of like a genetic algorithm, but with the power of human minds directly engaged, instead of being engaged through a programmed computer.

I would suggest that both capabilities be made available, but that solutions be markable in a way that can't be undone by the user. When a solution includes another user's previously uploaded version, then the solution is marked and excluded from being further scored competitively. Perhaps some people would learn how to hack out the mark, but that's really not a reason not to try it out.

btbenj's picture
User offline. Last seen 6 years 51 weeks ago. Offline
Joined: 11/17/2007
Groups: None
Anjen, in fact the group

Anjen, in fact the group sharing feature was already in the game a long long time before the initial release. They had to take it off because with all the new people, there was no way their servers could handle all the data from shared solutions. Now it seems as though they have solved this problem.

I think that the single/group scoring method may be very useful, but I also believe that people who are playing in groups should have some sort of advantage, as they have spent time organizing a good team, and if everyone is helping, it is a very good cooperative effort.

zoran's picture
User offline. Last seen 6 years 37 weeks ago. Offline
Joined: 11/10/2007
Groups: Window Group
the intention of group play

Hi folks,

thank you for bringing this to our attention.  The group play was originally designed to be small "labs" where people would work together to discover solutions, and sharing interim solution is an important tool in this process.  It appears that groups have taken on a different meaning, where people from the entire countries end up in the same group: obviously at that scale, one may not want to share the results with everyone.  As you can imagine it is straighforward to let people either have their own group (ferzle's group for example) and focus on individual competitions, or choose to share solutions with others and thus focus on group play. 

The dynamics of playing both groups and individual play and not wanting to share your individual score with other members of the group is tricky.  We have some preliminary possibility, but it would be useful to find out what exact score dynamic would you like to see.  Please respond here with specific scenarios.

As always, thanks for your feedback.

Joined: 05/24/2008
Can humans really help computers fold proteins?

"QUOTE"

We’re collecting data to find out if humans' pattern-recognition and puzzle-solving abilities make them more efficient than existing computer programs at pattern-folding tasks.

If this turns out to be true, we can then teach human strategies to computers and fold proteins faster than ever!

- I was the Founding President of the Deutschland Touch Rugby Association, I stood alone in the middle of the English Garden practicing alone, breaking the law too!! I was in that paddock for weeks until someone got the nerve to ask me what I was doing with the odd shaped ball... and after explaining the rules, my passion and goals did they joined in... within three months, I had over one hundred members, I captained the German Mens Team against France in the First ever international match and started something I believed in fully, even as nieve as I was!!! It was what I made of it at first, but then the people made it so much better by being apart of it.

That original team are now running it, throughout germany with the same dedication and passion if not more.

They played in the World Champs and have gone onto something I never imagined would happen... but i believed if it did, everyone involved would share that same passion with everyone.

- Why do i share personal experience with you? -

Foldit is no different...

I stand here and say, together or alone, no matter the score... I am going to give it my best and bring as many people together as possible to give us a better chance.

If you join me, great, if not... great.

At the end of the day, no one knows I was the first person to create the German Touch Rugby Association(in 2002) it doesnt matter to me, but for all those who have played the game, they know how wonderful and amazing you feel after being apart of it.

Respectfully,

Renton Braden Mathew Innes
Kiwi & Child of the Universe.

Joined: 05/10/2008
Groups: None
group sharing and personal scores

Well, I tried to make my position as clear as possible in my original post. It seems that some who have responded understood my point exactly... others didn't. For the benefit of those who believe that this is about points and ranks, please let me clarify.

I started in foldit as a single player. It was my intention to remain a single player. In getting to know a few of the other players through group chat, a question arose as to whether or not some of the 'lone players' who had no background in these sciences could form a group that could competitively match some of the "professional lab groups" in folding and scoring. I was willing to see if that was a possibility. It seemed to me that utilizing both professionals and amateurs in this way could only enhance the information for which this project was developed. The best of both worlds!

I was EAGERLY looking forward to the solution-sharing capability for the groups. I felt that with us all being able to work together on the proteins there was an immense possibility of huge and unexpected amounts of further data for the project. It was an exciting premise and I was ready and waiting to "go to work". Upon my very first viewing of one of my group-mate's proteins, I noticed that it jumped my rank in PERSONAL play significantly and would not allow me to return to my own rank in PERSONAL play.

It is my own code - for ME - (no better or worse than anyone else's) that I make my way on what I have earned, not by riding on the shirt-tails of someone else or stepping over the backs of others who did not have the same chance or advantage that I did. I found it very disturbing that ONLY because I had viewed someone else's work, and totally without the knowlege that it would happen, I had suddenly risen over 100 points and 50 ranks above people who were working alone to fold the proteins. For that reason, I have not viewed any other work from my group members, therefore voluntarily - but with great concern for the lost possibilities - excluding myself from true group play on the current puzzles.

Foldit has been such a delightful experience up until now. The total support and comraderie in the chat room for beginners and for experienced players has been refreshing and helpful. There has been a very light and mellow attitude about sharing tips and designs and all seemed to be working together for the good of the project. AND, it has been great fun to compete against one another for scores and ranks, congratulating one another for each and every hard-earned point and playfully bashing one another when we are passed by someone else. We were all in it together and I have continued to be surprised and proud of the intense work and the hundreds (probably thousands now) of hours that people have put in for the sake of science AND for the sake of the game.

I did not hear a single complaint - not a one - when those of us who are now more experienced players signed in one day and found that we had been locked out of the beginner puzzles. We understood the reasoning on this - that the newer players needed to have a place to play and learn where they had an expectation of doing well based on their time and efforts, not being automatically trounced by people who had been playing much longer than they had. Obviously, the developers realized that this was a more equitable and productive way for everyone to compete at their own playing level or they wouldn't have made that change.

It seems to me that we are dealing with the exact same concept when addressing shared solutions and group play and their affect on personal scores. In a comment either on this thread or on the "swords and guns" thread, someone stated that the groups were 'small' entities that had banded together for better folding, or something to that effect. I am certain that that is the way it started out; however, if you will look at some of the group rosters, some of them contain 50, 75, or 100 people. There is NO possibility whatsoever for a single individual to compete against 68 people when those 68 are able to share and improve upon one another's work.

For those who are saying that it is for science only and it's not about the points so "why even worry about this?", I feel that you are either missing or neglecting the very strong part of human nature which seeks positive feedback and success. The point has been made that nobody is being paid for this work. Part of the reward in addition to our genuine desire to contribute to scientific knowlege is to see how well we can do in this project and to see how we are doing within a larger number of our peers from all over the world.

It is NOT "all about the points" but the points and ranks are, in a way, a source of "payment" for the very, very hard work that we are doing for the sake of science. For many of us who have given so freely both in time and effort on folding and time and effort in chat helping other players, it came as a shock and disappointment that we were playing against unknown advantages and tools for other players.

I am VERY excited about the possibility of a rebuild tool. I am happy that it is being tested because the faster it is tested, the faster we will have it. (Hooray!!!) However, it should have been tested WITHOUT the ability to effect scores in competition play. The same goes for shared solutions. They are a WONDERFUL resource for group play, but most definitely NOT fair in personal play.

For those of you who have written in and indicated that we who feel this way are just talking "sour grapes" and/or just worried about our scores and ranks, PLEASE go back and read again what we have written. It is NOT about the scores... it is about a PRINCIPLE: an "expectation of fairness" (a phrase borrowed from another foldit player) - a level playing field for ALL, not just for groups or folks who are testing new tools. For me, it's not the fact that, every time I sign in now, I have lost 20 ranks to group play, it is that the work I am doing seems futile and wasted compared to professional lab groups or new tools that are being tested. My time is more precious to me than to spend it feeling that I am merely struggling over a period of hours to reproduce work that a group has done in significantly less time.

I have a busy life. I have many more things to do that are much more important to my personal life than folding. I spend the time that I do in folding because it is important for the betterment of all people AND because it is fun. I have tried to be respectful and understanding of the opinions of those who feel that "all is fair so use whatever you can to get ahead". I would hope that you will be respectful of my opinions as well.

Benj, a personal note to you: You sort of got caught in the middle of all this by accident, I believe. That has concerned me greatly, particularly with you being so young. Twice in chat, I have told you very clearly that this is in NO way personal, that I have nothing against you, and that I am not in any way making any personal comments about you. I hope you will believe that. You are still very young. You see the world very differently than some of us older folks do. We saw the world very differently when we were your age, too. That's the way it works - that's the way it is supposed to be! I would offer to you that this can be a wonderful learning experience for you if you will allow it to be. Read and re-read what has been written about all of this, both the ones that agree with you and - most especially - the ones who do NOT agree with you. This is a perfect microcosm of 'the real world'. Learning how to share differing opinions in a respectful way can be invaluable to you as you age and venture further and further into situations that require cooperation and compromise. I hope you haven't been hurt by any of this. That would break my heart. I hope that you - and ALL OF US - can learn from it regardless of whether we are in agreement about it or not.

Folders and developers, I've explained where I'm coming from as well as I know how to. I am trying to regain my earlier excitement for the project and my willingness to "lose sleep for the further enrichment of personkind" (smile). I hope we can come to some "fix" on this that will allow everyone on both sides of this issue to feel that they have been heard and understood.

Thank you for taking the time to read this. Best of luck in your folding!

champuli

Joined: 05/13/2008
Groups: Penny-Arcade
Suggestion for a fix

I understand where everyone is coming from here...and it is annoying to see a ton of people from a group all tied for the same score because of a share. So I was thinking this might help...

What if only the first person in a group to reach a score is ranked. If you look at the shared solution, you get the points, but you are not ranked on the leader board because you didn't get that score originally. But then if you improve that score, you can then be ranked because you surpassed that group members score.

Example: Tommy gets 9100 on a puzzle then shares it. He gets ranked at #90 on the board. Jenny on his team looks at his solution and gets to 9100, but since she used the solution, she doesn't get that ranking displayed on the boards. Then Jenny plays with the solution and gets 9200. She is then ranked accordingly because it was her work.

This at least reduces the amount of ranks taken up by group share. Comments and suggestions on what everyone thinks...

Joined: 05/13/2008
Groups: Minions of TWIS
All for One... One for All

There are over 170 people in the "minions of twis" group. If one of us shares a top score, and everyone of us uses it, it would drop all other scores by more than 170 points... as awesome as that would be(go team!)... It does make the single player score seem obsolete. Everybody on that team get to jump right to the best design...

This isn't a bad thing...

If you are on this leading team, the starting point becomes the front design. What happens after that is what this game is all about. Finding the best solutions, fair shake or not.

However it might limit the creative aproach in one way. I may be stretching it here, but if a team follows a shared design early that sort of stalls out or gets stuck, the whole team worked on a dud design.

It may not be possible, but i would suggest having the shared solutions not available til say half way through the puzzle life. That way you would have maximum potental the combination of fresh ideas, and new ones...
Maybe lock in the single score at the half way point?
There's got to be some wiggle room in there for a solution that everyone will like...

Joined: 05/15/2008
Groups: Minions of TWIS
anjen is harsh

Position: Someone that loads shared protein shouldn't get that score as their own *unless* they improve upon it. Even if they improve upon it I don't think they should get the full scoring during the puzzle close tally for their global rank, but they should get something possibly even the full scoring if they improved the protein significantly. I also think that attribution should be given to the source of the shared protein (as I mentioned in the other thread) because not giving such attribution is plagiarism. The assumption made by teams would be made up of multiple people from the same lab was flawed. Teams are made up of people with varied backgrounds from geographically disperse locations. Welcome to the Internet dev team.

I think the client needs to have the feature to overlay the shared protein in the same manner that the tutorial had for guides. I think that's how people would want it for just viewing someone else's shared folds. Good luck aligning the proteins.

Back to the main topic -

Scenario: Someone shares an early fold for 8960 points. Another group member uses this fold and pushes on to 11020. I don't think anybody can complain that the second folder shouldn't get the rank for the 11020 fold because that's a substantial improvement and could likely be 100-200 rank difference. Anjen's "thou shalt get nothing if you used a shared fold" proposal is extremely harsh for this scenario. So how much of an improvement over the shared fold before people agree that the rank is deserved?

more scenarios to follow later.

Diderot's picture
User offline. Last seen 7 years 2 days ago. Offline
Joined: 05/09/2008
Here's a scenario...

...we all go back to the way it was before group sharing was turned on. We just forget that the last few days even happened.

It was more fun the old way for everyone, because everyone got to make some discoveries on their own. This meant more people would play, and they would play more often, too, which means better folding.

I suspect it's better science for two other reasons: It means that everyone must learn how to fold by themselves if they want to advance in rank, and this builds the pool of human skill. Also, it minimizes the path-dependence that comes with 170 people on a team all sharing a solution.

Now, I can't prove that it's better science -- I'm not a scientist, and I can't really evaluate the data. But it seems there's at least two or three good arguments for why it might be. I'd like to see team sharing abolished. If you want to show someone your protein, take a screenshot and send it!

Joined: 05/13/2008
Groups: Penny-Arcade
Science vs. Competition

While I think everyone agrees that the most fair way for scoring and ranking is to not allow sharing, but this is not the best solution for the science and discovery the game is based on. The sharing of ideas and discovery is integral to science. Someone might get stuck on an initial fold and give up, but they share that fold and someone uses that starting point to create a groundbreaking protein structure. Without sharing that scenario would never happen.

Removing sharing will severely hinder the potential of the foldit experiment. We all have to remember that this is a science experiment first, fun game second. They will find a compromise somewhere along the lines.

Joined: 05/21/2008
Groups: None
The first time I loaded

The first time I loaded another person's solution, I didn't really like the ranks I gained... I immediately thought that it would make the individual ranking meaningless (I have to say however that I also improved those scores afterward).

Maybe it would be better to have two distinct puzzle competitions (with 2 different proteins), one for the groups, one for the individual.

Or we could have 2 parts for each puzzle (with separate rankings and time): the first where sharing isn't allowed, and the second part where it is allowed : this way everyone can search for it's best individual structure at first, and in a second time all the members of the groups share their structures and try to improve them, this time as a group.

btbenj's picture
User offline. Last seen 6 years 51 weeks ago. Offline
Joined: 11/17/2007
Groups: None
It seems like a logical

It seems like a logical solution for this is having group scores and individual scores separate. That should not be too hard, and if you post it on the suggestions board, the developers will try to address it as soon as possible.

The only problem would be group scoring, would you get more points for improving a group score, or would it all go to the group points?

Joined: 05/10/2008
Groups: None
RE: Benj - it seems logical

Well, Benj... I actually laughed aloud with genuine delight when I read your most recent post. Thank you for saying in one short paragraph what I evidently could not seem to make clear in my previous (and much longer) oratory or in my original message.

My whole point from the beginning was to SEPARATE the scores so that groups can do what they were created to do (work together on solutions) and individual players can continue making their equally valuable contributions to this project. I am certain that, if given a chance, both playing processes will produce wonderful and unexpected data.

In an odd and ironic way, this has suddenly become hugely funny to me (humor... go figure... ). Your "it seems logical" post is, if I am interpreting it correctly, exactly what I have been asking for from the moment I was accidently ranked up with the first shared solution. Thank you for finally understanding that and stating it clearly.

zoran's picture
User offline. Last seen 6 years 37 weeks ago. Offline
Joined: 11/10/2007
Groups: Window Group
a potential group fix we are considering

Thank you all for your feedback.   we've been thinking about this for a while.  It appears that there are two problems with group sharing:  non-group members get a massive rank-down by virtue of many group members uploading a good score, the contributor of best score gets equalized to his group members and gets discouraged from further sharing.

As some of you point out, our goal is both to provide the engaging game play experience, and to eventually enable everyone to advance science.  Groups serve a purpose for both goals. We suspect that there a number of skills involved in the protein folding process, and groups would enable us to discover much better solutions by teaming up folks that are very good at finding initial good scores, and folks that have a knack for improving such solutions further.  This is why the groups are important.  Keeping the group and individual competition separate is not only logistically complicated with respect to the game, but more importantly it will tend to split the game community into two parts, something we would really like to avoid.    So, we really need a better game structure.

Here is the mechanism that we are currently considering:

  • Every time a player A saves a solution to share with others, together with the protein we also keep the highest current score of player A, say X.
  • A member of A's group, player B can download the solution, but will only be able to officially improve the individual score if they manage to attain a score greater than X by some amount. that is you to claim this solution as your own only if you notably improve it.
  • player B's own solution (that was created from scratch) will of course have update the individual score any time it is improved.  So player B can eventually chose to go back to their own solution and work on it instead.

This framework encourages both individual and group game play -- it allows everyone to take part in both improving other people's scores as well as working on your own .  We believe these things are important to allow the game to produce best possible solutions.  the massive collective rank-ups will be very unlikely (after all, getting extra points on a very good solution is very hard), and group members will not shy away from sharing their solutions (they will effectively issue a challenge: "here's my solution, try to best it if you can and you will be rewarded!").

we are also considering a scoreboard that can show best overall solutions, best solutions from scratch, and best shared solution refiners.  So if you're only interested in individual from scratch play, just set your scoreboard to only show "from scratch" scores.  At the end of each puzzle we will announce winners in each category, as well as the path of shared solution that lead to the best score.  Sort of like Tour d'France race, where we have multiple stages, and  different jerseys for best climbers, best sprinters etc.

Anyone see significant problems with this approach?

Diderot's picture
User offline. Last seen 7 years 2 days ago. Offline
Joined: 05/09/2008
Yes!

This sounds like a great solution. (Take this with a grain of salt, however. I'm the bonehead who thought group sharing would work out just fine, too...)

Joined: 05/10/2008
Groups: None
Zoran - possible score 'fix'

Zoran - as all the folks who know me in chat can testify, I readily (and frequently) admit that I am not computer savvy in any way and do not understand any of the intricacies of the foldit program. So... I do not know enough to say, "Yes, you're idea sounds like the perfect solution," OR "No, that's not exactly what I was looking for." (I must admit, I am sitting here laughing aloud at how incredibly dimwitted I must sound to all of you.)

Although I do not feel qualified to comment on the technical aspect of your note, I do want to say this: I appreciate - more than I can tell you - y'all's willingness to listen to the concerns that have been expressed on both sides of this issue and the work you have done to address them. My feelings about group folding are absolutely in agreement with what you stated, i.e. this project lends itself perfectly to group research and results and group play should be strongly encouraged. Group solution-sharing is an *essential* tool for that process. Additonally, individual players have made large and solid contributions to this process that were both unexpected and important to your stated goals. Again, the best of both worlds!

If your suggestions make the playing field fair for all - both groups and individuals - then I applaud them whole-heartedly. If it turns out that it's still not "quite on the mark", then I want to state very clearly that I fully believe we are now moving in a direction that will bring both sides of this issue to a common ground from which we can all continue to move forward on this incredible project.

To all of the developers, programmers, administrators, and various other foldit "magical beings"(-g-), thank y'all for all of your hard work on our behalf and thank you for showing that you are, in fact, listening. For all of the players who have put a lot of thought into the problem and written in with suggestions on how to solve it, thank you for your continued attitudes of community cooperation and support. This is how problems, differing opinions, varied solutions, and compromises are supposed to work. It is very refreshing and speaks well of all who are involved in this process (and bodes well for tremendous success of the project!).

Thank you. Happy Folding!

champuli

btbenj's picture
User offline. Last seen 6 years 51 weeks ago. Offline
Joined: 11/17/2007
Groups: None
Well champuli, I realized

Well champuli, I realized you had already said what I was saying, but I thought that the solution was really a lot more simple then all these people are talking about, and I was just trying to tell people that your solution was very simple, and could be easily put. The only thing I feel we should be talking about on here about that fix, is the technicalities of scoring (like I mentioned in my previous post, whether or not individual team members get credited with their discoveries for the team).

btbenj's picture
User offline. Last seen 6 years 51 weeks ago. Offline
Joined: 11/17/2007
Groups: None
And Zoran, your fix sounds

And Zoran, your fix sounds great.

Sitemap

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