10 replies [Last post]
Joined: 05/03/2009
Groups: Contenders

OK, many of you may not have seen the puzzle comments on 210, especially now it's dead, but maybe it's time to have a public free for all on the great Helix-Chase. Given the fact I was principally responsible for killing 210, I thought I'd go first.

210 was a broken puzzle, it should never have been released in that state, and it should have been pulled. That's principally why I kept pushing the score, eventually it became an exercise in discovering how many residues would kill my komp. BUT - I was not chasing points. I used my throwaway account CharlieFC, which I use for Chatzilla, and getting into newbie puzzles when group help is required. I also chased the helix in evolver mode, minimising any damage to other player's points. I am not a helix-chaser. I've done it twice, once to see what it felt like and again on 210 to get it shut down.

Helix-chasing is not new, but it's reappeared since the advent of these new mutatable freestyle puzzles. It's entirely a personal choice, but you can take the so-called moral high-ground and wearily get frustrated as you see opportunists flying past you, or you can give in to temptation and end up feeling ever so slightly soiled in your soul. Regardless, we all play our game in our own way, and that's why FoldClub works, we're all slightly different and we all bring variety to the table. I personally dont chase the helix, I dont see the point in skewing a puzzle, and I dont play FoldClub for the points and the rank, I play because it's absorbing, addicting and has a genuine relevance to the real world.

However, FoldCentral needs to step up and admit they got 210 spectacularly wrong instead of handing out a few sweeties. Changing the rules is a very unwise move, as slippery slopes tend to be steep as well. We all know what works and what doesnt, and we all had to learn this for ourselves, sometimes the hard way. Losing hours of work and 20 ranks because your score hasnt registered with the server, but in-game your scoreboard seems fine, then posting in feedback and not even receiving a reply, happens eventually to us all. You chalk it down to experience and carry on. Helix-chasing can be effectively controlled by building in constraints and limits. And if players are going to open freestyle puzzles, they should do so with open eyes, and remember that all helix combinations will probably score best. And dont forget Freud, sometimes a cigar is just a cigar, and a helix is the answer. Just probably not one that's 752 residues long...
Anyhow, I'm off for a beer.
Feel free to throw in your 10 cents worth.
CFC

Joined: 04/13/2009
Groups: Contenders
some thoughts

CFC, we've chatted about the issues surrounding helix-chasing (and 210 in particular) already, but thanks for writing out your thoughts in such a coherent way.

I'm fairly new to the whole process of helix chasing (didn't even know how to rebuild with helixes for the longest time!) so FWIW, here are my thoughts:

First, I'm not entirely clear to what degree the process of building and perfecting these elaborate helices is representative of a real protein folding technique OR is an artefact of the game. In other words, are elaborate helical structures "well folded" in "real life"? Would they be good solutions to a protein folding problem in a lab? Or ... is there something artificial about how the game counts points that is inflating the "well foldedness" or "good solution-ness" of these structures in the game? If the former, then it seems that even if helix-chasing is boring and repetitive, it should be a valid way to fold that is well-rewarded (for those players who enjoy it, of course). If the latter, then there's no two ways about it - helix-chasing is a game exploit and something needs to be changed in the game mechanics to address it. Not just for fairness, but for the value of our folding to science, since we are generating invalid (or non-optimal) solutions.

Second, I know sometimes you're dissatisfied at how the devs/admins handle issues like this. I guess I just want to point out that this is a beta program that is produced and maintained by a university research group. I don't know for sure, but it seems likely to me that there are a couple of poor overworked grad students at the core of the maintenance (or what usually comes down to damage-control) process. So personally, I don't really expect that every single time there is a screw-up (which ... hey, it happens) they are going to spend a lot of time making amends and telling everyone how sorry they are and correcting all the repercussions. I think they just make a note of ... "oops! Shouldn't have done that!" and then move on. As a side note, it may be that they don't have a set-up by which they can easily go in and correct scores (for example, remove all scores associated with puzzle 210 and adjust everyone's global score). They should be doing this, but it might be too much of a PITA for them. Another side note -- would there be MORE complaints if they removed scores ("admins, why did my score go down? Boo hoo ...").

Third, I think that part of what is going on with this problem and other issues that folks have with the game is that sometimes we want to have more control over the process. We love folding and we want to live, breathe, and eat it (or I do :) ). We want to be able to make core decisions about the game development and mechanics. We're invested. I know lots of folks have repeatedly asked why the project is not open-source. I don't really think they are going to go this direction, but I think we can and should talk about enhancing communication between players and devs, increasing information openness associated with the project, and setting up structures by which players get to enact control over limited game-related domains. I typed out that sentence and realized I really don't know what I mean by that ... but it's an interesting thought -- short of foldit being open-sourced, how can ordinary people out there (like yours truly :) ) participate in the development and evolution of the game in some way?

(I have a feeling that you're going to say that we're all pushing on our end and the devs are doing little on their end -- i.e., not spending enough time reading forums, responding to feedback, etc. This is definitely a significant issue and I'm not sure what we can do about it. I have a feeling it's mostly related to limited resources (resources meaning overworked grad students ;) ) rather than some kind of core disrespectfulness toward players or intention to shut us out of the process.)

And that's my rambly .02.

Joined: 04/13/2009
Groups: Contenders
Another note-- Just read

Another note--

Just read the actual debate on the page for 210 (don't know how I missed that).

I do have a thought about the particular problem that happened there (and may have happened in the past?)

It seems that a player (Milward) noticed there was a problem with the puzzle within a day or two of its release and posted a comment. An admin (AustinDay) responded a day later and said they were going to "swap it out". Then, as far as I can tell, it was not swapped out. No idea why.

IF as soon as players and admins realized and came to an agreement that a puzzle had a problem, they IMMEDIATELY removed the puzzle, everyone would be happy. OR ... alternatively, a puzzle could be kept up but clearly MARKED in some way so that players know that they can work on it if they want, but it will not effect their global scores (or that its effect on global scores is TBD by admins) -- Players would in essence be freely choosing whether to fold a "broken" puzzle that may not be scored.

This strategy would provide *fair warning* to players and be relatively simple to implement. It would be much simpler than taking the time and going through solutions and figuring out which ones properly adhere to "the spirit of the game" (!).

Joined: 05/03/2009
Groups: Contenders
Almost true...

infjamc actually posted a feedback first, and I picked up on it and pushed the score and it started cropping up in global chat. This situation also occurred on 174, and it got pulled early, but if memory serves, the top 20/25 were all Chasers, and scores stood. It then was re-released in a fixed state. And it IS possible to close puzzles without them scoring, it happened in the recent past when I think 192? appeared for about 3 hours. If a puzzle is obviously flawed, no-one's gonna waste their time if they can be certain that it will be pulled without scoring. Problem solved.

And your earlier post, there are very good reasons why helices are so good at compacting space in secondary structure. DNA is the obvious example, a double helix, that forms super-coils of coils of coils. And like I said earlier, sometimes the answer is a helix, plain and simple.

infjamc's picture
User offline. Last seen 2 years 49 weeks ago. Offline
Joined: 02/20/2009
Groups: Contenders
More on helices...

My guess is that helices are an efficient way to get points in freestyle design puzzles because you can get as many hydrogen bonds as you want. Alpha helices, after all, hydrogen-bond between the (i)th and (i+4)th residue, so you get on average 2 hydrogen bonds per residue except for prolines and terminal residues of the helix. Add that to the fact that you can get even more points from favorable side chain interactions, and it becomes obvious that it can be a dominant strategy in certain cases.

Take Freestyle Design 60 (#185), for example. My design actually had very low complexity-- it's a single giant helix that includes only glutamate and arginine, with EEEERRRR repeats in the middle. Why use these? Because they're large hydrophilics, plus that you can get two hydrogen per glutamate-arginine pair bonds and favorable electrostatic interactions if you put them 4 residues apart and that the side chain configurations are correct. And that was enough for 10045 points (i.e. +34.08 per residue in addition to the 8000 that you get for free) and 3rd place in the soloist competition.

My past experience also shows, however, that EEEERRRR repeats may not be the optimal strategy in the "helix craze" category. A two-helix design with a short loop in the middle, with WWEEWWR repeats for both helices, seems to work better for larger proteins, if only because the packing scores for tryptophan can exceed 100 in certain cases.

infjamc's picture
User offline. Last seen 2 years 49 weeks ago. Offline
Joined: 02/20/2009
Groups: Contenders
Correction...

The previous post should read "1 hydrogen bond per residue" instead of 2 due to double counting.

And while we're still on the topic of helices... technically, the hydrogen bonds in helices are (on average) weaker than those of beta sheets. So why aren't we seeing designs comprised of mostly beta sheets that are competitive in terms of FoldIt score?

Mark-'s picture
User offline. Last seen 38 weeks 13 hours ago. Offline
Joined: 08/05/2009
Groups: Contenders
My original post on the

My original post on the 'comments' was only to find out if 210 was a throw-away puzzle or not. Whether I posted too late or Admin just didn't see it in time I don't know. I think, as has already been said, that if we had known it was a 'throw-away' with no points to be gained either way then I may also have decided to see how big a helix I could build, or I may have just carried on for the fun of it. Who knows?

The fact that it was not immediately pulled, and it was awarded points, seemed to me blatantly unfair to the majority of competitors, so I voiced my disappointment. It was Admin who suggested that they could award points depending on structure. That is up to Admin to decide, not us. We can only voice our opinions and hope that they agree with us.

It was never my intention to cause such problems or to create extra work for Admin. For that I apologise. What I will not apologise for is highlighting errors or unfairness, as I see it, in the puzzles.

I think it's certain that the next design puzzles will have more constraints attached, so I'm sure some good will come of this eventually.

Take the scores away or leave them as they are. It's history now, and I'm already too involved with the next puzzles to worry about it anymore.

Mark

Joined: 05/03/2009
Groups: Contenders
Thanks Mark

Cheers Mark, thanks for the input. I hope more Folders will chip in and express an opinion. The simplest solution is already becoming clear, if a puzzle is flawed, let it be known it'll be void and unscored, and no-one needs waste their time.
CFC

Joined: 10/10/2008
Groups: None
Hi All, Again, sorry for

Hi All,

Again, sorry for the trouble the puzzle has caused. At least I won't be making that mistake again. :) In terms of the scoring though, it seems that it would be much more time consuming to pick and choose the solutions to give points to than I anticipated. I appreciate the points made about how we should look to the future and in light of that, I've decided to leave the point distribution as is. That would at least be consistent with the previous puzzles which had this kind of error. But I'll use that saved time to bring you all more exciting puzzles! Just you wait!

-Austin

infjamc's picture
User offline. Last seen 2 years 49 weeks ago. Offline
Joined: 02/20/2009
Groups: Contenders
On second thought...

There is actually a way to avoid the helix craze:

In puzzles where one cannot add or remove residues, a possibility is to use a shaken-and-wiggled giant helix as a constraint so that people cannot get credit if their solution is too close to that.

Susanne's picture
User offline. Last seen 4 years 24 weeks ago. Offline
Joined: 11/09/2008
Groups: SETI.Germany
Helix Craze/#210

Usually, after a few futile attempts at folding large freestyle puzzle into sheets that actually want to stay adjacent and produce hydrogen bonds, out of sheer frustration with my efforts, I resort to an assortment of helices to compact the puzzle quickly and begin to score points. I have no idea what goes on chemically and cannot plan ahead other than for compactness/shape. If there would be a tutorial or detailed description regarding successful sheet folding in freestyle puzzles, I would be very keen to learn more as I much prefer to work on sheet/helix combis.

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