Cloned accounts

It has come to our attention that some Foldit players are creating multiple accounts and using them competitively in puzzles. To clarify, each player should have only one competitive account. Multiple accounts that are used for chatting are okay. We will soon begin taking action to prevent cloned accounts from competing.

EDIT: This is the result of the recent community chat discussion about clones (http://fold.it/portal/node/989454).

(Thu, 05/05/2011 - 18:33  |  30 comments)
Joined: 01/19/2011
Finally!

Good, there are certain people that are blatantly cloning and they are just making it more difficult for me to gain a rank...

Hanto's picture
User offline. Last seen 1 day 7 hours ago. Offline
Joined: 05/10/2008
Groups: None
I only want to know the real

I only want to know the real name of the individual creating this thread, his/her position within the fold.it organization and a full financial disclosure statement indicating names and addresses of supporting organizations or governments. It seems odd that no one authored this thread, and it even looks like it might be just a bit threatening in nature. I want those disclosures made to me before any action is taken by Fold.it. I have a very strong feeling that many supporting organizations and government entities will find fault in this type of effort to intimidate.
On a personal level I only feel shame for those more worried about rankings then about curing sickness and furthering scientific knowledge.

Joined: 02/23/2011
Really, sir?

Besides knowing we are helping advance science, points earned and higher rank are our rewards for playing the game well. When certain people compete with more than one account and crowd out the upper ranks, it discourages other players from playing. I think such oppression would be much more interesting to any concerned 'supporting organizations and government entities'.

I, for one, don't find this thread the least bit threatening, but maybe that is because I don't have ~6 competing clone accounts.

xedrIRC's picture
User offline. Last seen 6 years 29 weeks ago. Offline
Joined: 04/13/2011
Thanks Hanto

Thanks Hanto for sharing your desires and feelings. Everybody is entitled to an opinion. But the proposed rule changes are a good idea. The score-skewing effect created by account cloning creates a real discincentive for other players.

In any event, the fold.it organization is free to set their own rules without having to account to you or any of the participants in this game.

Hanto's picture
User offline. Last seen 1 day 7 hours ago. Offline
Joined: 05/10/2008
Groups: None
Only 4 accounts

I have only 4 accounts of which only two are used at any one time, one reflecting one core on my AMD platform and one reflecting one core on my Intel platform. However at least two extra Hanto_???? accounts were created by someone as a prop to invade Go Science with. I have suspicions of who that was, but I'm sure that the Admin folks here could verify it wasn't me.
The only thing you guys are trying to do is intimidate especially since a numerically much weaker team is giving you a much tougher row to hoe.

Joined: 08/29/2010
And now where is your problem Hanto?

If you have not that much accounts the ips will show this for sure so i don't get it why you complain about this thing...

I also think for everyone one account is enough for competion and if multiaccounts aren't allowed for chatting the method of chatting must get redesigned within fold.it

Joined: 11/10/2007
Groups: Window Group
Right, clones were discussed

Right, clones were discussed at the recent developer chat, and the disruption to the ranking caused by them. Through the discussion with the community we decided that multiple cloned accounts should not be competing. This post is meant to clarify the position we are taking on cloned accounts.

B_2's picture
User offline. Last seen 4 years 32 weeks ago. Offline
Joined: 11/29/2008
Groups: None
How?

From a purely technical point, how exactly will you determine cloned accounts?

Can't use email .. every account already has it's own email.

Can't use IP address ..

Where I live, there are 4 people, plus two neighbors using a shared wireless network, on 5 PCs. Two or more people could share each PC.
They ALL appear to the outside world as being one IP address. It's the nature of modern networking.

Where I work, at a large educational entity, there are between 8000 and 10,000 PCs that all connect to the internet, and EVERY SINGLE ONE of them looks like it has the same IP address to the outside world. PC labs could have an entire class sharing random PCs - using internal IDs from the PC won't work either.

Banning an IP address could shut off an entire class of biology students who have been assigned here for homework.
(Well, there are a couple of classes from certain school districts who SHOULD be banned, and BTW, all those obnoxious students had the same IP address. I keep logs too.)

So, without having to provide a passport or SSN or some other personal identifier to compete, how do you propose to control this?

BOINC went through this same exercise years ago, and gave it up as impossible to police.

Banning similar sounding named accounts might work in the short term, but if people want to do it bad enough, they will find other ways. Anonymous proxies are always good.

And in the end, you create chaos, hate, (more) animosity between teams, and it will all be wated effort that could be used elsewhere on a project that certainly can use all the resources it can get.

Just thinking out loud...

B_2's picture
User offline. Last seen 4 years 32 weeks ago. Offline
Joined: 11/29/2008
Groups: None
I forgot to mention -

In this feedback, there is discussion on WHY duplicate accounts are needed. http://fold.it/portal/node/989405

Until the Restore Absolute Best is fixed, it's impossible to explore more than one solution to the same puzzle without doing manual saves every couple of seconds, and that quickly become impossible to maintain. All it takes is one recipe doing a "Restore Best" and hours or days of work can get overwritten by the high score from another path.

I hear the developers complaining about not enough exploring, and now they want to inhibit it further? Doesn't make sense.

B_2's picture
User offline. Last seen 4 years 32 weeks ago. Offline
Joined: 11/29/2008
Groups: None
It's clearly obvious from

It's clearly obvious from the people contributing here that it's Anthropic Dreams beating up on the tiny Go Science team, in retribution for some of the AD members leaving and joining Go Science.

They couldn't stop their precious recipes leaving with the users, so they have invented a new issue as a means to annoy Go Science, and continue trying to be the 800 pound gorilla team who HAS to get what they want in their childish way.

Every wonder why they left in the first place?

Joined: 08/24/2010
Hi

Attempting to relocate this to a new forum thread - http://fold.it/portal/node/989668

infjamc's picture
User offline. Last seen 3 years 1 day ago. Offline
Joined: 02/20/2009
Groups: Contenders
.

Re: vixynn

While I do not like the idea of cloned accounts either, I can see why some players find them necessary. As I mentioned in http://fold.it/portal/node/989405 , workarounds do exist-- but they have to be done very carefully and manually. The best one can do in a recipe/script is to use the quick-saves, and even then there's always the chance of wiping out an alternative trajectory via the restore very/recent best function.

This is exactly why I suggested a month ago that there should be an option of designating accounts as "secondary," non-scoring accounts that are linked to a main account (see http://fold.it/portal/node/989493 ), as this would allow cloned accounts to exist for ease of exploring multiple solutions without cluttering the scoreboard.

infjamc's picture
User offline. Last seen 3 years 1 day ago. Offline
Joined: 02/20/2009
Groups: Contenders
Clarification (WARNING: LONG POST)

The bottom line is that a distinction needs to be made between the following:

1. Running multiple accounts for the purpose of abusing the system, namely bumping other players down the rankings. I think most people would agree that this should be banned.
.

2. Running multiple accounts for legitimate purposes, such as exploring multiple solutions. Currently, this comes with the same problem as #1 (cluttering the scoreboard), which is why I suggested the secondary account idea as a stop-gap measure. But I still believe that the best long-term solution is to *remove the need for multiple accounts* in the first place by revamping the way that solutions are handled. One possibility is the following:

* At the start of each puzzle, one starts working with trajectory #1 by default.

* At any time, the player may choose to create additional trajectories. If this function is invoked, a copy of the current structure is made, and it becomes the recent/very best solution for the new trajectory. From now on, "recent/very best" refers to the high-scoring solutions of the new trajectory.

* At any time, the player may switch between the existing trajectories to work on for the purpose of defining which structures "recent/very best" refer to.
.

========================================
The actual use of this new system might look like the following:

1. User starts a puzzle for the first time. The starting structure has a score of 8000, which is automatically copied into the following autosaves:

Trajectory 001:
* "Autosaved Very Best Soloist Solution-- Trajectory 001" = 8000 (special condition not met)
* "Autosaved Recent Best Soloist Solution-- Trajectory 001" = 8000 (special condition not met)
* "Autosaved Credit Best Soloist Solution-- Trajectory 001" = 8000 (special condition not met)

[Current Trajectory = 001]
.

2. The user improves the solution to 9500. Unfortunately, the condition required for the score to count was not met by the time that the wiggle was completed, so the score that got recorded on the scoreboard was 9400, an intermediate structure. A rebuild was attempted on the 9500 solution, which resulted in a structure that scored 9450 that still fails the condition. The 9450 solution is then manually saved as a recent best. The three autosaves now contain:

Trajectory 001:
* "Autosaved Very Best Soloist Solution-- Trajectory 001" = 9500 (special condition not met)
* "Autosaved Recent Best Soloist Solution-- Trajectory 001" = 9450 (special condition not met)
* "Autosaved Credit Best Soloist Solution-- Trajectory 001" = 9400 [on the scoreboard]

[Current Trajectory = 001]
.

3. The user decides to start over by resetting the puzzle, followed by adding a second trajectory. The following autosaves are created:

Trajectory 002:
* "Autosaved Very Best Soloist Solution-- Trajectory 002" = 8000 (special condition not met)
* "Autosaved Recent Best Soloist Solution-- Trajectory 002" = 8000 (special condition not met)
* "Autosaved Credit Best Soloist Solution-- Trajectory 002" = 8000 (special condition not met)

[Current Trajectory = 002]
.

4. On the second attempt, the solution is improved to a score of 9480. Unfortunately, the condition that had to be met for the score to count was broken by the final wiggle, so the structure that received the credit is an intermediate that scores 9420. A rebuild was once again attempted, and the result was a solution that scored 9460 that still failed the condition. The user decides to save it as a recent best anyway, and the following autosaves now exist:

Trajectory 001:
* "Autosaved Very Best Soloist Solution-- Trajectory 001" = 9500 (special condition not met)
* "Autosaved Recent Best Soloist Solution-- Trajectory 001" = 9450 (special condition not met)
* "Autosaved Credit Best Soloist Solution-- Trajectory 001" = 9400

Trajectory 002:
* "Autosaved Very Best Soloist Solution-- Trajectory 002" = 9480 (special condition not met)
* "Autosaved Recent Best Soloist Solution-- Trajectory 002" = 9460 (special condition not met)
* "Autosaved Credit Best Soloist Solution-- Trajectory 002" = 9420 [on the scoreboard]

[Current Trajectory = 002]
.

5. At this time, "reset very best" will load the 9480 solution and overwrite the 9460, while "reset recent best" will load the 9460. If the user decides to designate #1 as the "current trajectory," then "reset very best" will load the 9500 and solution and overwrite the 9450, while "reset recent best" will load 9450.

========================================

==> The above proposed system comes with the advantage that all trajectories stay with one account. This is a huge advantage over using multiple accounts to work with alternative solutions, as the structures from the different trajectories would stay with the same account. The main difficulty I see is the sheer extent of programming required to make this system work. After all, for those who want to work with multiple trajectories simultaneously using a multi-core computer, there needs to be a function for saving/loading solutions associated with a specific trajectory rather than just the current one. Practically speaking, I don't see the fully-automatic multi-core processing of multiple trajectories (one per core) commonly done in Rosetta 3.x becoming reproducible in Foldit any time soon.
.

Another problem is that the sheer number of autosaves involved due to having a set of autosaves for every trajectory may make it difficult for the user to find a specific solution. One possible improvement might be redesigning the interface for loading/saving solutions by grouping the solutions by trajectory, such as the following:

[Local Solutions]
+ Trajectory 001
--- "Autosaved Very Best Soloist Solution-- Trajectory 002" / Score: 9500
--- "Autosaved Recent Best Soloist Solution-- Trajectory 002" / Score: 9450
--- "Autosaved Credit Best Soloist Solution-- Trajectory 002" / Score: 9400

+ Trajectory 002 <=== Current trajectory
--- "Autosaved Very Best Soloist Solution-- Trajectory 002" / Score: 9480
--- "Autosaved Recent Best Soloist Solution-- Trajectory 002" / Score: 9460
--- "Autosaved Credit Best Soloist Solution-- Trajectory 002" / Score: 9420

+ Manual Saves
--- T001-S9500 rebuild / Score: 9387

B_2's picture
User offline. Last seen 4 years 32 weeks ago. Offline
Joined: 11/29/2008
Groups: None
Let's fix the problem It's

Let's fix the problem

It's obvious there are valid reasons for using multiple accounts. infjamc and I have both pointed out weaknesses in teh client that require separate accounts, and Hanto has reasons for using separate accounts to be able to remotely monitor the activity on his multiple cores.

Instead of banning the use of multiple accounts which will cause a lot of bad feelings, please fix the client.

And , I believe it was AD who brought up this newly invented issue on the developer chat - so yes, AD started this thing.

Why would they kick out Hanto for doing something that wasn't against the "rules" of Fold.it. For that matter, what ARE the rules of Fold.it ?

Joined: 02/25/2011
Groups: None
I agree with B_2 that we should fix the problem first

and I think that infjamc's suggestion would solve everything

Players could have multiple accounts (which they all want for different reasons) but only their top scoring one would post on the scoreboard so new Foldit players do not become frustrated as easily
Foldit is frustrating enough already for new players

I don't think it matters who brought this up or who noticed this problem first, it might have been AD in chat but the feedback posting about it that B_2 referred to was created by CFC from Contenders fold.it/portal/node/989405

I don't think that pointing fingers or asking for someone's credentials or Long Form Birth Certificate will solve the problem ;)

infjamc's picture
User offline. Last seen 3 years 1 day ago. Offline
Joined: 02/20/2009
Groups: Contenders
Re: B_2

To be absolutely clear, my attitude toward the usage of multiple accounts is "I can tolerate the practice but would not encourage it myself." (Full disclosure: I personally have no "need" for multiple accounts since I only run Foldit on my laptop, whose CPU only has 2 cores.) Plus, the reality is that fixing the system either through the designation of secondary accounts or through my above-mentioned "work with multiple trajectories" option will require extensive programming. Practically speaking, this means that those who use multiple accounts, *even for legitimate reasons*, will have an unfair advantage over other players for the foreseeable future even though they might never intend to cheat the system.

==> So, for the short term, I would recommend manually implementing the concept of "recent best" and "very best" in your recipes/scripts. It's cumbersome, but at least it works. The following is an example:

====================
1. Suppose that this is the pseudo-code for your current script (I'm using Blue Fuse as an example here):

Set recent best
Process structure
Reset to recent best
Process structure further
Reset to recent best
.

This can be re-written by first re-inventing the wheel, namely re-creating the function of recent bests and very bests using quicksave slots.

--------------------
Function initialize():

Set traj to 1 //Trajectory number, can be 1, 2, or 3
Set s1 to (traj-1)*3+1
//Determines the quicksave slot to be used for the current solution; this will translate to 1, 4, or 7
Set s2 to (traj-1)*3+2
//Determines the quicksave slot to be used for the recent best solution; this will translate to 2, 5, or 8
Set s3 to (traj-1)*3+3
//Determines the quicksave slot to be used for the very best solution; this will translate to 3, 6, or 9

Quicksave to slot s1 //Further processing will be done on this structure
Quicksave to slot s2 //Set recent best
Quicksave to slot s3 //Simulates the "very best" solution
Score structure, then store the result as current_score, score_recentbest, and score_verybest
.
.

Function set_recentbest():
Quickload slot s1 //Current structure
Score structure and store the result as current_score
Quicksave to slot s2 //Overwrite recent best
Set score_recentbest to current_score //Update score of recent best
.
.

Function load_recentbest():
Quickload slot s2 //Recent best slot
Score structure and store the result as score_recentbest
Quicksave to slot s1 //Overwrite current structure
Set current_score to score_recentbest //Update score
.
.

Function load_verybest():
Quickload slot s3 //Recent very best slot
Score structure and store the result as score_verybest
Quicksave to slot s2 //Overwrite recent best
Quicksave to slot s1 //Overwrite current structure
Set score_recentbest to score_verybest
Set current_score to score_verybest
.
.

Function update_recent_and_verybest():
Quickload slot s2 //Recent best slot
Score structure and store the result as score_recentbest
Quickload slot s1 //Current structure
Score structure and store the result as current_score

If current_score > score_recentbest:
---- Quicksave to slot s2, and set score_recentbest to current_score
---- //Update the recent best if the current structure scores better

Quickload slot s3 //Very best slot
Score structure and store the result as score_verybest
Quickload slot s1 //Current structure

If current_score > score_recentbest:
---- Quicksave to slot s3, and set score_very to current_score
---- //Update the very best if the current structure scores better
// Note that this ends up with the current structure being loaded last whatever happens.
// This is intentional, to ensure that you're working on the current structure.
.
.
.

--------------------
While cumbersome, this is a one-time investment; for future purposes, one only needs to copy and paste the above functions.
The main part of the program is actually very short:

set_recentbest()

process_structure()
update_recent_and_verybest() //Insert one instance of this function every time before saving or loading the recent best or very best

load_recentbest()

process_structure_further()
update_recent_and_verybest()

load_recentbest()

--------------------

2. The bottom line is that you need one quicksave slot for storing the current structure, one for the recent best, and one more for the very best. Running multiple instances of the script only requires changing the value used for "traj" (1, 2, or 3).

infjamc's picture
User offline. Last seen 3 years 1 day ago. Offline
Joined: 02/20/2009
Groups: Contenders
Correction to previous post

There is an error in one of the functions. The correct version is given below:

====================

Function update_recent_and_verybest():
Quickload slot s2 //Recent best slot
Score structure and store the result as score_recentbest
Quickload slot s1 //Current structure
Score structure and store the result as current_score

If current_score > score_recentbest:
---- Quicksave to slot s2, and set score_recentbest to current_score
---- //Update the recent best if the current structure scores better

Quickload slot s3 //Very best slot
Score structure and store the result as score_verybest
Quickload slot s1 //Current structure

If current_score > score_verybest:
---- Quicksave to slot s3, and set score_verybest to current_score
---- //Update the very best if the current structure scores better
// Note that this ends up with the current structure being loaded last whatever happens.
// This is intentional, to ensure that you're working on the current structure.

infjamc's picture
User offline. Last seen 3 years 1 day ago. Offline
Joined: 02/20/2009
Groups: Contenders
Addendum (sorry for triple-posting)

For those of you who want to use the pseudocode I've posted above, please note the following:

Before calling any function that deals with the quicksave-simulated versions of recent bests and very bests, *make sure that you have saved your current structure by adding the line "Quicksave to slot s1." I didn't add that to the functions listed above based on the assumption that this will be done in the "process structure" step, but you might want to consider adding that line (even if this creates redundancy) to prevent over-writing your current solution by mistake. Again, debug thoroughly...

jflat06's picture
User offline. Last seen 1 day 3 hours ago. Offline
Joined: 09/29/2010
Groups: Window Group
.

A very simple option would be a button in the advanced UI which lets you completely reset a puzzle (including absolute best, quicksaves, undo timeline). It would effectively let you start the puzzle over from scratch. The only things that would remain would be your score on the scoreboard, and any manual saves you had done.

Is this a desirable option?

infjamc's picture
User offline. Last seen 3 years 1 day ago. Offline
Joined: 02/20/2009
Groups: Contenders
Re: jflat06

This would be a good idea. However, I think some players might want to keep a backup copy of the autosaves (just in case) as the puzzle is completely reset-- a task that currently must be done manually. Would that involve too much extra programming?

Joined: 05/03/2009
Groups: Contenders
Step in the right direction

Beta's uber-reset button will solve short-term requirements, but not everyone runs the advanced gui.

Big thanks should go to jflat for coming onto #global and chatting through some of the problems that currently exist.

I also believe that massively increasing the number of save slots available, to 80, or even 100 would help. Since the scripting 'status quo' is currently limited to 10 slots, having 100 available would allow you to run 10 clients simultaneously, with the same solo position, using the same script. You WOULD have to have 2nd, 3rd, 4th etc versions of each script to avoid overwriting slots, but it would give a little more freedom to those with the hardware to exploit true exploring.

It would also be a step closer to truly emulating what we do with hand-rebuilding. With high-end hardware being more and more common-place, the dev team have a chance to be pro-active, rather than reactive. 100 slots. Job done.
CFC

wld333's picture
User offline. Last seen 7 years 18 weeks ago. Offline
Joined: 02/15/2011
Groups: foldeRNA
Parallel vs. serial

I agree with CFC.

Some people are presumably using cloned accounts just so they can do parallel exploration of several alternate solutions simultaneously. These people may in fact have no desire to hog the rankings -- it's just a side-effect of the way they explore.

Allowing parallel (and simultaneous) exploration within a single account for those whose hardware supports it seems like a good solution.

If that's in place, along with a solution to whatever it is people want to do in chat, then it seems like cloned accounts could be eliminated without anyone having a legitimate reason to gripe about it.

B_2's picture
User offline. Last seen 4 years 32 weeks ago. Offline
Joined: 11/29/2008
Groups: None
That's a nice feature, if it

That's a nice feature, if it doesn't slow the client down even more by continually saving 100 slots.

Unfortunately, it still requires re-writing all the scripts in use to take advantage of it.

What I envisioned is a hotkey that will save and/or restore the Absolute Best for the solution I'm working on, so it needs to be related to the solution, not the puzzle. Right now, it's associated with the user/puzzle. If I save something as "Solution A" it needs it's own Absolute Best, and if I save a derivative as "Solution B" and run a recipe against it that uses "Restore Absolute Best" it doesn't restore the Best from "Solution A", it only restores the Absolute Best from "Solution B".

And, I don't always want to start over from the beginning, so the uber-reset isn't an option either.

It's really a branching thing - partway through a puzzle, I want to explore more than one evolutionary branch, and do it simultaneously. Right now, it requires separate accounts to do that.

B_2's picture
User offline. Last seen 4 years 32 weeks ago. Offline
Joined: 11/29/2008
Groups: None
Did this thread get shorter?

The front page says 20 comments, but the top of the thread says 24 comments.

Has there been some "editing" going on?

infjamc's picture
User offline. Last seen 3 years 1 day ago. Offline
Joined: 02/20/2009
Groups: Contenders
Re: B_2

"Did this thread get shorter?"

No, it's just that the post count takes some time to update itself.

B_2's picture
User offline. Last seen 4 years 32 weeks ago. Offline
Joined: 11/29/2008
Groups: None
Actually, I believe there

Actually, I believe there were comments from 'vixynn' which are longer visible.

beta_helix's picture
User offline. Last seen 1 day 2 hours ago. Offline
Joined: 05/09/2008
Groups: None
that is correct B_2

it seems that Foldit user vixynn has left us:
http://fold.it/portal/user/227611
and selected "Delete the account and all content" before leaving.

Thank you all for your feedback and suggestions on the parallel exploration issue, our goal is to implement a satisfactory solution soon.

jflat06's picture
User offline. Last seen 1 day 3 hours ago. Offline
Joined: 09/29/2010
Groups: Window Group
.

This is actually very similar to a solution which we will be rolling out very soon.

We've decided that a global reset would be a temporary solution at best, and it's in our best interest to solve the problem in a robust way.

It won't solve the issue with chat, but should solve the other issues.

spmm's picture
User offline. Last seen 47 weeks 6 days ago. Offline
Joined: 08/05/2010
Groups: Void Crushers
cloned accounts on 522b

still obviously being done on 522b

Get Started: Download
  Windows    OSX    Linux  
Windows
(7/8/10)
OSX
(10.7 or later)
Linux
(64-bit)

Are you new to Foldit? Click here.

Are you a student? Click here.

Are you an educator? Click here.
Search
Only search fold.it
Recommend Foldit
User login
Soloists
Evolvers
Groups
Topics
Top New Users
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