Input wanted on planned new algoritm from users perspective

[Brian the Roman]

Howard,
could you give us a quick explanation of how the structures of generation 0 are different from those of later generations. You say later gens will take longer per structure and one therefore assumes that you believe their probability of being better structures is higher. How are you doing this? Are you possibly using the best energies calcd from the previous generation and 'looking around the neighbourhood'?

ms

[reader50]

I live in California, the electricity is not cheap. And paid for out of my own pocket, not on a business ledger.

From a stats-keeping standpoint, restarting with lower totals (and ones that rise slower) would be nice. But I've pretty well covered the integer overflow problems.

My vote is to keep the stats units I've paid for, not to put them out to pasture. My total is not in the Top 10 (or anywhere near), but it still kept my bedroom noisy for weeks on end. Slowed my games down too. (serious sacrifice there )

[Grumpy]

This is the biggest problem with zeroing the stats, people who are not large folders have put in a lot of effort to get where they are. To suddenly wipe all their effort away disadvantages them a lot more than the top 100. Think of those on only one computer, a slow one, that have been in it from the start. They are about to hit 10 Million after all that effort...bang, stats zeroed. I abide by the referee, but I am in the keep the stats camp..Erm, can not you have both running at the same time.. Just have an option to choose Display Combined Stats or Display New Project Stats . Users can then follow whichever method they wish

[FBK]

8 Months ago I joined DF. Since then, I added 4 folding machines, to my existing 6 machines. My monthly electric bill jumped significantly. I monitored and tended my personal farm. I crunched, 7x24, since then, through slow genes as well as faster genes. I lost work from overloaded server(s), 904 errors, failed updates and a half dozen other reasons. I didn't expect anything in return. But, I am proud of my contribution, and that contribution is reflected only through stats.

8 months ago, many other people joined DF. Some are still folding. Some folded for a while, but have long since quit. Some bought and supported personal farms. Some put forth the gargantuan efforts to Borg work machines. Some NEVER FOLDED a SINGLE UNIT.

In a few weeks, apparently, we will all be equal on the stats.

Perhaps I'm a stats whore. Perhaps I'm not noble or righteous.

Somehow zeroing the statistics, and my contribution, just doesn't seem right.

FBK

[MAD-ness]

I am glad that some alternative viewpoints have been expressed.

I knew that some people would not be excited at the prospect of a stats zeroing but it took them a while to show up and voice an opinion.

I can go either way, neither will make or break me but for some people it is a larger issue. Understandably so for those who did a lot of work keeping networks running or footing higher energy bills to keep the home computers on 24/7 and crunching away.

[Goobee]

I would rather see the current points rolled over. My farm does not run 24/7 for free; in a manner of speaking, I paid for the points I produced. I do not participate in non-medical DC projects, only in projects that may one day lead to finding cures for illnesses/diseases.

In return for my expenses, I get stats. I like stats, many of us do. My points were difficult to come by. All of my machines are at home - if I could enlist a bunch of machines at work I would but I can't. My points are generated from electricty paid by me; if you ask around, many DC'ers are paying the electrical bills out of their pocket.

You may want to keep this in mind when you guys make your final decision.

[Grumpy]

As part of the Democratic Process at OCworkbench, I have started a poll on the issue. Members can choose to keep it the way it is, zero the stats, or have a system that has OLD+NEW stats counted OR just count the New stats ( So you would have 2 parallel ladders running and the individual or Team can decide which they prefer to follow). I will post the results when the poll closes.

Perhaps other Teams should do likewise and put their results here too

[OgreChow]

Don't you all see that once the stats are reset they will show exactly how much you are contributing RIGHT NOW?

This means that if you are upset because you are working very hard, all of a sudden you will be working very hard and everyone will know it!

As for concerns that it will stay this way, static - perhaps at first...until you add a computer to try to bypass the person ahead of you, or someone switches teams, or someone loses interest - in short, all the things that have led to stat-climbing and dropping today. It will just be a short sprinting lapse before we being the long-distance race in which we are currently engaged.

I personally think that the first few hours after the reset are going to be incredibly exciting - I know I am going to find a few more machines to crunch on and will be updating every few minutes

-OC

[PinHead]

I don't have a preference either way on the stats!

But if they are zeroed, I think the folders are looking for some type of recognition for the work that they have done. Either a static past results page or a trinket of some type by their name, to indicate their participation in phase 1. Maybe even just a number next to their name indicating phase 1 ranking.

Also, will this progie have new hardware requirements?
Higher minimum cpu or memory?

[Brian the Fist]

Well, you just gave me a somewhat wild idea there, inspired by our friends eBay. What if we zero the stats as described, but keep (internally) a record of your production in phase I. Then we internally add phase I + phase II production and call that total production.

Stats pages will show only phase II production BUT you will get a series of coloured stars or other objects next to your name based on total production (I noticed some 3rd part stats pages do something like this already). Thus you can quickly identify people who've been around for a long time and contributed a lot even though their phase 2 production may not be so much.

Lemme know, and if you have a preferred color scheme..

To Brian the Roman:

Initially, the generation N structures will be generated as 'near neighbours' to the lowest energy/RMSD structure from generation N-1, and so on. Later on this may change slightly, and generation N structures may be random combinations of fragments from generation N-1 structures.

[Faceless]

Whatever you want to do is OK by me. I don't want to be a whinner wanting the stats reset just because ExtremeDC is not on top of this project. The project is worth while and I'm planning to stick with it.

[cygnussphere]

Hum.... Me clicks on link to Dyys Top 1000 users page.

Scrolls to page 4. current position 168/1000

now clicks on handy sort by last weeks production button.

Wasza! scrolls back to pg 2 status now 76/1000.

Where do I find the reset button on this thing?


Actually, The dual recognition plan does seem like a good idea.

Fellow M.O.B. (minion of brian)

[Insidious]

First of all, I am here for the duration no matter what is done with the stats.

Is it correct that everyone here is talking about stats pages other than the one on the Distributedfolding.org site? I don't find the information everyone is referencing on that one.

(I use statsman.org and stats.zerothelement.com)

Doesn't all this discussion only apply to Dyy and whoever does Statsman?

My utopic stats page would open to dynamic stats of Phase II with a button that would link me to the (static) Phase I final results.

(I do like the star idea identifying a user who has Phase I input next to the name on the Phase II page)

-Sid

[grobinette]

This is getting better, I like the zero out the stats but with some form of recognition for phase 1 activity idea....

[Anarchy99]

leave the stats ppppllllleeeeeeeaaaaasssssseeee

"no disassemble" johnny 5

"we likes our precious stats we does" gollum/smeagol

"stats good hhmmmmm" yoda

"we like the stats the way they are pilgrim" john wayne

you could zero them but leave us a slightly transparent block with the previous effort made

"me thinks you a got it now" jar jar binks

"happy birthday mr president" marilyn monroe [but what a hottie]

some consideration to the members who cant climb as fast to show how long they have been at it, helps them justify the effort
and gives them a benchmark for the time they devoted.

but I do understand the guiding principle behind a zero move

A99

[cygnussphere]

stats.zerothelement.com/index.php

These would be the stats referanced previously

[Starfish]

Originally posted by Insidious


(I do like the star idea identifying a user who has Phase I input next to the name on the Phase II page)

-Sid

I too like the idea of a star indentifying who has helped during Phase I.

I can understand that people who have done 10+ Million or even 100+ Million (took me a ~ year just to break 5M two weeks ago ) are not really waiting for a complete reset.

Perhaps the star idea can be realized in such a way that there are a few star colors like you see in some statspages?

Example: Den's Distributed Folding stats :: Team Endeavor:
http://users.neotechus.com/~hruzaden...-df-stats.html

(especially see the overview on the bottom of the page)

possible solutions are:

color x every xx structures
color y every yy structures

or

color x between x and y structures
and color y between y and z structures

That would give the people who'll "lose most" some distinction instead of "just one Phase I star"


Whatever the decision:

It's great to see that we're about to move to a next Phase..Cheers to all who helped achieving this!

[FoBoT]

Originally posted by Brian the Fist
Well, you just gave me a somewhat wild idea there, inspired by our friends eBay. What if we zero the stats as described, but keep (internally) a record of your production in phase I. Then we internally add phase I + phase II production and call that total production.

Stats pages will show only phase II production BUT you will get a series of coloured stars or other objects next to your name based on total production (I noticed some 3rd part stats pages do something like this already). Thus you can quickly identify people who've been around for a long time and contributed a lot even though their phase 2 production may not be so much.

Lemme know, and if you have a preferred color scheme..

just can't stop thinking of a better way, can you?

i am sure that would be more palatable to the people that think moving to a new phase in the project is "zeroing" the stats

i can't wait!!

[Anarchy99]

hey I like the star Idea

[reader50]

Question1: if units crunched today are going to be tossed, why should I crunch today?

We are fighting off another team in Folding@home this week, but I was crunching 24/7 last week.

Question2: if our stats are taken away, why should we crunch Distributed Folding in the future? There would be a precedent for taking our stats away now and then.

[AMD_is_logical]

Well, I voted for continuing with the old stats, but if we go with the star idea, how about a sort of spectrum.

* - up to 5M
* - 5M up to 10M
* - 10M up to 20M
* - 20M up to 50M
* - 50M up to 100M
* - 100M up to 200M
* - 200M up to 500M
* - 500M up to 1000M
* - 1000M and up

Note that I used an asterisk as a star. There may be a better font+character abailable. The page shouldn't try to load a bunch of .gif files, though. That would make the stats pages take longer to load.

[Goobee]

While a few of us have lodge our protests with respect to the zeroing of points, that issue appears moot at this point.

It seems that will happen whatever is argued here - the only real issue is what to do with the old results once the new project begins.

I do not see why it will be so difficult to link the totals from the old database to the new database and show the aggregates??!! A child of five can write simple code to link two tables and produce a sum.

Who cares if the algorithms in the old project vs. the new project are different..............stats are stats!

Are you guys asking for our opinions or simply giving us a rhetorical question.

[KWSN Grim Reaper]

I REALLY like the star idea too, although I'd like us to continue with the current stats so that we don't loose anybody. Someone mentioned running a poll...but what would be the margin for zeroing the stats? 51%? That would leave 49% not very happy with it, what would be the acceptable 'loss' of participants? Clearly some people have a big problem with it

Ni!

[FoBoT]

Originally posted by reader50
Question1: if units crunched today are going to be tossed, why should I crunch today?

We are fighting off another team in Folding@home this week, but I was crunching 24/7 last week.

Question2: if our stats are taken away, why should we crunch Distributed Folding in the future? There would be a precedent for taking our stats away now and then.

they aren't being "taken away"
they still exist, you still contributed "X" amount to the first phase of the project, that will never change

its just different ways of accounting for different parts of the project

[FlowerKid]

Exactly right FoBoT. This phase is ending, so I think the stats should be reset to reflect that. Just a designation of your final standing in Phase I would be sufficient for me, personally. I will still be crunching in this project no matter what the decision, and I hope others do too.

[MAD-ness]

I like your avatar FK.

You ought to drop by and say hi to your old friends every so often.

[Grumpy]

Well, if 49% are not happy with the zeroing, logic dictates at least having an aggregate added to the stats page..currently at OCworkbench the result is 55% against zeroing.

[Starfish]

Originally posted by reader50
Question1: if units crunched today are going to be tossed, why should I crunch today?

We are fighting off another team in Folding@home this week, but I was crunching 24/7 last week.

Question2: if our stats are taken away, why should we crunch Distributed Folding in the future? There would be a precedent for taking our stats away now and then.

Answer1: I don't think that the actual work is going to be tossed away....the project is just coming close to a new Phase and the question is: what are we going to do with the stats now that the way of generating structures is going to change?

Answer2:
No matter what happens to the stats, the project has a certain set of goals/a cause it wants to achieve.

It's up to you if you find those goals/the cause more interesting then possible resets of the statistics in the future.

Personally: I like stats..it's nice to have a 'dynamic playing field' around you while you're generating structures...but the cause of the project is what keeps me running this client... if the stats were taken away things would become less fun...but someone has to do the work

[StrategyFreakAMD]

I am all for zeroing. This would help to get rid of all the people with many stuctures but not producing so the smaller teams actually have a chance. It will also get rid of the thousands of inactive people who haven't turned in a structure at all. Finally, if the top 10 do have enourmous production, they will have no trouble getting back to their old positions.

If you were to set up a poll about the stats, at least 2/3s of the people should be for zeroing the stats.

[Bionic_Redneck]

I agree it won't make a difference if they split the stats. It might even get others involved in DF.

[pointwood]

Reader50, if you're only doing this for the stats, then why did you choose DF? Why not RC5 that have a client which have smaller requirements? I know that if I where only doing it for the stats, then RC5 would be my choice.

Right now we have a large lead and a retirement of phase 1 would give others a good oppertunity to challenge us. While it is nice to be outproducing everyone else with a large margin, there is not much fun happening. It was much more fun when we where trying to catch Free-DC.

With that said, I don't really care whether the stats are being reset. You, Howard, should do what gives the best options for making the project grow in the future.

I really like the way, you, Howard, handle this - makes me feel appreciated And I think just the fact that it has been discussed here will go a long way to satisfy most people, nearly no matter how you end up making it.

if the stats are reset, then I really like the "star idea". That plus having a page with the stats from phase 1.

[Brian the Fist]

Originally posted by reader50
Question1: if units crunched today are going to be tossed, why should I crunch today?

We are fighting off another team in Folding@home this week, but I was crunching 24/7 last week.

Question2: if our stats are taken away, why should we crunch Distributed Folding in the future? There would be a precedent for taking our stats away now and then.

a) because I didn't officially say stats will be zeroed yet (note I cleverly avoided actually stating this )

b) because now you will get a cool star colored based on your production since the start.

For the stars I was thinking of blatantly stealing them from eBay (would that be illegal or something???) cause I really like the eBay feedback stars, they're cool (come on, someone here must use eBay besides me..)

[reader50]

Wish more people would jump into the "No" side of things. I'm looking like a troll by carrying most of it.

pointwood, I usually evaluate a project first, see if it is doing something worthwhile. Then crunch for the stats if it is. If the project does not seem worthwhile *RC5*, I'll still crunch a few units to get a feel for it, so I can write our Install pages and build stats for it. Sometimes, I even jump in when the team is under pressure. But that's about it.

Stats are the intangible return for our volunteered resources. Besides recording our contribution, they also allow us to play games on the side such as KWSN and others sometimes do. This is ok.

The project may see stats differently, but I see them as "ours". They got the free CPU time; we got the bills, noise, heat, and our stats. Since the stats seem tied to our contribution rather than a particular project codebase, I have yet to hear a good reason for nuking them. A bonus spelling-bee star is not what I had in mind. Not even if it's in color.

Stats accounts are usually zeroed for cheating. IE: you submitted fake units to SETI in an effort to rise up the charts faster, or to push your loser team above more honest teams. Such accounts should be zeroed. But all my units were crunched honesly on my single home computer. On all projects, not just Distributed Folding. I didn't cheat, and put considerable resources into reaching my present position.

I want to keep the stats I worked for. Taking them away means I have to either drop DF, ignore a pitiful position at the bottom of the team, or return and crunch upwards for several months. Time that could go to other projects that are also worthwhile. This would certainly be nice for the DF project, but I already put that time in.

Side note for Howard: Why do foldingathome.com, foldingathome.net, and foldingathome.org all point to the DF project? Shouldn't they point to Stanford?

To others who wish to keep the stats they contributed: Please post. Contrary to popular opinion, Free-DC no longer requires you to join their DC teams before you can register on their boards. Also, the per-post charge is waived in the Distributed Folding project forum. Jump in. I can't carry it all, it's not worth getting banned.

[pointwood]

Reader50, I certainly don't consider you a troll (and I don't think others do either).

As I think I've said before, I don't really care what Howards decision ends up being, I just don't care enough about the stats

The reason to do a stats reset IMHO is (at least partly) because it is something similar to 2 seperate projects.

As I understand it, we are starting from scratch (well, it builds upen the current version, but it is much improved upon and works a lot different too) and that make it somewhat natural to start from zero with the stats too.

[FoBoT]

Originally posted by reader50
Side note for Howard: Why do foldingathome.com, foldingathome.net, and foldingathome.org all point to the DF project? Shouldn't they point to Stanford?

uh , maybe howard was smart enough to pay $15/yr for those domains and the *people* (edited out a mean name calling word ) at stanford aren't internet savvy enough to get those domains?

[FoBoT]

howard is in a no win situation here

if he starts a new set of stats for DFII , X% of the current crunchers will be pissed off

if he comes up with some magic formula to equate DFII production to DF1 production, then X% of the current crunchers will say it is unfair because , either

A- the new scoring is biased towards the old phase 1 people or

B- the new scoring is biased towards the new phase 2 people


howard, i don't envy you dude, X% of the current people are going to crap on you regardless of your decision

i still think DF is #1 DC project going right now and am excited to see what the science of the NEW system will bring (this coming from a moderate stats ho!)

[Insidious]

my impressions (without even looking at a stats page to back them up) would be that it seems odd that a participant would have such difficulty with being able to view what they accomplished in Phase I on a page and being able to view what they are presently doing in Phase II on a different page.

Could it be that there are users who did a lot for a while, but have decreased their commitment to a point that if only today's effort was on the page they would not look as good?

I have no sympathy for anyone who thinks there is actually much of an accomplishment to view if it all depends upon the past. I see nothing glorious about resting on ones' laurels!

I personally like the idea of those who have reduced their commitment to DF being displayed as such.

In other words:

If you were doing alot in the past, it showed then.... That was your 15 minutes. But why should Howard provide you a false front to make you look good if you are no longer putting in that kind of effort?

[Aegion]

Originally posted by reader50

Side note for Howard: Why do foldingathome.com, foldingathome.net, and foldingathome.org all point to the DF project? Shouldn't they point to Stanford?

The two projects were concieved of oringinally at about the same time. I believe Howard was going to call his project Folding at Home, but Stanford took the name for their project first. Read some of the details of the official DF site info for more info on when the Distributed Folding Project was first conceived of.

[PinHead]

Why not the little origami pics from the website.

Clean them up, shrink them down a little and colorize them.


Anyway, if I am reading the emotions correctly, it is starting to sound like a current project stats page with a trinket and a second stats page with cummulative stats might make the majority happy.

I don't think that it is important why a person donated their computer time; the important thing is that they did choose to! So hopefully everyone can chip in and come up with some common ground.

[horus2]

What I would like to see is 3 different stats databases. One with the first phase stats, one with the new phase and another one which is just a sum of the two. The final database would not even have to be updated as often.

I think the stars idea is a little gimicky though.