Full DPAD Stats Soon to Come

[magnav0x]

Ok I just completed my climate prediction stats and I will begin working on the DPAD very shortly. As you should know, I have already got Free-DC only stats up at http://stats.dbestern.net

I'll keep you all posted on the progress.

[magnav0x]

Ok the stats engine is now live and complete. For those that have or plan on participating in the 4.3X client, the stats are located at:

http://stats.dbestern.net/muon1

We have just passed US Distributed :thumbs:

[the-mk]

Originally posted by magnav0x
Ok the stats engine is now live and complete. For those that have or plan on participating in the 4.3X client, the stats are located at:

http://stats.dbestern.net/muon1

We have just passed US Distributed :thumbs:

You want to be a stats god, huh? Great work man! :cheers: :thumbs:

Congratulations for passing US Distributed! :cheers:

[wirthi]

Oh shit, maybe I should invest some power again in this project or else I'll lose #1 position. In fact that's my only #1, so I should fight for it .... (although this is a bit unfair, others have done far more work for the team than I have, I guess the older stats for pre-4.x version were truncated or so).

Great stats man!

[the-mk]

Question: what is better for "Best Muon Yield"? High percent-numbers or low percent-numbers?

And for wirthi: how many time (CPU-time) and how many computers did you invest to this project to become number one?

Perhaps I should move to a project with damn good stats :D

[wirthi]

Like always (but unlike DF): the more the better!

If you run the graphical client you see (after some time) particles that fly along a path through cylinders. Those cylinders (that expose a magnetic field) controll the flight-path of the particle. The project tries to find an optimal answer how to position and size those cylinders so that a maximum number of particles reaches the target (at the end of the path).

When you watch the client in the beginning, a lot of particles will fly in the wrong direction right in the beginning, many will collide with the wall, and so on; only a minimal fraction will reach the target. After a log of calculations less and less will collide and the number of particles reaching the target will increase.

This project uses an evolutionary algorithmn just as DF is using right now. The first paths are calculated randomly, after that the best ones are taken and modiefied and testes if they give a better result.

The fun about this project is that you can somehow controll the results that pop out. If you run it with a high number of already good simulations a lot of other good simulations will come out. If you start right from the beginning you'll first get poor results but over the time they'll get better and better.

How much time I investe? Honestly I don't know. I'd guess about a months value of my 3 PCs at the time I used the project (Cel-400, P3-500, P4-1500). Could be much more!

[the-mk]

:Pokes: A suggestion for the stats (http://stats.dbestern.net/muon1/):

How about sorting "Potential Conquests" and "Potential Threats" by "Days Until Overtake" and not by "Position Difference"?
Especially on the user-pages (http://stats.dbestern.net/muon1/user2.php and http://stats.dbestern.net/muon1/user.php) and ev. on the teampages (http://stats.dbestern.net/muon1/team.php)?

Just a suggestion...

[the-mk]

I found something funny with your stats, in the headline of the page is an error:



DPAD Muon1 Statistics Engine by magnav0x[Free-DC.org]
Database Last Updated At: 11/02/2003 26:00 GMT


Do we have a new time system where nobody told me? 26:00 GMT is just a bit funny, isn't it?

[wirthi]

Originally posted by the-mk
I found something funny with your stats, in the headline of the page is an error:

Do we have a new time system where nobody told me? 26:00 GMT is just a bit funny, isn't it?
Great, gives me more time to do all my work :D

[the-mk]

How often do these stats (http://stats.dbestern.net/muon1/) update? :dunno:

[magnav0x]

Question: what is better for "Best Muon Yield"? High percent-numbers or low percent-numbers?

A higher Muon Yield is better

Yeah I noticed the problem with with the timestamp :blush: It's something I'll get around to fixing, but I have 2 other (work) projects I HAVE to get done this over the next two weeks, so I figure it can wait.




How about sorting "Potential Conquests" and "Potential Threats" by "Days Until Overtake" and not by "Position Difference"?


That is something I may look into doing in the future....see above answer :rolleyes: It also poses a problem with doing that, because the overtake dates are not stored on in a database, they are calculated on the fly. It just so happens that I'm pulling them out of the database before the calculation based on total Mpts of the potential threat/conquest. It would, to say the least, require a database/engine overhaul.



How often do these stats update?
My stats are updated daily at 6PM PST (2AM GMT).

Offical team stats can be found at (these are supposdly updated every 2 hours):Pokes: :
http://www.stephenbrooks.org/muon1/v5.html?hide=1&teams=1


There are official individual stats in there too at: (scroll to the bottom of the page)
http://www.stephenbrooks.org/muon1/

I don't think they actualy get updated every 2 hours, but I never realy watched it. BTW you can check to see when and how many results you sent by checking sendlog.txt in your muon folder. It will only show how many results you sent and now how many Mpts they are worth. To get an idea on how many Mpts you get per results go into your results.txt. After completing each one it puts all the information in there, including Mpts value. It varies quite a bit per result. Sometimes I get one that's 7 Mpts and other times they are 100+ Mpts.