No, you’ve done it, when you click on the team it has it’s
members and there stats, but you have brought up a point.
A link to each teams forum ? 
Still no starting date !? and my :bigdump:ing finger is already starting to itch ! :lol: 
I’m still addicted to SETI 1!
It’s all quiet from Sir Ulli 
I think that TPR being so up for a race has :scared: them all
away :lol:
[b]
COUNT ME IN
[/b]
Unless I missed the start already :rolleyes:
:lol: Some wish that it had started as they had huge cache’s
Anyway, I have to finish pottering around and get me son up so
he can take us to the airport.
something is going on so look here
http://forum.oc-forums.com/vb/showthread.php?s=&threadid=184940
and all about caching
regards
Sir Ulli
did I say I’m in??!?!?!?! 
LWD is ready to run,
with just a few suggestions to make it more fun:
[ul][li]The Last Dance:
[/li][li]Start: 15 June 2003 00:00:01 UTC
[/li][li]End: 30 June 2003 23:59:59 UTC
[/li][li]Caching allowed.
[/li][li]Even the score for large and small teams by calculating Lowest Average CPU Time to declare the Winner.
[/li][li]Multi-Mega-Crunchers (Avg WU per day > 500) not allowed. (May not be necessary if we use Lowest Avg CPU Time)
[/li][*]If you join during the race, advise how many WU you currently have when you announce IN THIS THREAD that you are joining. All Racers start with ZerØ WU. [/ul]The only problem is that SETI Team won’t let me gather data that includes CPU times quickly, so can someone grab those data and make a stats page?
Hi bud, I think Len’s stats could handle the CPU stuff.
He’s VERY good is our Len :hail:
Lets see what he says.
Originally posted by LWD_TOMaxwell
[B]LWD is ready to run,
with just a few suggestions to make it more fun:[ul][li]The Last Dance:
[/li]> [li]Start: 15 June 2003 00:00:01 UTC
[/li]> [li]End: 30 June 2003 23:59:59 UTC
[/li]> [li]Caching allowed.
[/li]> [li]Even the score for large and small teams by calculating Lowest Average CPU Time to declare the Winner.
[/li]> [li]Multi-Mega-Crunchers (Avg WU per day > 500) not allowed. (May not be necessary if we use Lowest Avg CPU Time)
[/li]> [li]If you join during the race, advise how many WU you currently have when you announce IN THIS THREAD that you are joining. All Racers start with ZerØ WU. [/ul]The only problem is that SETI Team won’t let me gather data that includes CPU times quickly, so can someone grab those data and make a stats page? [/B][/li]
Berkeley gives us
- total cpu time
- average cpu time
The average cpu time from Berkeley makes no sense here because it includes all previous crunched WU times. Any math geek w/ an idea how to use this numbers for calculating new Wu average times only?
Total cpu times are weak numbers. Why? This time is shown in years (w/ 3 digits) for big crunchers, but in hrs and min for small crunchers. What does this mean? For big crunchers only every additional 8.76 hrs are showing up in the total, while each min of a small cruncher is visible. Is that still ok?
I think any combination of number WU’s and average time might be better or a team could win with only 1 but fast WU only. 
I think your best judgement would be enough I would say Len
But obviously, the collective teams would need to agree.
:cheers: Len
I agree…basing each team’s results on average workunit time regardless of the number of participants actually discourages production…since you don’t want any of your slower (>3 hour) crunchers participating. The winning team would likely be the smallest team utilizing the fastest computer(s).
One other objection I might voice. It’s been stated that caching would be allowed for this gauntlet, but no finite ‘start date’ for caching stated, as was decided for the last gauntlet (everybody started once the captains said 'game on). Caching workunits from before a gauntlet is even announced is akin to starting a 1 mile race at the half mile mark. Basically this allows those with a “I must win at all costs” mentality to boost themselves compared with everybody else.
With the rules as stated, it would appear that it’s being conducted for all the wrong reasons. I don’t and can’t speak for anybody else on my team, but I’ve grown somewhat tired of SETI; I will continue crunching with my steadily-diminishing farm until the project ends, but I have little interest in such a gauntlet again.
Originally posted by Len
I think any combination of number WU’s and average time might be better or a team could win with only 1 but fast WU only.
I think a good measure on the caching lark is “Getting a Team
together, then caching can be measured from the day that the
team is finalised with reagard to the race starting a few days later”
This way, there is only a few days of caching. Len can measure
that there isn’t more than a few days worth of caching.
The race should be open to all team members.
But it’s the way that measures fairly and equally is the sticking point.
Originally posted by Len
[B]Berkeley gives us
- total cpu time
- average cpu time
The average cpu time from Berkeley makes no sense here because it includes all previous crunched WU times. Any math geek w/ an idea how to use this numbers for calculating new Wu average times only?
Total cpu times are weak numbers. Why? This time is shown in years (w/ 3 digits) for big crunchers, but in hrs and min for small crunchers. What does this mean? For big crunchers only every additional 8.76 hrs are showing up in the total, while each min of a small cruncher is visible. Is that still ok?I think any combination of number WU’s and average time might be better or a team could win with only 1 but fast WU only.
My suggestion was based on racing team totals, i.e.,
[ul][li](Ending Total Racing Team CPU Time/Total Racing Team WU) - (Starting Total Racing Team CPU Time/Total Racing Team WU),
[/li]OR
[li](Ending Total Racing Team CPU Time - Starting Total Racing Team CPU Time)/(Race WU)
[/li][li]but other suggestions are solicited[/ul]
[/li]
It’s not easy figuring out a way to make it competitive, fair and doesn’t-leave-a-bad-taste-in-your-mouth. 
It would be great if it encourages N00bs to join and inactive accounts to fire up one last time. As far as caching, you can’t really stop that anyway, because some people have no choice but to cache (e.g., Fido has to burn a CD so he only sends comp’d WU every other week, and on dialup, it takes a whole workday) Many of us aren’t able to cache too.
Originally posted by LWD_TOMaxwell
My suggestion was based on racing team totals, i.e.,
[ul][li](Ending Total Racing Team CPU Time/Total Racing Team WU) - (Starting Total Racing Team CPU Time/Total Racing Team WU),
[/li]> OR
[li](Ending Total Racing Team CPU Time - Starting Total Racing Team CPU Time)/(Race WU)
[/li]> [li]but other suggestions are solicited[/ul]
[/li]>
It’s not easy figuring out a way to make it competitive, fair and doesn’t-leave-a-bad-taste-in-your-mouth.
I might be wrong (late, tired, etc.) but
(Ending Total Racing Team CPU Time - Starting Total Racing Team CPU Time)/(Race WU)
would not fix the problem of winning with only 1 fast WU.
How about:
(Race WU’s)/(Ending Total Racing Team CPU Time - Starting Total Racing Team CPU Time)
So many WU’s would give you a high number, but with shorter crunching time you can equalize this again.
Many WU’s with high crunching times can lead to a lower score than less WU’s with low crunchung times.
This would allthough reduce the effect of big time caching.
Am I completely wrong here? Please check and let me know what you think.
How about :
(Race WU’s)/(Ending Total Racing Team CPU Time - Starting Total Racing Team CPU Time)
this sounds good to me
regards
Sir Ulli
That seems a good way Len.
I know I certainly couldn’t have worked it out.
Does it mean you’d have to redo your code a fair
amount ??
Originally posted by Sir Ulli
this sounds good to me
It works for me 
Originally posted by dalethfc
Does it mean you’d have to redo your code a fair
amount ??
Adding 1 or 2 columns to the rows, sorting by score instead WU’s, changing output … simply some additional coding.
I am loading the full data of the teams down from Berkeley anyway; so there is no prob if I don’t get it finished before the race starts (on the 15th?).
Additional points (nothing new;)):
- I suggest to have a team account for each team and late joiners are using those; else I would have to poke each one into the stats by hand.
- Please no name changes or leavers. Can’t be so hard to stay in the race with your account names for 2 short weeks. The stats only show missing names yet but the automatic handling of their results is still missing.
The only thing I have done on the code since the last race was the page layout (TPR style now). Wish Berkeley would show a unique id for each user; without that a change in the list of names is causing alot of guessing.