## Questions on Ranking ...

**Moderators:** Stardust@home Team, DustMods

Sorry. I certainly didn't mean to bite your head off. Maybe I'm spending TOO much time on this thing and should consider giving myself a break. Besides, the last time I checked, Switzerland was still a neutral country. I'll see what I can do about showing you my file. But don't expect to learn too much since I don't have any more of an idea about what is going on than anybody else. I just happen to be one of the lemmings who can run a little bit faster than the rest that's all.

### Here is a tip!!!

Hi There,

You can copy your info and send it only to who you want by private message, to that person and not on the posts!

Hope this helped?

Howie

You can copy your info and send it only to who you want by private message, to that person and not on the posts!

Hope this helped?

Howie

If you ask whether I am neutral and fair or not, of course "we" are. Always exerted in order to keep a level head and to rest easy all together, but also, for example, to battle for naval supremacy - worldwide, thanks to the millions from the (private) economic system.foxranch wrote:Greuti,

Sorry. I certainly didn't mean to bite your head off. Maybe I'm spending TOO much time on this thing and should consider giving myself a break. Besides, the last time I checked, Switzerland was still a neutral country. I'll see what I can do about showing you my file.

### Higher Rankings

.

___

___

**Answers on Ranking...****Rank**______**Score**

18500______9

15000______15

12000______32

10000______57

9000_______72

8000_______92

7000_______117

6000_______154

5000_______203

4000_______280

3000_______412

2000_______661

1500_______931

1000_______1490

900________1650

800________1860

700________2148

600________2466

500________3000

Last edited by fjgiie on Tue May 29, 2007 5:51 am, edited 1 time in total.

Well,DustSabre wrote:Has anyone tried plotting the exponential curve associated with this data and the continuing data up through the top 100?

Also, what is the score at 1000 these days?

using the data of the Top 100 list, some recently postet stats (down to rank 2100) and fjgiie's data leads to the formula:

y = 1E+06*x^-1,0043

The current score at rank 1000 should result at about 970, rank 500 at about 2000. At higher ranks the results will be undervalued.

Hope that helps.

Nice job coming up with the function. Did you do exponential regression to get that? In reality, however, that function is really just a particular solution to some differential equation that describes the rate of change of the score at any level. If you could find the differential equation, you could then solve it for any time you wanted, thus being able to predict what the same exponential curve would look like in the future.The current score at rank 1000 should result at about 970, rank 500 at about 2000. At higher ranks the results will be undervalued.

Hope that helps.

Thank you, the moon, for your input which is surely the most precise answer to DustSabre's initial question.

Thank you, DustSabre, for your reply.

I am afraid but I'm not a mathematician and solving differential equations had never been my hobby. I just let excel generate a trendline based on the available data and picked up the equation. A quick job. I'm not quite sure what excel does in the trendline mode "potential" but this mode secures the best correlation coefficient R^2.

Unfortunately there is too much data at the top rankings and not enough on the other side.

The data of the top 100 (from today) processed as described result in

y = 275825x^-0,5835 which would promise a score of 5039 at the moon's rank (953)

So I tried it with +/- equal intervals from rank 100 to 2500, using the given data and putting some interpolated data into the pool to describe the middle range of the scale.

y = 2,5E+06*x^-1,082

Here are some results using that equation:

Rank / Calc. value / reportet value

100 / 17137 / 17100

117 / 14460 / 13879

128 / 13120 / 12750

183 / 8912 / 9021

298 / 5258 / 5018

953 / 1495 / 1557

2002 / 670 / 660

Just a try! Any better ideas are welcome.

Thanks.

Thank you, DustSabre, for your reply.

I am afraid but I'm not a mathematician and solving differential equations had never been my hobby. I just let excel generate a trendline based on the available data and picked up the equation. A quick job. I'm not quite sure what excel does in the trendline mode "potential" but this mode secures the best correlation coefficient R^2.

Unfortunately there is too much data at the top rankings and not enough on the other side.

The data of the top 100 (from today) processed as described result in

y = 275825x^-0,5835 which would promise a score of 5039 at the moon's rank (953)

So I tried it with +/- equal intervals from rank 100 to 2500, using the given data and putting some interpolated data into the pool to describe the middle range of the scale.

y = 2,5E+06*x^-1,082

Here are some results using that equation:

Rank / Calc. value / reportet value

100 / 17137 / 17100

117 / 14460 / 13879

128 / 13120 / 12750

183 / 8912 / 9021

298 / 5258 / 5018

953 / 1495 / 1557

2002 / 670 / 660

Just a try! Any better ideas are welcome.

Thanks.

Last edited by stardust1 on Wed May 23, 2007 10:55 am, edited 1 time in total.

Wir leben in einer Zeit vollkommener Mittel und verworrener Ziele. (Albert Einstein)

The average duster has 467.33623 points and ranks #2715.

There is not enough information to find the rank of the duster such that one half of the points are above and half of the points are below. However, we have not to date let something like lack of information stop us from making a good guess. Around the member with a rank of #240, one half of the points are above and half are below. Why is that important?

The members higher ranked than #240 as a group have the opportunity to find one half of the tracks made by interstellar dust. Also dusters ranked between #1 and #240 have the same opportunity.

Thanks,