Weighted score of Diwali contest

Submitted by admin on Mon, 11/29/2010 - 17:36
Here is the weighted score of the rangoli results. It doesn't change the outcome of the winning entry, although there are some changes in the placements, according to the votes received. The top entries in both the lists are more or less the same. And we must add that the weighted score can get skewed too. It all depends on how you slice and dice it. :)
EntryWeighted Score
3586
3337
3189
3078
3070
2956
2940
2877
2843
2785
2741
2568
2553
2417
2408
2312
2243
2242
2234
2215
2212
2137
2122
2087
2085
2051
2030
1926
1848
1808
1789
1688
1666
1651
1639
1631
1626
1566
1558
1551
1525
1513
1474
1428
1418
1349
1233
1163
1098
1081
902
898
jkmrao

Very Interesting and rich data! How did you arrive at the weight, is it like what I mentioned earlier? In this case, it did not alter the result at the top because that entry is perhaps of a very high quality. But in other cases, it might and in other years, it might have. I looked at the two sets of results for the first twelve places, which is an interesting set. In both the sets there are identical entries, except for placing. The maximum change is seven places, in one case. Perhaps in such cases, one of them was voted by people who voted for almost all and the other by a large number who voted only for a few. In the ikOlam case, where there are large number of entries and a large number of voters, the effect may not be appreciable. But in the other case, where these numbers are small, this is significant. latAjI, one request. Please consider my observations as purely academic and as a mathematical problem. We, in my field of interest, viz. crystallography, treat lots of data and statistical analysis is a very important facet in our studies. For example, in protein crystallography, a few hundred thousand reflections are recorded experimentally and thousands of atomic positions are considered. My observations are not meant to hurt any one, either the contestants are the administration. Regards! - mOhana
Tue, 11/30/2010 - 17:10 Permalink
deepa ram
hip hip hurrah, i rank 5th ,thank u all for voting me,,when is the marghazli contest Lata mam,i will try this time to rank in top three!eagerly waiting for it!
Tue, 11/30/2010 - 21:31 Permalink
kameswari
Thank you Latha for taking so much pain in declaring the results of all entries and I also thank all the members who voted for the entries and the encouragement given.
Wed, 12/01/2010 - 00:33 Permalink
Lata
Yes Mohanaji, the placement does change in the weighted method , as represented in the picture posted by you. We also found that the weighed score can fluctuate in any direction; for example in one of the models, we found that all it took to influence the results differently, was just four voters who knew how the weighting method worked! We're glad to be able to analyze the statistics and learn more about these type of calculations Mohanaji, and we welcome more input from anyone who can shed more light regarding this matter. However, I must also add that, our perspectives are a bit different from yours. You see, from what you've mentioned here, sounds like in your field of work, there are millions of life at stake (not to mention money too), and everything that members in your department do/work on, comes under the lens, both figuratively and literally. But at ikolam, we look at the big picture and don't have the time or the energy to further slice and dice the data, every which way. Also, we don't get paid like the way the members in your department do, probably since no one's life is at stake here. :0 So that pretty much explains why we, stop after a certain point, and stick to declaring the top winner followed by other positions. In conclusion, I can say this much, that we do constantly improve our processes :)
Sat, 12/04/2010 - 21:02 Permalink