diagonal argument

I’m happy today that we’re launching Suruk, a dream app that I always wanted to use. A few weeks back, I talked to Anenth about this idea and he got thoroughly excited. He then teamed up with Sandeep to finish the whole app in just about a month! We only had a couple of meetings in coffee shops and a couple of demos using SharingGlass add-on for Skype. The alpha client is fully functional supporting a digital meter, driver rating, ride pooling and emergency SMS. Please download it and use. We’re looking forward to your feedback to make it more usable. Let us urban commuters join hands to make our commutes cheaper, safer, and together!

Based on my hyper-coloured vision and non-training in psephology, following is what I think will emerge on May 16. They’re not necessarily what I want to happen.

All-India level
1. BJP becomes the single-largest party. The analysts come up with all theories from the weather to the staggered election schedule to what not. But, I think it may have to do with BJP’s strategy of silence and letting the UPA fall on its own weight. (Probability: 60%)
2. Left rout in Kerala and West Bengal, a drop by at least 50% of their current seats. Overall seats might be maintained in other states like TN. (Probability: 70%)
3. UPA still emerges as the largest pre-poll alliance. Woos everybody under the sun. (Probability: 70%)
UP and AP could create a flux.

Tamil Nadu
No landslide for any combine. DMK, VCK win most of the seats they contest. Congress lose in many seats to the AIADMK and the left. PMK loses badly. DMDK might spring a surprise by winning a seat. BJP might, with a probability of 10%, get one seat. Also, to be noted is Jayalalitha’s implicit “acceptance of defeat” in starting to blame the EVMs.

Post-polls, Stalin becomes Deputy CM of TN, Congress might join the state government (currently minority) if they get bargaining power.

South TN
Madurai – MK.Azhagiri wins, AIADMK accuses that money power won.
Dindukkal – AIADMK
Theni – AIADMK
Thoothukkudi – DMK (AIADMK hurt by anti-Sethu Samudram stand)
Virudhunagar – Vaiko could lose to Cong (Probability: 60%)
Thirunelveli – Congress
Nagercoil – DMK, but BJP could be a surprise winner here
Sivagangai – P.Chidambaram wins

What do you think?

You can find my Delicious bookmarks at http://delicious.com/sundar

Wow! I’m proud to see my finding a regex bug helped Compete correct its stats and set the record right for Yahoo!. As a result Yahoo! Search marketshare is now ~19%, which is up from the wrong percentage of 13.3%. Yahoo! surely needs some positive buzz at this point than any other. That too this one they have the right to.

I’m happy for my alma mater (aka ex-employer). At the same time, I’m mildly disappointed that neither Danny (@SearchEngineLand) nor Jeremy (@Compete) chose to acknowledge me by name or link to my post. Even other commenters have felt so. Glad that Webpronews managed to dig the link out from my comment at Danny’s post.

Edit: Compete has linked to my post now. Max from Compete has commented below saying that they meant to give credit to me. Thanks guys.

[Update: Compete ups stats: Y! up from 13.3% to 19%]

Steve from Compete has posted a comment confirming that their regex did have the bug pointed out by me. He, however, added that for capturing clicks on result pages, they use other regexes which don’t have such a bug. He assured that they’ll do a recount and update their numbers. I don’t know how much difference it’ll make to search engine market share numbers. But, any increase will do good for Yahoo!, I suppose.

[Update: Compete has confirmed that it was a bug.]
[Context: My previous post]
I had earlier suspected that Yahoo! Searches are severely undercounted by Compete because of a regex bug. Now, my suspicion has been confirmed. Danny Sullivan of the SearchEngineLand has confirmed that those regexes were indeed given by Compete.

By some estimates, refinement queries constitute over 70%? of the total queries. Is Compete, a widely quoted stats provider, undercounting Yahoo! Searches by 70%?OK, in a hurry, I might’ve dramatised the impact, but some significant chunk of searches are undercounted. Doesn’t it amount to a huuuuuuuuge difference to Yahoo!’s market valuation?Well, I hear that other Stats are more widely quoted than Compete.