boomerang jus keeps cummin bak

Tuesday, January 26, 2010

Sultan of Swing !?!

Innovation is not just one idea; it is a lot of them put in series. There are a lot of failures, ideas and many moments of problem solving involved to achieve that one thing, which ends up overshadowing it all (which actually is the purpose). Scott Berkun here delivers a great lecture on innovation or, as I put it, how to develop any idea. I have had my share of making, breaking, re-making and often reorganizing my thoughts with one of the ideas at work in past 12 months. I don’t believe it would change your world much, but it has definitely changed mine.

After having attempted unsuccessfully a number of times to develop something out of cricket, an unknown time ago, I copied some code from Google Visualizations into an html file and put it in front of my teammates in Mumbai. We discussed if that was possible, what could be done to reach there and possible resources. For next few days we would discuss it during our brunches at ICICI, but it was not before one of the rock climbing trips that one of them told me he has some hands on web data mining and if we could use it. I said maybe we could mine internet for data and do what we initially planned. I forgot it over other things at ICICI, but Sandy came up with the first skeleton app that made it possible! As soon as we shared the idea with the rest of the team in Bangalore, everyone was on board. It felt easy, such was the intuitive gut that I had that time. And with initial results from the actual models made by our team, it seemed in a hands approach. But as it turned out later, it wasn’t.

There is a direct relationship between the time you spend doing a thing and the amount of learning you are going to get from it. Though everything looked great from a high level, things almost fell apart when we analyzed them in pieces. To make things worse, we had exhausted a lot of time in developing this phase and then were left with only option – quick fixes. Quick fixes have a way of not working sometimes exactly between the times when you need them, they will screw your original solution so bad; you would feel shooting your foot was easier. Though we had something substantial in hand, it wasn’t ready for sure. We sat together again, challenged each other's ideas, helped implement promising ones. In one of these sittings, we almost totally scrapped earlier approaches, and still got good results. Mashed it up with the old approach, added a lot of common sense, challenging our own ideas this time. Many brain storming sessions and thousands of lines code later, it seemed to be doing fine (did I say fine, it rocks!). We bounced back as a team, I am quite sure it has, in pieces, something from everyone in Marketelligent, which makes it even more special.

Last week, we launched the first version of Mswinger, a dial displaying which way a cricket (ODI) match is swinging. Not saying Mswinger has 100% accuracy in gauging the swing, but we are not far from that figure either. We sure have set the bar pretty high for anyone else to match. You can track the upcoming ODIs with it and let me know your opinions.

My takeaways are very high from this project (as for everyone else IMO), innovation is a sequence of events done every day and never settling for something in between.

Friday, September 4, 2009

The Approach

You don't have minute details here, but this pretty much summarizes my belief of work.

Wednesday, June 24, 2009

Kabhi screw driver nahi dekha kya?

A sudden call from brother took me to almost a ritual trip to Dadar today. Dadar station has kind of become a meeting joint for us when we are in a hurry or when seeing off someone.

You get a seat easily in trains towards Churchgate during evening, but you seldom get two men, taking seats off! As it turns out, the chaps were trying to take out a mobile which slipped from pocket during those profound sleeps. Entire compartment had taken various position to not miss the smallest bit of the rescue, some people had even started making video (between did anyone tweet about it??). The owner of the mobile has skipped five stations now and desperately wanted to get off the train WITH the mobile. In the process, the fellows managed to unscrew some of the seats but couldn’t take off what lied beneath it (seat is not wood deep my friend!). It was a hard steel sheet, having the least spaces on sides, I am sure even Aamir Khan wouldn’t have slipped through it.

Some geek had suggested earlier to keep calling the mobile which might let us trace it through the otherwise opaque sheet and might eventually cause the vibrations to make it slide (wherever). But even after stripping the local of 3 seats, the scheme failed, only victory was the location of mobile with an error in accuracy of about 5 cm (that is a big error just in case you were wondering how tall is Aamir). The shear effort was visible in the guys clothes , made me jump into the scene (I love hogging attention). Now all eyes on me (some decently attractive female eye balls included), I saw an obvious solution to the problem – IF we could get a screw driver of a particular size, the problem was solved (my exclamation was like telling Pakistan that terrorists are their problem), they all shrugged in synchronization as though we were playing in an Opera. A lot of people made heavy sighs, and suggested that he get down on next station as the train wont go to yard for another 6 hours . I kept trying (cynics said I was trying to impress the girls) and in a moment of disgust, I just pulled the bottom of seat so hard that the vibrating mobile slid elegantly into my palm (yes you geeks, you can give we-said-so looks now). A lot of people sighed again (cynics said no girls were impressed, but I beg to differ). The guy took the mobile as if it was stolen (wait a min) and in one swift action of legs, jump outside the compartment, pushing a teenager, with a screw driver in his hand, getting up the train vigorously in the process.

Yes, he had the same exact size screw driver in his hand. Everyone started chuckling moments after realizing what has happened. The poor chap sitting on the just-kept-there seat remarked "Kabhi screw driver nahi dekha kya?"

PS: Hey, we all love Aamir, he wont be mean to me. In fact he wont be even 85% of what I am (because then he would have to be taller!!)

Sunday, June 21, 2009

Play em smart

My take on what might be a better way to take in form Pakistan bowling:

• Mohammed Aamer should not be taken lightly, he has taken crucial (better opener, high strike rate, good average) breakthroughs in the first 6 overs in all the matches except when they played SL.

• Aamer provides perfect platform for Afridi to bowl sliders, if a wicket falls around this time, promote a left hander (depends on the kind of pitch as well).

• Ajmal has taken 10 out of his 12 wickets after 13th over, he is not the man you want to go after at the end, take him on in the beginning itself when he is not bowling dusras.

• Gul has been near perfect with his full deliveries, play as straight as possible, however Pakistan have had problems with fine leg in the circle, get the drift?

• Now the most interesting part, whenever Pak has to push back the spin for couple of overs and bring Gul up, they end up losing the match (warm ups included). Teams have lost to Pakistan while stepping up between 7th to 13th overs, 3 of the lesser scoring shots matches between these overs resulted in Pakistan defeat.

Bottomline: Don’t throw wickets in the first six, get left handers to nudge ball around with Afridi & take on Ajmal here. Make Pakistan bring Gul as soon as possible (by not loosing wickets and maintaining a decent run rate).

PS: Of course on a given day Umar Gul can bowl out entire SLan team in 10 balls. This is a plain strategy under assumptions.