## Profile

Rags Srinivasan
Worked at Salesforce.com
138 followers|107,108 views

## Stream

### Rags Srinivasan

Shared publicly  -

"In true life, we can’t all be geniuses. As if to prove otherwise, he keeps telling us that Charles Darwin was no good at school – but from that it doesn’t mean that the modern louts leaving school with no GCSEs today will go on to write an On the Origin of Species."﻿
2

### Rags Srinivasan

Shared publicly  -
1

### Rags Srinivasan

Shared publicly  -

#measure #analytics #statistics #Pinterest #RevenueModel﻿
1

### Rags Srinivasan

Shared publicly  -

You don't see Bayes in popular media that often. It tells you how important it is to explain such complex theorems in simple words.
"he footwear expert made what the judge believed were poor calculations about the likelihood of the match, compounded by a bad explanation of how he reached his opinion. The conviction was quashed."﻿
Bayes' theorem is a mathematical equation used in court cases to analyse statistical evidence. But a judge has ruled it can no longer be used. Will it result in more miscarriages of justice?
1

### Rags Srinivasan

Shared publicly  -

Power of metrics driven decisions: 59% of those who did WSJ's Blindfold Bracket got 2/3 of their predictions correct. All their predictions based only on metrics.
I am in the top 15% with 27 out of 36 games correct.﻿
1

4 out of 4 today so far. Waiting for regression to the mean to kick in.
If anyone is really asking, here is my simple algorithm for my Bracket choices

I treated all the metrics as linear interval scale from 1 to 5.
I ignored the metric on Hot streak. It didn’t matter (teams will regress to the mean)
Offense and Defense are my first criteria. Both got equal weight with an exception. When a team with better offense faced a team with better defense I picked the defensive time unless Offense is 2 points better than opposing team’s defense.
If both Offense and Defense are equally balanced I picked based on experience.
I mostly ignored 3-point shooting because I did not bother to check prevalence of 3-point shooting in NCAA.﻿

### Rags Srinivasan

Shared publicly  -

My Blindfold Bracket based on metrics presented by WSJ stands at 22/30. My first ever Bracket!﻿
1

### Rags Srinivasan

Shared publicly  -

My thoughts on Risk Taking

I believe one's action deserves the label Risk Taking only if one has multiple options available, each with its own outcome, and one picks an option whose outcome is most advantageous.
If one picks the one and only option available to them, however uncertain the outcome is, it is not risk taking.
As a corollary one should take risks when one can most afford it and not as the last action.

Using NFL analogy
- Onside kick on opening kick-off is risk taking.
- Onside kick when down 3 with 30 seconds on the clock isn't risk taking

- When you have a successful product line introducing a new one with potential but unproven market and may even cannibalize your current one is risk taking
- When your current offering is commoditized, introducing the only possible product isn't risk taking

Next time you are interviewing a candidate about a time they took risk, ask them them whether it was onside kick on opening kickoff or in the last 30 seconds.﻿
1
1

### Rags Srinivasan

Shared publicly  -

How Big is Infinity?

Seven mins thirteen seconds of awesomeness.

Using the fundamentals of set theory, explore the mind-bending concept of the "infinity of infinities" -- and how it led mathematicians to conclude that math itself contains unanswerable questions.

You do not need to be science or math savvy to enjoy the video. It is for all of us!

#watch #letyourmindbeblown
.﻿
1

### Rags Srinivasan

Shared publicly  -

When Social Media studies report likelihood of purchase ...

This is a study from CMB, http://goo.gl/SsQZc, that found
"The real question on our minds was does engaging with consumers via social media sites like Facebook and Twitter make consumers more likely to buy or recommend the brands they follow? 51% of Facebook fans and 67 percent of Twitter followers reported they are more likely to buy that brand since becoming a fan or follower."

First this is a stated preference survey not based on actual behavior. You can't read anything into this report.

Let us take this as true. What the study says is
P(Customer | Fan) = 0.51
In the previous paragraph it says, P(Fan | Customer ) = 0.49

P(Customer | Fan) = P(Fan | Customer) X P(Customer) / P(Fan)
So what this means is,
P(Customer) (probability that a person is a customer) is almost same as the probability they are a fan P(Fan)
In other words, anyone who is a fan is likely a customer as well and vice versa. What CMB should say is, don't bother.﻿
1

### Rags Srinivasan

Shared publicly  -

Sweet 16 - Got 6 out of 8 games right. Now in top 10 of Blindfold Bracket. Not bad for a first timer? Metrics Rule!!
BUT probability helped. Metrics only point to likelihood of one team winning over another. That is not given however. Chance played a role in pushing me to the top 10. But basing decisions on metrics (took less than 15 minutes, although consuming more now due to bragging) gets you roughly right over spending hours on myths http://projects.wsj.com/blindfold-brackets-2012/viewscore.php﻿
1

### Rags Srinivasan

Shared publicly  -

The WSJ writes, "The most diligent students of basketball, people who've spent hours gaming out their brackets, saw them instantly shredded."

I am proud my first ever based on simple metrics is in Top 4% with chance to be correct in rest of games.

Little bit of metrics beats tons of lore

http://iterativepath.wordpress.com/2012/03/15/using-metrics-over-myths-in-hiring/﻿
1

### Rags Srinivasan

Shared publicly  -

I recently talked to CEO of Boulder based SurveyGizmo on their pricing simplification. Here is my conversation with Christian Vanek﻿
1
Story
Tagline
Working on it
Education
MBA