As a reminder, following Statistical Ideas is easy to do, through @salilstatistics, facebook, YouTube, LinkedIn.
Columbia Adjunct
Georgetown Adjunct
NYU Adjunct
Editorial Board in American Statistical Association (+ peer-reviwer & honoree)
Kids' Statistics
Former two-time Obama Admin executive. Executive consultant with investments or exits by firms like BlackRock and DFJ; >12 tech growth firms and nonprofits in varying sectors. This is the leading educational probability site in the world, with over 220 thousand followers, including celebrities, and 32 million reads. It's also integral to the curriculum of at least five leading universities and also followed by central banks.
Academic research contributor in arXiv (scientific), SSRN (social science) and RePEc (economic), and this site is also added to a selective academic aggregator. Abstracts have been among most downloaded. And an RSS subscription icon is on the right toolbar of this web log.
Thanks much for the message manoor. You should consider following the blog using the form above (or RSS icon on the right panel of this blog site). I will also be posting some new blog entries soon, mostly around central tendency and distribution fits. Both are related to recent applications that may be of interest.
ReplyDeleteSalil, Do you have a twitter handle ? Regards -Ashok
ReplyDeleteThanks much Ashok. I don't have a Twitter account yet, but one can always connect with me on here, or via facebook!
DeleteSalil, I plugged your blog on my blog today. Thanks for the insights. Scott http://medicalevidence.blogspot.com/2014/04/underperforming-market-why-researchers.html
ReplyDeleteWhat are you basing the fact that Buffett underperformed or outperformed the S&P 500? Book value or Buffett's adjusted book value? I know that Buffett uses that, but the comparison seems odd these days. As Berkshire has become more of an operating company, Berkshire's book value is more comprised of corporate assets than financial assets. And corporate assets are subject to cost basis accounting rules that really don't have a lot to do with the underlying performance or value of those assets. Couldn't that be what's going on? And why there seems to be a sudden change in Buffett's performance? That change seems to match around the time the company bought Burlington Northern, an asset intensive business with a lot of depreciation. Thanks.
ReplyDeleteThanks much Stephen. My primary emphasis is on teaching probability and statistics, and only looks at specific, probability-related comments that were highly visible in the report to explain large investment ideas (e.g., rationale for when BRK underperforms, and coin flips). This is a very useful set of skills people can use to think about a number of types of decisions they need to make in life or among choices afforded to them. Other people can have an opinion on whether for BRK specifically they would like to now use a different standard or hypothesis despite what statistical performance data and some statistics commentary continues to appear in the annual report. And going down that road, along with other explanatory ideas radiating out from the financial press, only muddles the general conversation about how one should think about their own performance chances.
DeleteBuffett has outperformed by a wide margin over his entire career, whether or not you inject a different personal hypothesis for Buffett's recent (4 or) 5 years. I am pleasantly in agreement with everyone else on that. It still doesn't change that for most people, the idea of having an investment vehicle outperform (let alone with skill) a specified ex-ante, market benchmark over the long-haul is still a challenge.
Hi Salil,
ReplyDeleteI would like to know further about your academic course, Can you please provide further detail on that more from Stock Trading perspective.
Thanks
Abhay
Thanks much Abhay Dodiya. My university graduate courses focus on mathematical statistics, analytics, and biostatistics. They can provide business students the tools for financial and investment analysis, though it is not designed to be a stock trading course.
DeleteOn an aside, you might be interested in my new book, Statistics Topics. Here is some info:
https://sites.google.com/site/statisticalideas/statistics-topics
Thanks for your feedback.
DeleteHi Salil- I've recently come across your articles, and am enjoying them a great deal, but they raise more questions :) I have a couple of mechanical questions for you- what tools do you use for your analysis? Just plug everything into xls or a more sophisticated package (eg, SAS\SPSS) or do you create from scratch (eg, data in SQL, create your own queries, leverage graphics libraries)? And what do you use as your data source(s)? I'm curious how much effort is required to recreate\expand some of your work (I'm not a dev\stats guy btw, but don't mind trying to dive into the details fm time to time).
ReplyDeleteRegards
Quinn
Thanks much Quinn. The focus of this website is to demonstrate the development of probability theory and discuss how it connects to real world applications. So I am glad you are savoring them, though more than half of my articles should involve no software at all. For the others I do have access to all of the packages, and teach a few of them currently at Georgetown. I like to rotate among all of them so that the visuals are fresh for the audience. The data sources are usually official sources (e.g., government), unless I link to a specific private party that contains data sources. And at times I link to and provide data, so that anyone else (usually other academics or business professionals) can analyze them. Those are stored collectively here: https://sites.google.com/site/statisticalideas/
DeleteEnjoy!
Nice- thanks!
DeleteQuinn
The largest website in the world, eh? Source?
ReplyDeleteLook, mere attacks are going to get removed. I clarified my comments on this site to state statistics website, but would appreciate you writing something new that is constructive in the comment section, as opposed to taking a negative view of my intentions.
DeletePretty high opinion of one's self ...
ReplyDeleteAre there some facts you disagree with, anonymous purveyor?
DeleteSalil- as a long time researcher and scientist who has recently altered many of his viewpoints based on data rather than just "feel good" narratives, I recently came across your blog. I have to tell you it is such a breath of fresh air, since there is so much bloviation going on, especially on the net, and on so called social media, which while pandering to emotions and predetermined beliefs, do little to actually view what is really going on under the hood. Thanks so much, and please continue writing. Your posts are informative, and helpful, even the few that i find diverge from my own pre-determined beliefs. But even that is critically important. Thanks, again. Barry Newman, MD, MBA
ReplyDeleteHi Barry, just now seeing this. Thanks for the kind words, and please be well. BTW the blog is nearing 100k followers on twitter alone (@salilstatistics)! Drop in additional comments, as you like.
DeleteWe hear about people getting bumped from their airline in the news a lot now. I would like to see the other side, how many times did an airline 'oversell' a flight, and the airline DIDN'T have to bump anyone because there were a couple no shows, like the statistics predicted. I bet the 'oversells' work out in the airlines favor (with no one getting bumped) quite frequently, but I can't seem to find the success rate anywhere.. My guess would be that somewhere in the neighborhood of 80% or 90% of the time, when the airline is in an oversold position, there are enough no-shows that no one has to get bumped at all...
ReplyDeleteGreat comment and this data is somewhat available. I have to go back to the data but don't think there are any underselling of a flight (or one where simply less ticketed passengers show up for a plane capacity). Usually it is the NERVOUS bartering over monopoly money to get "volunteers" off the plane, though enough terrible outcomes have happened that is causing airlines to simply remove this insane practice that is exclusive to just this one industry (car rental and hotels are not operating like this and don't plan to). http://statisticalideas.blogspot.com/2017/04/fine-if-all-are-re-accommodated.html
DeleteSorry, I didn't read down far enough to see your April discussion. That got close to what I was thinking about...
ReplyDeleteThat's great Anonymous. Hope all is well!
DeleteHi Salil,
ReplyDeleteI'm doing some work on AVM accuracy and very much enjoyed your piece on "Fat tail Zillow errors". I'd be grateful if you could clarifiy something.
You say "...average error (when a consumer is in the outer 10% of consumers) that is as little as 16%, and as high as 80%!" could you clarify what "outer 10% of consumers" means, thanks.
Some of symbols do not reproduce very well - any chance of clarifying - I'm looking into applying EVT.
Great article.
Best,
FG
Thanks Farmer George and good luck. Check out the graphic provided in this case. In each county we look at the distribution of absolute errors, and mark the cut-off for the top 10%. Then the graphic shows the distribution of these cut-offs among all the counties. We see the low-end of this graphic is 16%, and the high-end is 80%.
DeleteHi, I've responded on the Zillow thread. Thanks.
DeleteHi Salil,
ReplyDeleteIt's Gordon (from twitter). Long time no see! Hope you're doing well, I'm finally getting around to checking out your content (there's quite a lot). Are you still teaching at NYU? If so, I would like to sign up for your class in the spring...
hope you and your family are well!