Corner of Berk-Scraper and Fun-Facts

[prunes]

A few random facts for you guys:

 

What has been the best performing investment idea?

MGIC (7.6x multiple since thread creation)

 

What user has posted the most investment ideas?

Viking (22 ideas indexed)

 

What user's investment ideas have performed the best?

This is a much more difficult question to answer, and the answer depends upon how you evaluate the question. But, possibly Packer16.

 

~

 

I've been working on a webscraper that crawls the board and tracks investment performance. I thought you guys might find this interesting and wanted to share it with the community.

 

It's by no means perfect right now but it's at the point where I feel comfortable showing it to everyone. Depending on how much value the board gets out of it, I will continue to improve it. Note also that I'm not a professional coder. I just do it for fun.

 

I'd love to hear everyone's feedback and whether there are any analyses people would like to see, features people would like added, etc.

 

http://cobff.chrisdrane.com/

[constructive]

This is pretty cool, thanks!

 

So the average return is 29.8% and median return is 15.8%. What would those numbers be annualized, versus the market?

 

I wonder how this will compare with the subset selected for the CoBaF Fund: https://docs.google.com/spreadsheet/ccc?key=0AivVdWOTQE2JdGQzYkFSQUg2eUR3UVRTcThFYWpBelE#gid=21

[valueorama]

This is a very nice tabulation and lots of work went into it.

 

Personally, I like this site because I get to see the process of analysis more than the results. Yes results are important, I feel process drives results.

[boilermaker75]

Very nice, Thanks!

[LC]

Awesome, I love it. If a suggestion is possible...would it be feasible to add in a cell which notes the date & annualizes the results?

 

[mikazo]

Very nice! I've used Python/SQLAlchemy/Bottle before as well, and the type of things you can do with it are amazing! I'd love to see work on this continue!

[mcliu]

Wow, this is really cool. Thanks for sharing. :)

 

Is the return calculated from the date of the first post? Annualized return would be cool to see.

[watsa_is_a_randian_hero]

Extremely cool!  I had a suggestion to Parsad a while back that the board should somehow incorporate feedback so a viewer could quickly sort posts (because there are so many) to find the ones that would be considered most valuable (from the perspective of that viewer).  This would be very helpful for that! 

 

My original suggestion was something where other viewers rated (thumbs up/down) posts and posters.  One issue with basing feedback merely on thread creator is that multiple posters can be long the same idea and contribute to the analysis on the thread, but only the thread creator gets analysis under this construct.  Tough to solve.

 

Wow, this is really cool. Thanks for sharing. :)

 

Is the return calculated from the date of the first post? Annualized return would be cool to see.

 

May I +1 this, and also suggest adding the S&P 500 return since the first post as well?

[Parsad]

Very neat idea...I like it!  But I don't think it's accurate other than the germination of ideas.  If you go simply by the rankings, Ericopoly has made zero contributions and mine would be Chanticleer!  ;D

 

Maybe some tweaks...number of posts in a specific thread gets some weighting relative to the stock price at that point in time, since most people would not have held on to their BAC or especially BAC warrants without Eric's analysis and comments.  Otherwise the results indicate that Eric is perhaps the greatest coat-tail investor in history...which is not accurate at all.  Cheers!

 

 

[twacowfca]

Very neat idea...I like it!  But I don't think it's accurate other than the germination of ideas.  If you go simply by the rankings, Ericopoly has made zero contributions and mine would be Chanticleer!  ;D

 

Maybe some tweaks...number of posts in a specific thread gets some weighting relative to the stock price at that point in time, since most people would not have held on to their BAC or especially BAC warrants without Eric's analysis and comments.  Otherwise the results indicate that Eric is perhaps the greatest coat-tail investor in history...which is not accurate at all.  Cheers!

 

Actually, if I'm not mistaken, Eric has said that he doesn't have original ideas, but simply identifies the most compelling ideas that others have come up with.

 

Apparently, good ideas like pointing out the merits of  a whole class of stocks that are posted under the General category apparently won't make the list and shorts or stocks with downsides that are posted under investments may  appear to be failures.

 

Nevertheless, the results are interesting.

 

 

[Parsad]

Very neat idea...I like it!  But I don't think it's accurate other than the germination of ideas.  If you go simply by the rankings, Ericopoly has made zero contributions and mine would be Chanticleer!  ;D

 

Maybe some tweaks...number of posts in a specific thread gets some weighting relative to the stock price at that point in time, since most people would not have held on to their BAC or especially BAC warrants without Eric's analysis and comments.  Otherwise the results indicate that Eric is perhaps the greatest coat-tail investor in history...which is not accurate at all.  Cheers!

 

Actually, if I'm not mistaken, Eric has said that he doesn't have original ideas, but simply identifies the most compelling ideas that others have come up with.

 

Apparently, good ideas that are posted under the General category apparently won't make the cut, and shorts or stocks with downsides that are posted under investments would muddy the water.

 

Nevertheless, the results should be interesting.

 

That's my point.  People always come up with an idea, but what do you do with that idea?  How many would have sold BAC when it hit $5, rather than buy BAC at that price?  I think the tool has no way of indicating results based on the actual most opportune time to purchase, and instead simply shows when and whom germinated the idea.  That in itself is fascinating, but I think undermines what the true goal is...to get a real gauge for investment ability and execution.  It's why Eric's results are stratospheric...combine his leverage/concentration with his ability to buy at the most pessimistic of times...you get extraordinary results. 

 

If the tool could somehow record the number of posts by an individual relative to the historic prices and then weight that relative to the future outcome of the investment, that may provide a more accurate gauge.  I'm sure that is a hell of a lot more complicated...but I wouldn't want investors suddenly weighting their investments based on the tool as is...I think they may find the end result different than what the historical numbers may indicate...as there is no weighting for conviction and temperament.  Cheers! 

[Guest hellsten]

Thank you. Very interesting data. The forum needs something like this that summarizes and presents the investment ideas in a better way than this page:

http://www.cornerofberkshireandfairfax.ca/forum/investment-ideas/

 

Although you've already spent time worth thousands of dollars on this, the following additions would be nice to see:

- list of names and how many times they are mentioned, e.g. Francis Chou, Chou count as Francis Chou and Warren, Buffett, Warren Buffett, Warren Buffet count as Warren Buffett

- export to Excel

- a page that displays the data in a sortable table

- if it makes sense, additional data such as P/B, market cap, top 5 shareholders

- amount of replies listed next to returns. Packer16's threads get little replies, but returns are excellent.

[petec]

Also...is this total return or just share price?

[Hielko]

Nice job on the scraper, and too be honest I'm not surprised that the best ideas have few replies. The stocks with the most replies are almost always large cap names such as Apple, Microsoft: hard to have a significant edge in those names. They get a lot of replies because everybody knows those companies.

[Uccmal]

Amazing what can be done with data.

 

For the next project.  Go back to the inception of the board and determine how many pages discuss macro economics relative to investment ideas, how many were right, and whether there is any point to it all......

 

Like Sanj., and Eric, I haven't had an original idea in years.  On the old board I posted dozens, and it wasn't pretty. 

[Sportgamma]

Damn you Aperam Stainless Steel!

[prunes]

Thanks for all of the kind and constructive feedback guys.

 

@mcliu: Yes, returns currently use the initial posting date as a starting point.

 

@petec: Returns were calculated using Yahoo's adjusted closing prices, which are supposed to account for splits and dividends.

 

@Parsad: The point you raise is correct and I tried to address this in the FAQ. I think it is wrong for someone to interpret these results as providing basis that one person is a better investor than another. Hopefully you guys find some utility to this project this limitation not withstanding.

 

@hellsten: Great suggestions, thanks.

 

Comparing against the S&P appears to be the most frequent ask and should be relatively easy to do.

 

I would love to do textual analysis and/or correlate particular users' activity in particular threads. However, I think this would be difficult to do using a webscraping approach due to the amount of HTTP queries one would have to make. To do this right I would think someone would need access to the existing backend. However if anyone here thinks differently and would like to collaborate please do reach out to me.

[blainehodder]

Any idea on the average claimed returns of boardmembers?  Has anyone's internet "returns" underperformed in the history of the board?

[bookie71]

The problem is I never bought MIG, I only asked if anyone had thoughts on it, and referred to a quarterly earnings report. On CRVP I got out a long time ago before it dived so much. 

[onyx1]

I think the tool has no way of indicating results based on the actual most opportune time to purchase, and instead simply shows when and whom germinated the idea.  That in itself is fascinating, but I think undermines what the true goal is...to get a real gauge for investment ability and execution.

 

Agreed.  The real value of this board comes from understanding the dozen or so individuals that have real insights and well-considered recommendations.  These happen at opportune times on threads that are often already created.  One example (of many) is Twacowcfa who doesn't show up on the list, yet has offered some very high quality recommendations including MMPIQ, WSBASE market short, GSE preferred, or the epic call to lever BRK when it was trading at the buyback price. 

 

[Parsad]

I think the tool has no way of indicating results based on the actual most opportune time to purchase, and instead simply shows when and whom germinated the idea.  That in itself is fascinating, but I think undermines what the true goal is...to get a real gauge for investment ability and execution.

 

Agreed.  The real value of this board comes from understanding the dozen or so individuals that have real insights and well-considered recommendations.  These happen at opportune times on threads that are often already created.  One example (of many) is Twacowcfa who doesn't show up on the list, yet has offered some very high quality recommendations including MMPIQ, WSBASE market short, GSE preferred, or the epic call to lever BRK when it was trading at the buyback price.

 

Yes, totally agree.  How is Twacowcfa not on the list?  You have some of the most successful investors and prolific analysts excluded from the scraper.  I think like any tool or ratio, it gives a static impression on very strict parameters.  Boardmembers should view it in that light, and I'm sure Chris will continue to refine the scraper over time.  Cheers!

[ItsAValueTrap]

The scraper is very cool.  Unfortunately I think it will be very difficult to overcome certain limitations.

 

- Sometimes ideas have a very specific timeframe or the author may recommend an exit of the stock.

- Some ideas may be short ideas.  Some aren't ideas at all... the person simply wants to know about a certain company.

- Dividends and spinoffs can be tracked, but you'll probably need to use another data source.

 

Perhaps look at valueinvestorsclub.com  Its tracking system is more mature.  I'd be curious if you scraped data from there and looked at the performance of VIC.

[prunes]

You all raise great points. I certainly don't want to misrepresent anyone. In the site's current incarnation, it does imply that someone could have advocated or purchased a security that they did not. The optimal solution therefore has to be more nuanced than what I put together. Unfortunately I don't know if there is a feasible way to do this.

 

Off the top of my head, the following options exist:

 

i) Self reporting

ii) Manually checking each thread

iii) Some sort of removal feature (although most people don't want to email me asking me to fix things nor am I eager for the administrative burden)

iv) There's also the idea of using some sort of weighting mechanism based upon activity within any given thread. I don't know whether that's something that is feasible for me to implement.

 

None of these solutions seem like good ones to me.

 

So I guess I'd put the question back to everyone. Is the project worth salvaging? If so, I'd love for anyone with a programming background to PM me on how they would approach it.

[Hielko]

Things don't have to be perfect to be interesting. You can't make it perfect because you simply lack a lot of data that isn't available.

[woltac]

Prunes,

 

I love your scraper tool.  It makes me look like an all star investor.  My one pick is up 71% in two years!

 

woltac