Blog Week in Risk (February 18th)

[David Harper CFA FRM]

Some of the new content this week

• New P1 Practice Question: P1.T4.802. Stress Testing and Other Risk Management Tools (Siddique) https://forum.bionicturtle.com/thre...d-other-risk-management-tools-siddique.13507/
• New P2 Practice Question: P2.T9.803. Machine learning (van Liebergen) https://forum.bionicturtle.com/threads/p2-t9-803-machine-learning-van-liebergen.13513/ As a practicing data scientist, I think GARP selected two good data science readings, although I think it's impossible to really engage these topics without writing at least some code
• We updated the FRM Historical Pass Rate chart to reflect November's results https://forum.bionicturtle.com/threads/what-is-the-pass-rate-for-the-frm.10093/#post-57919

We are building out our free youtube channel playlists. Here are the FRM playlists (our sequence approximately follows the syllabus):

• Youtube T1 Foundations playlist http://trtl.bz/bt-yt-playlist-frm-t1-foundations
• Youtube T2 Quantitative playlist http://trtl.bz/bt-yt-playlist-frm-t2-quant
  
• Youtube T3 Products playlist http://trtl.bz/bt-yt-playlist-t3-products

Selected discussion in the forum this week

• [P1.T1] You do want to understand why a bank's balance sheet is inherently fragile to drops in asset value, and the implications of maintaining constant leverage https://forum.bionicturtle.com/thre...e-global-financial-crisis-brunnermeier.10076/
• [P1.T4] More good questions about hybrid (aka, age-weighted) value at risk (VaR) https://forum.bionicturtle.com/threads/calculating-revised-var-hybrid-approach.9857/
• [P1.T4] I'm glad @nikogeorgiev asked this question, as you should be able to find the variance of EDF from first principles https://forum.bionicturtle.com/threads/variance-of-edf.13514/
• [P1.T4] Thank you to @lRRAngle for helping me to further refine my explanation of GARCH(1,1) https://forum.bionicturtle.com/threads/garch-1-1-model-long_run_volatility.10509/#post-57804
• [P2.T5] I finally replaced two of the old fixed income mapping questions (T5.63) and simultaneously we've had some new discussion; it's not easy to grok Jorion's three mapping approaches https://forum.bionicturtle.com/threads/l2-t5-63-fixed-income-mapping-jorion.3617
  
• [P2.T5] @Stella.gkotsi found a typo in Meissner's text (that we missed), nice! https://forum.bionicturtle.com/thre...otes-p2-t5-market-risk.8758/page-3#post-57962
• [P2.T5] The simplified modified duration for par bonds and, related, yield-based duration implicitly assumes a parallel shift in the spot rate curve (even if the spot rate curve is not flat) https://forum.bionicturtle.com/threads/fixed-income-mapping-jorion.13511/#post-57893
• [P2.T6] Continuous versus discrete solution to the risk-neutral default probability https://forum.bionicturtle.com/threads/p2-t6-301-default-probabilities-malz.6773/#post-57950

Some external links I found to relevant to the FRM, or just interesting

• [Market correction triggered Feb 2nd]: I continue to appreciate Matt Levine at Money Stuff, he's written some smart commentary on the correction over at https://www.bloomberg.com/view/topics/money-stuff
• [Market correction] Also here is Matt gloriously on inverse volatility products https://www.bloomberg.com/view/articles/2018-02-09/inverse-volatility-products-almost-worked
• [Market correction] Harvard, Hawaii Gambled on Market Calm—Then Everything Changed (wsj) http://trtl.bz/2HryeaB
• [GARP] How have banks fared in the era of Dodd-Frank? http://trtl.bz/2sENhKM
  
• [governance] 2018 Proxy Season Preview https://corpgov.law.harvard.edu/2018/02/14/2018-proxy-season-review/
• [P1.T1] The new ISO 31000 keeps risk management simple https://www.iso.org/news/ref2263.html
• [P1.T1] 10th Edition of Global Fraud & Risk Report by Kroll https://www.kroll.com/en-us/global-fraud-and-risk-report-2018 Copy here at http://trtl.bz/kroll-10th-ed
• [P1.T1] Worldwide Threats Briefing: 5 takeaways from Russia to China https://www.wired.com/story/worldwide-threats-briefing-russia-election-china/
• [P1.T3] Primer on make-whole calls (BofA ML) http://trtl.bz/0218-make-whole Make-whole call provisions are discussed in FRM assigned (T3.R23) Fabozzi on Corporate Bonds
• [P1.T4] Primer on duration and convexity (BofA ML) http://trtl.bz/0218-primer-duration-convexity Nothing here new (or deeply technical) to FRM syllabus but a fine summary
• [P2.T7] Basel III: Are we done now? (keynote speech) https://www.bis.org/speeches/sp180129.htm
• [P2.T9] Overseeing cyber risk https://corpgov.law.harvard.edu/2018/02/18/overseeing-cyber-risk/
• [P1.T8] Book Review: The End of Theory https://blogs.cfainstitute.org/investor/2018/02/15/book-review-the-end-of-theory/
• [P1.T8] Data Science Will Transform the Investment Industry: Are You Prepared? (CFA Institute) http://trtl.bz/2szX8kW
• [P2.T9] How Blockchain could disrupt banking https://www.cbinsights.com/research/blockchain-disrupting-banking
• [P2.T9] Artificial Intelligence, Machine Learning, and Deep Learning: A Primer http://trtl.bz/2ECowQM
• [P1.T9] Excellent summary of cryptocurrencies http://trtl.bz/0218-decrypting-cryptocurrencies
• [P1.T9] Crypto Canon by a16z https://a16z.com/2018/02/10/crypto-readings-resources/
• [data science] Six reasons to learn R for business http://www.business-science.io/business/2017/12/27/six-reasons-to-use-R-for-business.html As some members know, I've been training data science for over three years (using R). I have about 30 course certifications, mostly at datacamp (highly recommended!) and coursera. Two weeks ago, I attended the rstudio conference, where I participated in a two-day workshop conducted by the talented Charlotte Wickham (topic: data science in the tidyverse). I've already registered for next year's conference in Austin, TX!

 

[QuantMan2318]

Dear @David Harper CFA FRM

Glad to see that this is back after a while

What do you think is the better avenue? Machine Learning or Data Science in R? I have been juggling in my limited time with these two. I have taken many certifications from data camp after your guidance here, but can you suggest ways so that I can apply what I learnt and I won't forget?

Thanks

 

[David Harper CFA FRM]

Hi @QuantMan2318 Thank you for supporting WIR, I was trying to enjoy my Sunday for a bit there but from this point thru the exam I'll be working pretty much everyday so I think the WIR is back for now! So glad to hear you are learning on datacamp, do you like it? I met a few of them in San Diego at the rconf, such a cool company.

Re: What do you think is the better avenue? Machine Learning or Data Science in R? Well, I don't know. I've gone the "data science" route simply because, like you, I have limited time; and I want to record youtube videos with R, such that I feel like I will be able to do that (much) more proficiently with the more basic layers of skills (or, maybe I should say: my current path tends to be breadth vis à vis tidyverse, rather than depth). Maybe also I wonder if ML is a big mountain to climb, a major commitment (?). Ergo I haven't engaged deeply at all in ML. I just have a couple of other areas of intellectual interest; e.g., I am very interest in geographical/spatial visualizations (e.g., choropleth) . Maybe because I teach, just the visualization methods in EDA fascinate me.

The best advice that I think I've gotten so far, with respect to keeping skills sticky, is to simply start a blog. So I am going to do that ASAP (this forum just makes it very difficult), I think with blogdown; see https://bookdown.org/yihui/blogdown/

My other keen interest is time series, fwiw. At the conference, one of the super smart guys at Business Science reviewed their tibbletime (http://www.business-science.io/code-tools/2018/01/04/tibbletime-0-1-0.html ) and I'm really keen to sharpen my time series skills. Maybe see if i can find an excuse to use that in an investment capacity! Thanks!