He also uses a number of other studies which I dont understand(regressions, runs and turning point tests, arima estimates, variance ratio tests, and non-linear extensions), I'm still waiting for Covel to address these instead of hiding behind his typical 'he blew up'
I believe reflexivity exists, it is a market inefficiency and trend following is a great strategy to exploit it. That's why hopping on every trend worked (historically): folding many losing trades and letting a few big winners ride. Covel does a great job of illustrating the fundamental idea: accepting many small losses, and riding a few big winners until the very end. The simlarities to poker by the way are fascinating. Folding many bad hands for a small loss and betting the right amount on the big hands. This is counter-intuitive to most. In poker just as in trading.
Why doesn't VN instead address those 100 SSRN papers discussing momentum profits? How many scientific papers of VN have been accepted substantiating his hypothesis?
As far as I know, he agrees that some systems were able to capture trends in retrospect like those studies, he just doesnt think that can keep on going as markets tend to change. It appears to me that he is making a distinction between permanent edges and temporary edges. His obscure statistical studies seemed to be trying to find permanent edges, while the SSRNs just backtested things and found something that works for a certain period and it remains to be seen if it will in the future
Just finished reading chap 3 of Practical Speculation. Its on technical analysis, he says his quant teams tested just about every indicator and they werent profitable. Also it included a number of 'patterns' that TA books like to promote and they typically failed. The NYFed tested the H&S pattern and it did not produce profits, the counter argument from some ET mod was that you were supposed to short before the neckline broke, he is arguing one needs to trade before the pattern has been formed, so the statement 'H&S pattern does not work' is actually correct even if his method works
I do find it very interesting how that Anekdoten guy could claim to make tons of profits trading a historically mean reverting asset(S&P500 futures) with a TF system. But that is just part of it, to me some red flag were his claims -Only having 7 losing days in 2007 with a hybrid system that scalped and did intraday swings at the same time -Saying he is a 'failed' systematic trader because he could not codify his method and therefore no one could test it(specially given that his method was so simple, I'm talking with ET quant ironfist about this in the moment) Interesting, how most of these gurus back off laying out their methods in a specific way that allows it to be shown to not work. I have no idea if he is a liar, lucky or the real deal but I find it a strange 'coincidence' how these daytrading gurus cant codify their magic
Chap 6 on Ben Graham argues value investing doesnt work. "In 1965, prompted by gains in value stocks and by Samâs desire to offer clients an alternate rating system, Value Line started selecting the best value stocks in their universe. The firm divided its sample of 1,500 companies into 10 groups based on value, updating the selections each month. The selection system used the same key variables that almost everyone else uses to define value stock (i.e., price to sales, price to book, and price to earnings). But there was one difference that today, almost four decades later, makes all the difference in evaluating the stocksâ long-term performance. In contrast to retrospective studies of value stocks, the Value Line companies were selected contemporaneously. Thus, the Value Line value stock performance record is a real-life unintentional experiment with no back-testing, survivor bias, or armchair quarterbacking. The procedure has remained the same since 1965" "We report the results of the experiment in Figure 6.1 and Table 6.2. Value Line Group 1, consisting mainly of growth stocks, quickly outpaced the three value stock categories. By March 2002, the growth stocks had returned almost 28 times the return of the best-performing 'value' group'" Even though I dont practice value investing what originally prompted my interest on the subject was this http://www.tweedy.com/resources/library_docs/papers/WhatHasWorkedInInvesting.pdf It seems that VN is claiming those studies are flawled. Some of the studies do suffer from selective time periods but some of them are free from survivorship bias and some others are quite broad in terms of time. I have no idea what VN means by 'armchair quarterbacking'. At this point I'm uncertain about what to make of the Tweedy studies
One thing I'm sure though is that if someone is a great stock picker he can outperform the averages by doing 'value investing'. Its possible that they tend to go up less than growth in the long-run(if the Value Line study is accurate) but they also tend to be easier to understand, high valuation stocks tend to be of the technology types which are hard to predict And this why the best years of Warren Buffett were in the 50's when he was doing 'cigar butt' investing, buying garbage stocks that went too low and selling when they reached a resonable level. Now maybe the market has changed and those opportunities are harder to come by but as I understand Buffett didnt give up that practice because Munger suggested it didnt work(as VN claims in his book) but rather it was because they had to manage greater sums of money and in larger caps those opportunities were scarce Monish Pabrai does the cigar butt value investing because he says he is mimicking the 50's Buffett, the guy knows buffett better than his wife, I find it doubious he would make a mistake of trying a strategy buffet did not think it works. And I seem to recall another buffettologist sending a letter to Buffett asking which one was better today(cigar butts or great companies at good prices), Buffett stated that 'either one was fine'. So it appears the VN made a mistake in his book by suggesting Buffett stopped garbage businesses at too depressed valuations because "Charlie Munger pointed out to Buffett, such companies tied up capital endlessly, consistently lost money, and were never able to be sold at higher prices than Buffett had paid for them, because nobody else wanted them."
I remember reading about Buffet that back in the 50s and 60s he did a lot of risk arbitrage, i.e. buying merger and acquisition companies after the buy out announcement, capturing the spread. Buffet was saying that back then he had less capital, so he could participate in many of the most attractive smaller mergers with bigger spreads which he can't now, managing multi-billions in assets. Doesn't mean it's impossible to do though. Daniel Loeb has been in many merger arb positions (as a value investor with $1-$2bln under management, obviously a fraction of Buffet's play money these days), according to his monthly fillings. Loeb wrote in a quarterly letter in 2007 (or 2008?) that he always waits for sell offs/market dislocations before he engages in big leveraged merger arb trades, because then the spreads blow out, leaving many opportunities on the table as other leveraged players have to liquidate positions to meet redemptions. Another manager who did merger arb from day one is John Paulson. Prior to founding his own firm Paulson was a partner at a merger arbitrage hedge fund. I remember reading articles he was in ROH (the buyout that nearly sunk DOW) and now also is in Cadburry (which has an offer from KFT): http://www.guardian.co.uk/business/2009/nov/11/paulson-raises-cadbury-stake IndexIQ is launching a MergerArbitrage ETF this month I believe www.indexiq.com Gabelli is running a closed-end merger arb fund, NYSE: GDL Finally, SSRN has a gazillion papers on merger arb & the efficient market hypothesis.