Offbeat "value" idea, boxing?

[LC]

wonder if Mr. Ben Graham did not do the value crowd a disservice when he drew a line between investment and speculation

 

Not sure I would agree but perhaps a step down memory lane could be useful:

https://blogs.cfainstitute.org/investor/2013/02/27/what-is-the-difference-between-investing-and-speculation-2/

 

Even Graham in The Intelligent Investor came to accept the necessity of speculation. “Outright speculation is neither immoral, nor (for most people) fattening to the pocketbook. More than that, some speculation is necessary and unavoidable.” But Graham was quick to distinguish between “good” and “bad” speculation. “There is intelligent speculation as there is intelligent investing. But there are many ways in which speculation may be unintelligent,” wrote Graham.

 

I think I'd agree with the above. Occasionally the horse races may present a great opportunity but by-and-large the house is the winner.

[Guest cherzeca]

Isn’t that supposed to be the “good literature” on the Kelly criterion? Or do you want something more academic? In that case I’d simply start at the Wikipedia footnotes.

 

seems to me that there is less than meets eye re Kelly Criterion.  the math is remarkable, but the estimation of one's edge as compared to market odds is the crux, and this is of course in the eye of the estimator.  since I am not a math maven, reading Thorpe's scholarly stuff wont help me and so I am left thinking that like anything else, there is a garbage in garbage out problem...which likely applies to any system

[Gregmal]

Investing is just intelligent and disciplined speculating. Nothing more, nothing less. There is a thread of the wager inherent in any opportunity. If there wasn’t there would be no variance in rates of return.

 

The Kelley criteria is no different than any other ideological framework. Choose and apply as it suites you. No successful market participant should be married and/or permanently divorced from any specific strategy.

[writser]

Isn’t that supposed to be the “good literature” on the Kelly criterion? Or do you want something more academic? In that case I’d simply start at the Wikipedia footnotes.

 

seems to me that there is less than meets eye re Kelly Criterion.  the math is remarkable, but the estimation of one's edge as compared to market odds is the crux, and this is of course in the eye of the estimator.  since I am not a math maven, reading Thorpe's scholarly stuff wont help me and so I am left thinking that like anything else, there is a garbage in garbage out problem...which likely applies to any system

 

Yes, very true. It's way more suited for black-and-white stuff like blackjack. Kelly also assumes the bet is settled instantly and doesn't take into account opportunity costs. Both not true in the stock market.

 

Still, I occasionally like to tinker a bit with the Kelly criterion. Especially with binary situations like mergers / CVR's / regulatory approvals required (i.e. biotech). For example, assume a simple takeover bid at $10, stock trading at $9, pre-announcement it was trading at $5. You can plug these numbers in the Kelly formula, play around with a range of probabilities and tinker a bit with the downside in case of a deal break. That way you get a pretty good feel for position sizing. Look at the huge difference between optimal sizing when downside is very small vs. very large: basically the mathematical proof of Buffett's rule #1: never lose money.

 

A short read by the CFA institute about applying Kelly in practice in the stock market: link. I thought that was a decent primer.

 

The Kelley criteria is no different than any other ideological framework.

 

It's not an ideological framework. It's math .. If you play blackjack in the casino you can fool around with any 'ideological framework that suits you' but in the long run you will never beat somebody who uses mathematically proven optimal bet sizes. In the stock market it's a little different because, as cherzeca pointed out, the difficult part is estimating your edge. Still it is good to know the theory and not something ideological that you should discard if it doesn't work a few times in a row.

[Gregmal]

The difference between the blackjack cheat sheet and applying some mathematical cheat sheet to the markets is that the markets are always changing. The 52 cards in a deck don’t. One persons 42% odds are another persons 66% odds in the stock market... and both can be wrong. Whereas a 9/4 against a face card has the same “odds” every time.

 

I’d also add that even an indisputable %/odds(which is rare if not nearly impossible by itself) can change instantaneously, because of some black swan event coming out of left field, rendering the entire premise useless. Such as the for profit prisons a few years back. Over reliance on tone deaf/non malleable  mathematics is probably the single greatest area of stealth wealth destruction for otherwise generally intelligent folks in the stock market.

[writser]

Sure, if you misapply the formula, don't know its limitations or make garbage assumptions you can lose money. That doesn't mean that the math is wrong, it just means that you are being stupid. Unfortunately no amount of math can change that.

[Gregmal]

Sure, if you misapply the formula, don't know its limitations or make garbage assumptions you can lose money. That doesn't mean that the math is wrong, it just means that you are being stupid. Unfortunately no amount of math can change that.

 

If the inputs are subjective, and the variables ever changing, then it’s no different or more reliable than any other product or paper put together on how to “beat the market”...

 

But there’s nothing that will ever stop the Excel sheet crowd... it has zero to do with the math being wrong or right. It has to do with calculating things that aren’t linear or regularly constant in terms of their predictability. I mean Mr. Einhorn is still absolutely convinced that Amazon and Tesla can’t possibly fit into his 476 tab spreadsheet that justifies their valuations.. a decade later. And if you ask him, the “math” is telling him he s 1000% right...

 

Dillard’s, was an easy way to use math to make money. But you can’t openly apply that successfully without applying something subjective in determining whether the idea fits the criteria for application. And even then, there was the spreadsheet crowd, totally fixated on “they’re got bad operating metrics”.

 

Many of these folks are too married to their own preconceived notions and incapable of evolving with the markets. It’s to their own determinant not mine...but if you’ve cracked the markets with some mathematical formula.... more power to you.

 

[Guest cherzeca]

@writser

 

you're merger arb example was helpful, thanks.  I also have a somewhat binary investment and have tinkered around with it, but cant get beyond the point of thinking that my sizing options were really just a manifestation of my fundamental analysis (viewpoint driving the math) which may be overly biased...though I ask whether there is anything wrong with confirmation bias when one has put a lot of work into something...:>)

 

I do credit the Kelly Criterion for offering a serious attempt at avoiding gamblers/investors ruin, and I found it interesting that it was at the core of thorpe's blackjack system (increasing bets when odds become more favorable)

[LC]

In the world of investing, all the Kelly criterion tells you is that to maximize your expected wealth, you should concentrate in your best ideas. That is the only lesson.

 

In scenarios with finite solutions it can tell you how much you should concentrate, but these scenarios are rare outside of casinos and can be largely ignored for our case.

[Gregmal]

In the world of investing, all the Kelly criterion tells you is that to maximize your expected wealth, you should concentrate in your best ideas. That is the only lesson.

 

In scenarios with finite solutions it can tell you how much you should concentrate, but these scenarios are rare outside of casinos and can be largely ignored for our case.

 

In which case it’s really no more insightful than anecdotes like “buy low and sell high”. Giving greater weight to your highest conviction ideas is really just common sense. But it’s also again, subjective. Buffett, Icahn, and Tepper all have different top ideas. Some will do better than others. When all else fails, look at your results.

[writser]

@writser

 

you're merger arb example was helpful, thanks.  I also have a somewhat binary investment and have tinkered around with it, but cant get beyond the point of thinking that my sizing options were really just a manifestation of my fundamental analysis (viewpoint driving the math) which may be overly biased...though I ask whether there is anything wrong with confirmation bias when one has put a lot of work into something...:>)

 

Would you mind sharing the idea? I'm curious. PM is fine too. I have a small position in the Bristol Myers CVR. Trades at $3, pays out $9 in certain circumstances, worthless in all other cases. I think that that is a typical situation where you might want to tinker around a bit with the Kelly formula, just to get a feel for what an optimal position size would be if your assumptions are correct. What usually happens is that you think: I make these and these assumptions and I'm comfortable with a 2% position. Then you plug your assumption in the formula and it tells you you should increase your position tenfold. The interesting question then is: is your position sizing way too conservative or are your assumptions too optimistic? In practice I usually end up with a much, much smaller position though. What that implies I'll leave to the reader.

 

I do credit the Kelly Criterion for offering a serious attempt at avoiding gamblers/investors ruin, and I found it interesting that it was at the core of thorpe's blackjack system (increasing bets when odds become more favorable)

 

Back in the days I used to play a lot of poker. An old married couple of regulars at the table had some cool stories about even longer ago, when counting cards in single deck blackjack was allowed and profitable. Usually the blackjack tables were filled with a mix of card sharks and casual players. They'd all have some fun, banter a bit and play for minimum bet size. Then at some point the remaining cards in the deck were favorable and all the card sharks would suddenly switch to maximum bet size, i.e. suddenly bet 200 instead of 5 or whatever. The dealer obviously knew what was happening, wasn't fazed at all and continued bantering with the card sharks. At some point all the sharks switched back to minimum bet size and the game continued as if nothing had happened, even though some card sharks just lost or won a huge amount of money. Rinse and repeat. Apparently this really freaked out the casual players who didn't have a clue about what was going on. Would have loved to be there.

 

In the world of investing, all the Kelly criterion tells you is that to maximize your expected wealth, you should concentrate in your best ideas. That is the only lesson.

 

Not at all. If you find attractive 'lottery ticket' stocks or options and make some very basic assumptions about them the Kelly criterion will show you that you should keep your position sizes small, completely the opposite of what you are saying. The Kelly criterion also shows you that, in terms of position sizing, it is very important to look at downside protection if you are wrong, rather than solely looking at the upside if you are right. Something that a lot of investors tend to overlook. In general, I think the Kelly criterion can be a useful tool to challenge your own position sizing in certain scenarios. I think that that is a useful exercise every now and then, to avoid getting too attached to your own preconceived notions about sizing. The alternative is solely relying on gut feeling and experience.

[Jurgis]

@writser

 

you're merger arb example was helpful, thanks.  I also have a somewhat binary investment and have tinkered around with it, but cant get beyond the point of thinking that my sizing options were really just a manifestation of my fundamental analysis (viewpoint driving the math) which may be overly biased...though I ask whether there is anything wrong with confirmation bias when one has put a lot of work into something...:>)

 

Would you mind sharing the idea?

 

I am pretty sure he means the longest thread on CoBF:

https://www.cornerofberkshireandfairfax.ca/forum/general-discussion/fnma-and-fmcc-preferreds-in-search-of-the-elusive-10-bagger/14520/

 

I've done Kelly on this couple years ago.

 

The key part here is couple years ago.

As you've observed:

 

Kelly also assumes the bet is settled instantly and doesn't take into account opportunity costs. Both not true in the stock market.

[Jurgis]

A short read by the CFA institute about applying Kelly in practice in the stock market: link. I thought that was a decent primer.

 

LOL. I realized the same thing this guy did: that a lot of places use bad incorrectly simplified formula. There are some places that have correct one that I also used.

I think I wrote about it on CoBF at some point. But not as nice an article as CFA institute did.

 

At least I feel smart and corroborated. 8)

And I now have a link to give anyone who uses wrong formula without writing an article myself.  8)

[Guest cherzeca]

@writser

 

you're merger arb example was helpful, thanks.  I also have a somewhat binary investment and have tinkered around with it, but cant get beyond the point of thinking that my sizing options were really just a manifestation of my fundamental analysis (viewpoint driving the math) which may be overly biased...though I ask whether there is anything wrong with confirmation bias when one has put a lot of work into something...:>)

 

Would you mind sharing the idea?

 

I am pretty sure he means the longest thread on CoBF:

https://www.cornerofberkshireandfairfax.ca/forum/general-discussion/fnma-and-fmcc-preferreds-in-search-of-the-elusive-10-bagger/14520/

 

I've done Kelly on this couple years ago.

 

The key part here is couple years ago.

As you've observed:

 

Kelly also assumes the bet is settled instantly and doesn't take into account opportunity costs. Both not true in the stock market.

 

seems to me that you can create binary outcomes as a simplifying heuristic on almost any investment...set a stop loss low and a sell-take profits high, and then assign probability values to each...again this is is highly subjective, but it seems that if Kelly is applicable here it is applicable almost anywhere...and if it is only a buttress to common sense, then that can be of utility

[LC]

 

In the world of investing, all the Kelly criterion tells you is that to maximize your expected wealth, you should concentrate in your best ideas. That is the only lesson.

 

Not at all. If you find attractive 'lottery ticket' stocks or options and make some very basic assumptions about them the Kelly criterion will show you that you should keep your position sizes small, completely the opposite of what you are saying. The Kelly criterion also shows you that, in terms of position sizing, it is very important to look at downside protection if you are wrong, rather than solely looking at the upside if you are right. Something that a lot of investors tend to overlook. In general, I think the Kelly criterion can be a useful tool to challenge your own position sizing in certain scenarios. I think that that is a useful exercise every now and then, to avoid getting too attached to your own preconceived notions about sizing. The alternative is solely relying on gut feeling and experience.

 

 

I should have been more clear: "Best" being defined as a scenario with (1) a higher probability of winning and (2) a higher payoff ratio.

 

A lottery ticket has (2) but not (1). An investment with both of those characteristics (i.e. maximum expected value) should be the higher priority.

 

It is very useful to break out investment opportunities into these 2 factors and do your best to think about them objectively.

"What is the chance I am right or wrong here?"

"If I am right, how much am I going to make?"

 

In the world of investing, all the Kelly criterion tells you is that to maximize your expected wealth, you should concentrate in your best ideas. That is the only lesson.

 

In scenarios with finite solutions it can tell you how much you should concentrate, but these scenarios are rare outside of casinos and can be largely ignored for our case.

 

In which case it’s really no more insightful than anecdotes like “buy low and sell high”. Giving greater weight to your highest conviction ideas is really just common sense. But it’s also again, subjective. Buffett, Icahn, and Tepper all have different top ideas. Some will do better than others. When all else fails, look at your results.

Yes it is common sense and all Kelly did (really Bernoulli before him) was quantify it where you can explicitly define the odds. As you correctly mention it is subjective in other areas outside of a casino because in these areas we cannot discretely identify the odds.

 

You can make assumptions to help (i.e. measure various historical volatilities to define both p and odds) but now you are transferring the subjectivity not to the odds themselves but how you are measuring the historical odds. This is what cherzeca alludes to above or more accurately is essentially a barrier option. Pricing barrier options is  how the financial industry has attempted to re-define and price these scenarios using different probability distributions/density functions. Slightly more accurate but still subjective:

 

E.g. if we are both taking a side on a 1-month vanilla and I am using historical monthly local vol and you are bootstrapping using weekly or daily or even hourly, well then we will be on different sides of the trade all else equal.

[Cigarbutt]

1-On the quantitative side of the equation(s)

Fancy math may give a false sense of precision and the idea is it to incorporate all potential outcomes (good and bad). The same principle applies in weighted scenarios analysis in certain circumstances when it will take some time to see the outcome (ie launching a new product, change of strategy, what happens if interest rates rise, who will get elected etc). In Mr. Hagstrom's The Essential Buffett, there is a nice little section on the Kelly Optimization Model and the simplified equation is 2p-1=x, with p being the probability of 'winning' and x the allocation that should be devoted to the specific investment. It is then only a simplified tool to evaluate how one is confortable with the potential outcome (expected value) and the standard deviation of the outcome and how this can be incorporated into one's portfolio. In the 3-page section, the author describes how one may decide to use a fraction of x, as in the half-Kelly or fractional-Kelly model (more on that in 3-).

 

2-On the mysterious investment

If the investment is GSE-related, an area I consider myself too stupid to make a reasonable assessment, when I tried to come up with a range of potential outcomes, I used a series of critical steps and put a reasonable % of being right for each step. If one establishes five or six critical steps, even if one uses an 80% chance of a positive outcome for each step in the sequence, one ends up with a very low % of a positive outcome (and a low weighted expected value) after all steps, an analysis that, given my significant analytical limitations, resulted in putting the investment thesis in the too hard pile. I'm looking to be educated here.

 

3-Going back to the initial spirit of the thread and position sizing

Both Mr. Mayweather and Mr. McGregor had obvious comparative edges. 1)The bestial factor. Anybody who had watched an ultimate octogonal fight would be impressed with the importance of raw power and grit. Advantage McGregor. 2)The age factor. A lot of 'fans' felt that this was a fundamental factor. Mr. Mayweather was quite old for a professional boxer and the typical scenario involves an exponential decline of ability when the time has come. Advantage McGregor. 3)The you play my game factor. The fundamental factor had to do with the fact that Mr. McGregor had to play according to Mr. Mayweather's rules. This would be like forcing a hedge fund, or hedge funds or funds of funds to enter a value contest where they would play the value game for a pre-defined long-enough period against an older (and apparently) declining old-style competitor. Of course, we know how this ended and the Master likely put only limited effort into the contest. However, even if, especially in retrospect  :), this appears to be all too clear, Mr. Sigsbee, who appears to be a master at his own game and who felt that he possessed an edge for both evaluating the handicap and the unsophisticated millennial 'fan' factor, only made a relatively small bet on his friend, suggesting that he likely uses quite a fractionated Kelly model.

[Gregmal]

3-Going back to the initial spirit of the thread and position sizing

Both Mr. Mayweather and Mr. McGregor had obvious comparative edges. 1)The bestial factor. Anybody who had watched an ultimate octogonal fight would be impressed with the importance of raw power and grit. Advantage McGregor. 2)The age factor. A lot of 'fans' felt that this was a fundamental factor. Mr. Mayweather was quite old for a professional boxer and the typical scenario involves an exponential decline of ability when the time has come. Advantage McGregor. 3)The you play my game factor. The fundamental factor had to do with the fact that Mr. McGregor had to play according to Mr. Mayweather's rules. This would be like forcing a hedge fund, or hedge funds or funds of funds to enter a value contest where they would play the value game for a pre-defined long-enough period against an older (and apparently) declining old-style competitor. Of course, we know how this ended and the Master likely put only limited effort into the contest. However, even if, especially in retrospect  :), this appears to be all too clear, Mr. Sigsbee, who appears to be a master at his own game and who felt that he possessed an edge for both evaluating the handicap and the unsophisticated millennial 'fan' factor, only made a relatively small bet on his friend, suggesting that he likely uses quite a fractionated Kelly model.

 

Or.....enter black swan... not unheard of in the world of professional boxing. The fight is fixed and Mayweather takes a flop, and all of our observable data points were irrelevant. Those that wagered correctly, lose big. And then you get "the rematch", which pays Floyd and Conor for round 2! Often following the money/financial incentives is pretty useful as well. In which case McGregor's weighted odds would have been higher than the ones purely based around the data set above. Which one is the correct one? If this happens 100 times, what is the distribution? Subjective and interesting nonetheless.

[Guest cherzeca]

"2-On the mysterious investment

If the investment is GSE-related, an area I consider myself too stupid to make a reasonable assessment, when I tried to come up with a range of potential outcomes, I used a series of critical steps and put a reasonable % of being right for each step. If one establishes five or six critical steps, even if one uses an 80% chance of a positive outcome for each step in the sequence, one ends up with a very low % of a positive outcome (and a low weighted expected value) after all steps, an analysis that, given my significant analytical limitations, resulted in putting the investment thesis in the too hard pile. I'm looking to be educated here."

 

this is not how investing works in a political/legal/financial matrix.  if you have 5 "independent" variables of 50% probabilities then you would never invest...but suppose the positive occurrence of one variable favorably affects the odds of the next variable...ie the variables are not truly independent, but are influenced by the other variables outcomes, such that a 50% variable in the end of a line of time sequenced variables increases substantially above 50% if the outcome of an earlier sequenced variable comes out positive.

 

capiche?

[Spekulatius]

I still think that even if the absolute odds or probability is not knows, the Kelly formula is reasonably to determine

A) Maximum bet size

B) sizing bets relative to each other

 

Apparently, the input size are based on estimating the odds, probabilities for each investment and each investor cannot have a precise estimate of each, but they should have an estimate of odds and probabilities nonetheless, as it it better than following the guts so to speak. At least in my opinion it is.

[Cigarbutt]

"2-On the mysterious investment

If the investment is GSE-related, an area I consider myself too stupid to make a reasonable assessment, when I tried to come up with a range of potential outcomes, I used a series of critical steps and put a reasonable % of being right for each step. If one establishes five or six critical steps, even if one uses an 80% chance of a positive outcome for each step in the sequence, one ends up with a very low % of a positive outcome (and a low weighted expected value) after all steps, an analysis that, given my significant analytical limitations, resulted in putting the investment thesis in the too hard pile. I'm looking to be educated here."

 

this is not how investing works in a political/legal/financial matrix.  if you have 5 "independent" variables of 50% probabilities then you would never invest...but suppose the positive occurrence of one variable favorably affects the odds of the next variable...ie the variables are not truly independent, but are influenced by the other variables outcomes, such that a 50% variable in the end of a line of time sequenced variables increases substantially above 50% if the outcome of an earlier sequenced variable comes out positive.

 

capiche?

Devo anche dire che non capisco and that's one of the reasons why investing can be fun. You may (probably? :) ) have developed an informational/analytical edge that is not available to me, at least up to now. The last time I looked more seriously into this (two links below), it seemed that the determinant variables may be indeed correlated and key building blocks (most? all?) had been put in place. However, even if, as always, there is unsophisticated money involved, I find there is a lot of educated competition and the timeline could be extended. The odds may reflect the outcome if you see this investment through a bettor prism, which you should IMO at least to a degree. Enough said on a topic that is being well covered elsewhere. This is worth watching but I may end up being simply a witness of an eventual cheery consensus and that's fine. Good luck.

https://www.jchs.harvard.edu/sites/default/files/harvard_jchs_gse_reform_none_or_done_layton_2019_0.pdf

https://www.investmentbank.barclays.com/content/dam/barclaysmicrosites/ibpublic/documents/our-insights/gse/gse-reform.pdf

 

Reading Fortuna's Formula, you likely met Mr. Claude Shannon. In a related way and through his 'collaboration' with Mr. Thorpe, one of the informational theory insights that Mr. Shannon identified is that, from a Kelly criteria perspective, you have to stay in the game long enough (avoiding "gambler's ruin") to harvest the eventual gain related to the area of the curve within the calculated risk zone, given numerous 'transactions' and the law of large numbers. Events that occurred rarely or over long periods of time required an unusual confluence of assumptions.

 

In a related way perhaps, I took into consideration the comments that you provided in the potential PG&E bankruptcy thread. Less than a year ago, you mentioned (in relation to a seekingalpha article detailing potential outcomes): "this is a good discussion of possibilities, and they range from some equity value to some debt value to don't touch with a 10 foot pole.  this will be a firestorm is bk court as well, and until we know what California wants to do, I don't think this case is investable". I thought this was interesting (warning: limited value because anecdotal) since I felt that an absence of clear vision from the government was actually a favourable factor in the delineation of an entry and exit point during the fall of 2019 (with the help of Kelly criteria for position sizing). Sometimes, there is a difference between what a government wants to do and what is eventually done and the timeline can occasionally be challenging. I'll be looking forward to learn.

 

-----)Back to the spirit of the thread

 

 

[Guest cherzeca]

"I felt that an absence of clear vision from the government was actually a favourable factor in the delineation of an entry and exit point during the fall of 2019 (with the help of Kelly criteria for position sizing). Sometimes, there is a difference between what a government wants to do and what is eventually done and the timeline can occasionally be challenging. I'll be looking forward to learn."

 

bravo, I find that I agree with you (now), though in my defense I sense that there is a difference between what I earlier discerned re PG&E as absolutely uncertainty as to govt response (I am not a big fan of California politics in relation to a private corporation's dealing with constituents) and what I have sensed all along as to what the fed govt will do with GSEs (There Is No Alternative).  but to extent the uncertainty can be embraced, it is a feature and not a bug