Market is Thrusting Out of the Triangle (And Friday Could Be Wild)

An update on the running triangle of a couple of days ago: the thrust out of it has begun a new plateau between 1116 and 1127 (Dow 10480 to 10590) which EWP argues is resolving into a triangle that may have ended mid-day today and be in its final thrust. EWP expects Friday to be wild:

The pattern is pretty clear in the Dow, although not textbook. The implication is the thrust should go about the widest part of the triangle, or 100 pts, targeting Dow 10640. 

Curiously the S&P lagged and has not made a new high (above 1130). Its target sits right at 1130, completing a triple top back to the August and June tops. 

Several convergences may make Friday a bit wild. It is options expiration for a start. There are divergences across indices right now, beyond the one noted above about the S&P vs Dow. And as Walter Murphy notes in his newsletter, bullish sentiment has gone back over 50% of AAII analysts in a mere three weeks! It took five months off the 2009 low. Kenny is all over this, and provdes this chart where you can see the huge blue spike of sentiment during this September rally:

Comments

[HighPlains]

What was more impressive than the 50% AAII bulls measure was not only how fast it got there in three weeks, but the spread b/w bulls and bears is one of the largest in the series dating all the way back to 1987.

However, one should be careful not to read too much into a chart that only goes back three years. There were plenty of times in the 1990s when the AAII bulls measure was reached the 70%s.

[home for sale]

No one can predict the flow of the market but it is alarming if it goes up fast then after a while it will also drop.

[Dsquare]

Its drawn as a BS triangle. The triangle should be drawn from the b to the d peak. Its the b-d line that's important.

[Wave Rust]

flash crash up ?
and guess what's shrinking ,,,, go ahead

it ain't bullish sentiment

trade a couple more decades and you'll get to see how often sentiment is right, as opposed to when it's wrong.

back to shrinkage and stinkage ,,,, smells like burning bear hair

A=flash crash
B=bounce
C was abcde and the e finished at 1040 Aug 31

If the correction ain't over (like if it was an intermediate A), then get ready for a big old B for a flat.

either way it's bullish ,,,, unless, of course, i'm wrong

ok, lets act bullish for little while. if not, i'll just let ya talk amongst yourselves about those good old down days last month.

wave rust

dubl posted cus i forgot where it was

[jim]

loosersssssssss short short????

sp 1137 ha ha ha

[Edwin]

Mr. Partridge says "“You know, it’s a bull
market!”

short term bullish..of sure..No shorting until it fails..SIT TIGHT if you are long.

[Mamma Boom Boom]

BREAKOUT?

[Mamma Boom Boom]

Hank, don't forget to add to shorts. You promised!

[ed]

Oh Oh mamma no triple top, say it ain't so

[ROLF]

Buy, buy, buy momma! Sure it's a breakout!Quad witching and all.I'm sure JT is buying hand over fist.LOL!

[Account Deleted]

S&P 500 Broadening pattern

[Michael]

This is yet another PRIME EXAMPLE of why these young "bloggers" like Kenny, Daneric, Binve, as well as others have no clue...

When they go making a HUGE deal out of AAII bullish sentiment increasing, it's quite telling.

The retail investor is literally non-existant in this market and has been for quite some time!

These BS "bloggers" fail to understand that retail only makes up 11% of the daily volume in the stock market. They also conveniently ignore the fact that there have been net OUTFLOWS of roughly $48 BILLION from US equities since April of 2009.

More failed analysis from the "kiddies" out there.

Next.

[Hank Wernicki]

Sure did !

[Mamma Boom Boom]

Hank, good for you. I like to see a man keep his word.

http://www.timecities.com/Members/mamma/EMclapping.gif

[Bull trade]

NYSE cumulative advance decline is at RECORD HIGHS, eventually bodes well for higher bull market prices down the line.

Ignore bears who say shorting a Bull market is a wise strategy.

[Mamma Boom Boom]

>EWP expects Friday to be wild:<

Yep, my poor wittle heart is just going 'pity-pat....pity-pat'....pity-pat'.

[Michael]

Congrats to all of those who bought shares of EXAS earlier this year ( around $2.75 ) as my "Stock of The Year".

Traded up to $6.37 today.
Check the chart.
:)

[Mamma Boom Boom]

Ok Michael, you forced my hand:
'Good Job'.

[Dsquare]

Money, what is it, who should control it?‏ This is an interesting talk given to the House of Lords (2004) I came across at Zero Hedge.
http://www.monetary.org/houseoflordstalk.pdf

Yelnick, gold bugs, any comments?

[yelnick]

DSquare, a very interesting speech, thanks for sharing. In Andy Jackson's time and up to the creation of the Fed (1913) the issue was who controlled the Money Power, the risk being said back then that if it in the hands of oligarchical banks it can be used to steal from hard working people. He doesn't use that way of framing the issue anymore, because modern economics has generally ignored the issue of Money. The speaker contrasts "private" with "public" money, conflating gold and credit, both of which he throws into the "private' realm. The Money Power fight was between credit run by private banks and gold as People's Money, outside the control of government or banks.

His solution is to make the Fed a branch of the US Govt and have it issue credit only, removing the banks from creating credit out of nothing by requiring 100% reserves. The Fed credit comes into circulation via USG spending, so this plan would empower a leviathan state (which we already have).

The concept of 100% reserves will starve the private economy of growth capital. The Austrians (Rothbard) argue for 100% reserves but only for checking accounts. He may mean the same thing but he did not clarify this in the speech.

The old gold standard got around the problem of starving a growing economy of capital by the use of Real Bills, which can be leveraged 20x or more yet always get paid in gold accounts.

It may be the next macro economics genius solves the money issue.

History tells me we need three forms of money:
- the Real Bills sort for financing commerce, which could be based on gold or centralized reserve accounts in the central bank
- consumer credit, especially the sort that is collateralized (mortgages, cars)
- growth capital of the sort the shadow banking system provides (venture capital, growth equity, IPOs)

[HighPlains]

I think both the pattern and the sentiment will lead to higher prices.
Neely closed out his long trade today, from 1032 to 1120. Beautiful trade, however I think he was anxious to book the big win as more upside is in store next week.
my guess.

[JT]

"Buy, buy, buy momma! Sure it's a breakout!Quad witching and all.I'm sure JT is buying hand over fist. LOL!"

Dude, you sound like the typical BITTER BEAR. Sounds like you've been short during yet another 70-80 point rally in the S&P.

What a Loser.
LOL!

[Nick Radge]

Negative outflows from equities, yet prices keep going higher. it's also an upside breakout from the triangle. Not textbook in any way, but does suggest a retest of the April highs potentially as part of a larger, more elongated flat pattern.

STU 'wild Friday' - wrong again
STU 'bearish gold' - wrong again

Give it up you idiots and start trading price action for what it is and not what you want it to be.

[Baby Bull]

Wonder what the STU will say tonight after the SPX traded an intra-day high of 1131.47

LOL!

[DW]

In 1994 Dow was roughly 3500.

Prechter said in an interview with Barrons that if the Dow ever broke above 4000, he and EWI didnt know what they were talking about and to never listen to him again.

1994 Barrons, look it up!

[Wave Rust]

DW,
hi Don
you been killin it nicely

Prechter is the perfect personification of why traders lose. When they (like BP) are wrong in their analysis and/or trade, they can't/won't/don't change. Foolish and sad.

It's an amazing feat of marketing that Prechter is still in business.

You are still the best trader I have ever seen, Don. Bar none!

Tip of the hat to you.

wave rust

[vipul garg]

michael,
good pick there with exas.

though you have mentioned exas a few times for benefit,
my trading style doesnot permit me to be a positional trader.
i like to take large positions for bursts.

in anycase, great pick.

[yelnick]

Baby, the STU simply said that the Dow closed its last gap, but has not exceeded the Aug high, while the S&P only broke it intraday, not closing. Most indices remain below. Hence, multiple non-confirmations. So they remain bearish. Prechter launched an EWT today as well, to keep the faithful in line, As usual, Prechter is a very good analyst and scores lots of points, particularly on the psychology of a wave 2, where near its end the optimism often exceeds that of the prior top (same with a B wave in a flat) since traders are convinced they caught the bottom. 

What is most telling in his psych approach is the big sideways move for the past four months, which is not a bullish pattern. You can argue it is a consolidation before a new rise, but it came about because the Stimulus was visibly fading and the economic news began to roll over. No one talks of a strong recovery - the V is long gone & the debate is all about avoiding a double dip and skimming along the bottom. Not the conditions for a new bull market.

Now, if they had dropped their nested 1-2s and followed my Fractal Finance sideways count (whether a Triangle or a double flat machts nicht), they wouldn't be worried at all about changing their count. We sit at a triple top, and need a real break above not a trivial false break to change the bearish outlook. And we remain far away from the April highs. 

Take a look at Tony Caldaro over the weekend; I think he will make some interesting points about the bullish case. This week proved nothing - instead was sideways after a brief spurt up to launch September. Next week may be when September revels itself. 

BTW while everyone talks stocks, binds continue to fall as is the USD. The action is over there. 

[yelnick]

Nick, the STU did not say Friday would be wild, but an ewave blog site.

[Chico]

Michael,

I know you've been touting EXAS all year. Great pick. How much upside do you think it has left?

[DG]

"i like to take large positions for bursts."

How are you determining your position size? If you can handle drawdowns psychologically and can conceptualize risk primarily as "risk of ruin", the Kelly Criterion is the way to go, so long as your method has a positive expectancy. I was just looking at some stats on a Kelly-driven portfolio using my trading logic and found that 35% of the time, it was drawn down 75% or more from its maximum, 41% of the time it was at less than 50% drawdown from its maximum and 16% of the time it was making new highs.

Psychologically, that's nearly impossible to handle. But, that account, if you had started it with $10K on April 3, 2009 and didn't have any limits on the amount of ES futures you could trade (actually this is the most practical limitation on the strategy), would have grown to $381 million as of Friday, but, along the way, you would have experienced a drawdown from $399 million to $30 million, which is still an enormous amount of money. Anyway, just to give you a sense of why I like Kelly. My idea is to automate a trading platform to trade my logic on the Euro, seed the account and then not look at it for a year.

I'm also working on a new long-term trading logic based on Weekly polywaves designed to take advantage of longer-term trends. It's more or less the inverse of the logic I developed earlier this year, simply as a non-correlated complement to it and for the purposes of allocating to a non-trading portfolio. The logic would have gotten you in to the SPY at 76.40 after the March 2009 low and your exit would have been 106.59 in early February 2010. One trade for almost a 40% gain. Looking at the decline to the March 2009 low from the September 2008 high, I see that getting in to that trade would have required probably 4 attempts, so risking 1% on each of those prior to the trade actually taking hold means your total return would have been around 34-35%, with only a 20% "winning percentage". Now that is "risk-reward" being in your favor, although my gut tells me that 2009 will have been an extreme case.

This logic would have gotten you short in April just prior to the "flash crash". There would have been a couple of long and short trades in the meantime and right now you'd be long from 1079.18 SPX with a stop at 1121.34 and you'd go short (assuming the 1131.47 high holds) at 1097.41 with a stop at 1131.47.

Anyway, it looks promising as well.

[vipul garg]

DG,
( kindly see your blog for a 'detailed' reply)

i primarily trade indian stock market.there,it is actually difficult to quantify how i do my position sizing because it is not rule based but based on my analysis of the technical position of the market and the stock and how much volume the stock future etc is generally there.
i am unable to trade on 1-2% risk model.or maybe i havent tried enough to do quantified risk based trades where you know how much you will loose if the position goes against you ,and is almost fixed for every trade.

the reason is largely why i am trading. a return of 30% is not acceptable to me and is not my aim.so i am looking for the casino effect .actually its more like skilled poker.you take a position when market, in your analysis, has fewer hands to play based on simple technical analysis.
and this is quite contrary to what every trade management book will say,i do understand that ,but so far it is a conscious decision on my part.

god has been kind and so far it has worked.
my trading style carries risk of ruin , since it is margin trading, and a large gap against my position can wipe out the trading account. it is for that reason i
have kept my trading capital almost constant.

my weakest point is the timing of my exit of a trade since large position and absence of rule based stop out makes handling volatility a psychological challenge.
recently i have started working on developing a positional trading style with quantified risk. however a large amount of hard work is needed to incorporate that which i havenot been able to do.

[Michael]

"Michael,

I know you've been touting EXAS all year. Great pick. How much upside do you think it has left?" - Chico

EXACT SCIENCES is looking at the biggest opportunity in medical diagnostics in quite some time... We are talking about a $1.2 BILLION dollar market place and that is only assuming 30% penetration.

The stock currently trades at a mere $250 million market cap. I was highlighting the "name" back around Christmas of 2009 when the market-cap was a little less than HALF that.

While space is limited here to address all of the powerful fundamentals of this opportunity, I will try to touch on a few significant points...

Suffice to say, that they will have NO competition in the non-invasive colorectal cancer screening market given their patents on DNA extraction from stool AND the combination of methylated markers that they are using for detection of not only colon cancer, but PRE-CANCEROUS flat-lesions that an invasive procedure like colonoscopy has literally no impact in detecting when it comes to the right/descending portion of the colon.

My suggestion to you is to listen to the audio/slide presentations that the CEO gave this past week in NYC. You can find them archived on the Company's website under the Investor Relations section.

Remember, colorectal cancer is the #2 killer of Americans over the age of 50. It is also showing up in huge numbers in the 40-50 crowd, just ask CBS's Katie Couric.

60% of all who are diagnosed with CRC are diagnosed in Stage 3 and Stage 4.

It is a very slow moving cancer that can be stopped if caught early. Trouble is, no one wants to undergo the bowel prep the night before for a colonoscopy, nor do they want to undergo such an invasive procedure, go off meds, take an entire day off work, risk colon perforation, a sloppy procedure, etc.

12 million Americans annually undergo a cheap $10 Fecal Occult Blood/Immunochemical Test (FOBT/FIT) that requires you to spread "poop" over a piece of cardboard for 3 days during a week. The only thing that this test does is detect BLOOD in the stool. It really doesn't have much to do with cancer detection, only blood detection, which can come from all sorts of things besides colon cancer. (ie. hemmoroids ). As a result, there are lots of false positives and people are sent off for needless colonoscopies.

The stoolDNA test by EXAS has been guided by management to be 85% sensitive for CRC, and at least 50% sensitive for pre-cancers, which is HUGE.

Remember, we have virtually erradicated cervical cancer with the pap-smear test and when it first came out it only had 50% sensitivity, but because women take this test on an annual basis, cervical cancer deaths have dropped from 35,000 per year, to roughly 3,300 per year.

The EXAS stool DNA test will most likely cost $300 and command 50-65% gross margins. I believe that it will also be recommended to be used once every 3-4 years. Currently, the American Cancer Society recommends one obtain a colonoscopy once every 10 years after one turns 50.

If the test simply replaces a small percentage of the FOBT/FIT tests that are taken on an annual basis ( 12mm ) I have price projections (even assuming between 52-55 million shares vs 40 million currently ) that are at least $50 per share.

Remember, this is not a drug. It is a LAB test that will be using commercially available PCR assay procedures. There might be some software "tweaks" made at the labs, but that is it. Also, the lab will be seeking insurance reimbursement, not EXAS.

So in essence, EXAS becomes a licensing Company for a lab test that costs the United States 50,000 deaths per year at a cost of $12 BILLION per year. And Medi-Care winds up paying for 70% of that number, annually!

As with most preventative medical procedures under ObamaCare, there will most likely be no co-pay by the insured, just as there will be no co-pay for Mammography or Colonoscopy (starting this month). In fact, currently there are 23 states that mandate insurance reimbursement when it comes to CRC screening.

I don't believe that the big funds ( like Fidelity ) have even begun to get involved in this stock. I believe that the big funds are awaiting the presentation of the stool DNA "Validation Study" results which will occur in Philadelphia on Oct. 28th at a major medical conference, which will then be used as leverage to become published in a major medical journal such as the NEJM.

The current management team came onboard 18 months ago. They have a proven track record given their previous Company - Third Wave Technologies that developed an HPV cancer screening test that eventually received FDA approval just 10 months before Third Wave was acquired by medical diagnostic company, Hologic for $568 million.

A parallel Company for one to do some due-diligence on would be CYTC, which developed the "ThinPrep" pap smear test and was bought out by Hologic for a most generous 7.2 times price/sales and $6.2 BILLION dollars!

Interestingly enough, a lot of big time mutual funds ( like Wellington, TCW, and Fidelity ) have piled into biotech names without doing effective due-dilligence.

Recently, these big funds took a huge hit this past week in the likes of Arena (ARNA) with their diet pill.

And Fidelity once had a 10% position in Sequenom last year when the stock was trading at a $1 BILLION dollar market-cap even though the Company depended on an "internally" generated Validation Study that was frought with all sorts of "irregularities". Even with this study having to be tossed out the window, SQNM still trades at a HALF-BILLION market cap!

If EXAS were to trade at a half-billion market cap, the stock would be at $11-$12 per share.

Meanwhile, Dendreon (DNDN) currently sports a $6 BILLION dollar market cap all because they have a prostate cancer drug (Provenge) that has recently completed Phase III trials and extends a patients life by only a few months at a cost of nearly $90,000.

My point is that if you are unable to comprehend what the Bio-Tech space is like, and what kind of market-caps are common in this space, you will miss out on understanding that EXAS is not just a "trade" as some have admitted to falling prey to, but an INVESTMENT that has a long way to go given it's rather paltry market-cap of $250 million.

Again, the Validation Results are due out on October 28th. These results will be presented by none other than Dr. David Ahlquist of the esteemed Mayo Clinic.

These results will not have been generated "internally". They will come from an independent 3rd party, the Mayo Clinic.

Once these results are presented, I am highly confident that the stock will attract the attention of large mutual funds like Fidelity.

Right now, the Big Boys have yet to even take a position in this stock.

As in all cases, please do your own homework and due-diligence as there are always risks in the investment field.

That having been said, those who try and merely TRADE this Company's stock price in an "in and out" fashion will miss out on the Big Picture. You will be left behind, and not be able to get back in . . . on the way to potentially much higher prices.

Exact Sciences has an awful lot going for it. And the new management team is quite impressive.

Good Luck to All!
:)

Disclaimer: I am long shares of EXAS and all of the above information is publicly available.

[Michael]

PS. Colorectal cancer kills 50,000 Americans annually. But on a global basis it kills 650,000 per year. The discussion and data disclosed above only pertains to the US domestic market. As a result, it's not too difficult to see why this could be a terribly exciting break-thru in the medical diagnostic marketplace!

[Chico]

Michael, thanks for that very detailed response. I appreciate it!

[Chico]

DG,

I don't know if you've read any of Ralph Vince's work, but the Kelly Criterion only applies to situations where all wins are for the same amount and all losses are for the same amount. Of course, this is usually not the case when trading, and why optimal f should be used instead.

[Nick Radge]

Yeah - Ralph Vince , good in theory. No good in reality.

[JohnEBGoode]

Contracting triangles and no Fed sponsored QE2 could spell disaster for bulls.

http://content.screencast.com/users/texana44/folders/Jing/media/552afd5e-3b2e-40c9-872c-8215e92a7d3e/2010-09-19_0903.png

http://content.screencast.com/users/texana44/folders/Jing/media/bd00c7cd-69a9-4b41-9770-8d60b3aeb99c/2010-09-19_0826.png

[Eventhorizon]

Chico,

A small correction: the Kelly Formula applies as you have specified, the Kelly Criterion simply means maximizing the geometric return over a given holding period (at least, according to Vince).

I just finished Vince's "The Handbook of Portfolio Mathematics" - it would be a tour de force if not for the typo's! - I am a few pages into "Leverage Space Trading Model". What do you think of his work?

Nick - Vince makes the point that it doesn't really matter what you think of the reality of optimal f, there's no getting around the fact that any trader is operating in the "leverage space" whether he likes it or not. It's one thing to understand optimal f and consciously choose sub-optimal leverage for reasons not captured in objective of maximizing geometric growth of capital.

Vince is explicit on using real risk (draw-down rather than variance of return) as the constraint on maximizing growth. So his objective becomes: "Maximize geometric growth subject to limiting the probability, y%, of a draw-down greater than x%". I don't see how this is "No good in reality", it strikes me as exactly what traders try to do in reality. What do you see as the flaws in Vnce's work?

[DG]

Chico,

Yes, there is that subtlety involved in the Kelly Criterion, but if you recalculate Kelly after every trade you can capture the new information regarding wins and losses, percentage and ratio-wise, for your next trade and position-size accordingly. Kelly is "myopic" in that way and doesn't take into account serial correlations, which is both good and bad. And, again, one simply HAS TO conceptualize risk NOT in terms of drawdown or risk on a single trade but in terms of risk of ruin.

One of the best discussions of Kelly I've found is here:

http://epchan.blogspot.com/2010/04/how-do-you-limit-drawdown-using-kelly.html

There is also a "continuous Kelly" which differs from the "discrete Kelly", but that's more about how much leverage to use rather than how much of your portfolio to risk on a single bet, as you probably know.

Then, there are all sorts of "fractional Kelly" applications based on the assumption that whatever estimates you have of winning percentage are too high. One thing you could do is use a "rolling Kelly" based on the last X number of trades and if that is smaller than your overall Kelly, bet the smaller percentage. Or, if it's larger, bet the smaller. The only thing to realize is that those are sub-optimal in maximizing portfolio growth. But, as many have argued, that is not the utility function most people are trying to satisfy in their portfolio construction. Some people like to sleep at night!!

But, as my stats above show, even using Kelly it is only a matter of time until you reach a new high-water mark, no matter how deep your drawdown, assuming your system retains positive expectancy. Also, since my trading method trades just about every day, the chance of my not reaching a new high water mark every few months is low. So far, the 8 months between one high-water mark and another have been the longest. Prior to that, it was 2 months. I've modeled out the deterioration of my "edge" over time and can forecast the day when, at current long-term trends, the edge would disappear. But, within the long-term trends, there are sub-trends whose rate of deterioration is slower and would allow the edge to continue for longer. But, August was my best month yet, which tells me that the edge can also improve from a trough. This tells me that the edge for this method varies, which I am pretty sure is tied to the nature of the wave structure unfolding in the market. But, I always had hypothesized that anyway, which is why I am now developing a method which is the inverse of the one I developed earlier in the year.

In my backtesting example, the Kelly-based portfolio hit a high-water mark in January of this year, drew down 90%+ and hit a new high-water mark a couple of weeks ago. So, the question then becomes, do you have the psychological fortitude to wait 8 months to hit a new high-water mark after enduring a 90% drawdown? Can you sleep at night if your portfolio is down more than 50% from its high-water mark almost 60% of the time?

An automated bot trading your strategy does, but most people don't.

[Eventhorizon]

Sorry, forgot to complete this thought ....

It's one thing to understand optimal f and consciously choose sub-optimal leverage for reasons not captured in objective of maximizing geometric growth of capital...... it's another to pretend that the math doesn't apply to you because you don't like the implications, or worse, dismiss it because it's too hard (compared to MPT) or challenges long-held institiutional dogma (MPT!!).

Progress is made by those willing to challenge the status quo.

[DG]

"Vince is explicit on using real risk (draw-down rather than variance of return) as the constraint on maximizing growth. So his objective becomes: "Maximize geometric growth subject to limiting the probability, y%, of a draw-down greater than x%". I don't see how this is "No good in reality", it strikes me as exactly what traders try to do in reality. What do you see as the flaws in Vnce's work?"

If you conceptualize risk as "risk of ruin", rather than "risk of drawdown of size X%", where "X%" is less than 100% AND if optimal f gives you a smaller betting size than Kelly, I would argue that's a flaw, assuming your trying to maximize long-term growth.

For example, if my largest historical drawdown is 90% using Kelly and then optimal f tells me that my actual bet size should be, e.g. 1/3 of the amount Kelly tells me, it seem intuitive to me that you would never actually maximize growth in that scenario AND you could still end up with an, e.g. 89% drawdown. So, yes, optimal f has decreased my maximum drawdown, but my high-water mark is far lower than it would have been using the full Kelly betting size.

Obviously, I'm not saying that optimal f would suggest a 1/3 Kelly-sized bet to achieve a 1% reduction in maximum drawdown, so that's just a scenario to illustrate the potential effect of focusing on drawdown as opposed to growth maximization.

If one were to utilize something other than Kelly, my personal preference is one that has historically been used by "Kelly traders" once their portfolio has reached a certain size, which is that they shift from geometric growth to linear growth as the target quantity.

I like it because it also seems sensitive to the "time value of money" criterion, i.e. I want to get richer, faster, by using Kelly and then I can shift to a position-sizing strategy which maximizes linear growth. With optimal f, I may eventually get as rich as I would with Kelly, but over a longer period of time. As an investor with a finite time horizon, that matters, even if I do eventually shift to linear growth because I don't want to lose whatever my Kelly bet size is on my last bet of my life.

[DG]

"As an investor with a finite time horizon, that matters, even if I do eventually shift to linear growth because I don't want to lose whatever my Kelly bet size is on my last bet of my life."

And, presumably, if your method's edge is enduring, you can set up an institution which lives beyond you to utilize it, which would argue even more for doing Kelly your entire life.

[DG]

"Obviously, I'm not saying that optimal f would suggest a 1/3 Kelly-sized bet to achieve a 1% reduction in maximum drawdown, so that's just a scenario to illustrate the potential effect of focusing on drawdown as opposed to growth maximization."

And, to just add one more thing to this though, I would argue that "drawdown" is a psychological inhibitor, not a financial inhibitor, at least not until it hits 100% and you are bust. Obviously, that's assuming you can still place a trade or can add enough incremental capital to place a trade while your method still has positive expectancy but has hit enough of a string of losses such that you can't place the next trade.

What if Livermore had decided to quit trading the first time he got down to his last dollar?

[vipul garg]

a large rally is entirely possible on spx.

[JT]

So much for a WEAK September with the SPX already +7% for the month which is pretty close to the best September since 1950 (+8% in 1954).

I wouldn't be surprised by a "pullback" this next week (especially given how poorly crude has traded and a typical post-expiry drop), but once again all of the Perma-Bears have been proven wrong and their short positions blown up once again.

[Eventhorizon]

Hey DG,

Excellent thought process as always. My response was directed to the usual criticism of optimal f, which I assumed was implied by Nick Radge's comment: optimal f works mathematically but no "real" manager could actually use it because of the draw-downs implicit in the method.

Your argument in no way conflicts with optimal f. The Kelly formula is a special case of optimal f. The risk of ruin concept is a special case of optimal f (max permitted draw-down = 100%).

What I like about Vince's approach is he leaves it up to you: if you are willing to maximize growth in an unconstrained manner, you can figure your bet sizes and live with the expected draw-downs (you will remember, I am sure, that the better the system, the greater the anticipated draw-downs at optimal f). If, on the other hand, you are a fund manager whose clients will simply not tolerate draw-downs beyond 15% then you can figure optimal bet sizing subject to that constraint. Point being, Vince's approach includes Kelly within its solution space - so what's not to like?

[Chico]

Eventhorizon,

Yes, the Kelly formulas yield the optimal fixed fraction that maximizes the growth of one's stake for certain betting situations, as stated by the Kelly criterion.

I am a fan of Vince. I first ready Portfolio Management Formulas about 20 years ago. I recently just finished The Handbook of Portfolio Mathematics and also just started reading The Leverage Space Trading Model. I'm not sure I'm going to implement his portfolio optimization techniques, however. I think the results are too dependent on the relative performance of each of the portfolio's components continuing to hold up in the future. I think it's more prudent to structure a portfolio under the assumption that you do not know which component will perform best in the future.

I think the main reason people criticize optimal f is because f was derived from an unrealistic series of outcomes. This could be caused by a number of things, such as using too little history that isn't representative of the market over time or an over-optimized system that has no hope of holding up in the future. Either way, you end up to the right of the peak of the optimal f curve. As Vince says, you need a system you can take into the foxhole with you.

[Chico]

DG,

I'm curious if you've calculated the return using optimal f. Vince is adamant that there is no other wagering scheme that will result in faster geometric growth (with the condition that the outcomes are independent). Also, the f that results in the highest stake in long run will also have the greatest drawdown. Someone that can't live with that can always dilute the optimal f (e.g., trade at half-f), with the understanding that while you reduce drawdown arithmetically you also reduce growth geometrically. I don't know how much you know about optimal f, but it is derived at through iteration rather than plugging numbers into an equation.

[DG]

Eventhorizon,

Yes, it's good to discuss something a bit more down-to-earth than wave counts and of course you're right that the Kelly is simply the "100% drawdown" case of optimal f.

Incidentally, it was really after collecting all of the Neely S&P trading data from August 2006, which I did at the end of last year, which got me motivated to investigate different trade sizing methods. Neely's Weekly trades have a Kelly up near 30% and expectancy near 45%, while his Hourly have a Kelly around 5% and an expectancy around 8%. The Daily numbers are somewhere in between that.