Forum Archive Index - February 2003

[Travis Morien <travismorien@yahoo.com>]

--- "tennyson@caverock.net.nz"
<tennyson@caverock.net.nz> wrote:
> >>Another claim :- "Trend following systems have the
> >>fatal flaw that they result in very high
> >>turnover..." Nonsense! Take another look at the
> TLS
> >>chart that began this discussion - ONE trade in
> over
> >>5 years! I have posted many charts of stocks in
> >>long-term uptrends, with the comment "it is all
> but
> >>impossible to beat buying and holding stocks that
> >>are in steady long-term uptrends like this".
> Anyone
> >>wishing to avoid the costs associated with more
> >>active trading can easily devise a system based on
> >>long-term trends that will signal very, very few
> >>transactions.
> > 
> > How many times is one faked in and out of these? 
> >
> >
> 
> None

This is the same logic as claims of the existance of
support and resistance which despite the claims made
by traders worldwide has never been found by an
unbiased computer... hmmmmmm.

By that i mean you've provided an example where it
worked and assumed that a single example proves the
rule.  Like it or not once again low turnover trend
trading systems can't be programmed into a computer
because as anyone who's ever backtested an indicator
in Metastock can attest the hit rates on any technical
indicator are too low for them to pay off.

No evidence has ever been shown that technical
analysis works.  For example the "Turtles" a famous
group revered in the mythology of futures traders once
demonstrated that their methods (risk management (loss
minimisation) and money management (trade parcel
sizing) are sufficient to be profitable *even with a
randomly chosen entry point*.  

Successful traders are much like successful gamblers,
they occupy themselves with managing their bet sizes
and limiting the amount they lose in any one session. 
No professional gambler fools himself into thinking
that he can predict the cards because he knows that
any system of prediction is destined to fail because
card sequences are not determinate.

I'm not claiming that there are no successful traders,
I am claiming that there are no successful technical
analysis methods and I claim as my evidence decades of
academic research which has systematically backtested
every popular (and unpopular) TA system and found that
none of them work.

I also claim as evidence the fact that players like
Paul Tudor Jones II, The Prediction Company, Merryl
Lynch, Goldman Sachs, Morgan Stanley, Fidelity
Investments, Renaissance Technologies spend tens (or
hundreds) of millions of dollars to develop their
technical trading systems, with mixed success.  Yet,
as is commonly seen on the net, thousands of amateurs
claim to be highly successful traders and they employ
only the most basic of TA techniques.

Either Jones et al really didn't do their homework by
overlooking these techniques or they realised
something was amiss with the simple happy world of
these little amateur speculators.

 
> If you look at Phaedrus's post on 31st January you
> can see his sell 
> point clearly indicated on the graph he posted.  His
> sell point looks 
> to be about $7.75, well below the resistance point
> of just over $9.   
> At this point the share goes into a trading range
> where the trend is 
> not present.   The whip sawing you talk about
> Travis, giving false 
> signals is occurring at this point but Phaedrus is
> not paying 
> attention to these false signals.  Once Phaedrus
> quits a trade he 
> will not trade again in that share unless a new 
> uptrend direction 
> is firmly established.  This usually means the share
> must break 
> previous resistance levels and it is clear it does
> not.   Once the 
> next trend is established it is clearly a downtrend.
>  Phaedrus does 
> not invest in shares that are in medium and long
> term downtrends.

Ok, I take it that telstra is the only stock Phaedrus
has ever traded since we're all focusing so closely on
this and ignoring the possibility that maybe he just
got lucky.

Unless someone would like to point out that this is a
general characteristic of Phaedrus' market calls, that
they are inevitably nice clean one way trips and that
volatility is never confused with trend changing??


> 
> >
> > 
> >The trends are unambiguous, apparantly work more
> than
> >50% of the time and do not generate high turnover
> >which means they generate a singal buy/sell signal
> and
> >not a series of whipsaws.  
> >
> >
> 
> It is not necessary to get a 50% strike rate to have
> a profitable 
> trading system.  The key is to make sure the value
> of your winning 
> trades more than cancel out the value of your losing
> trades which is 
> not the same thing.   For example, suppose you can
> show that by 
> jumping on share that is in a confirmed uptrend
> there is a one in ten 
> chance that a share will increase in price by 50%. 
> You identify ten 
> shares that are in an uptrend and buy them all.  One
> of these shares 
> does appreciate by 60%.  The other nine start to
> rise then fall away, 
> fall immediaetely or stay flat.   The average return
> from these nine 
> shares, after brokerage is -5%.  You don't lose more
> than this 
> because you have a tightly placed 'stop loss' on all
> of these shares.

This is money management and risk management, a field
quite distinct from technical analysis as it does not
involve chart reading, prediction of trends or
reaction to trends and instead focuses on the trader's
equity.

Just as with gamblers there is no suggestion that you
know what is coming up next, and this certainly has
nothing to do with support and resistance.

> Your single remaining share actually appreciates 60%
> but then falls 
> away and you exit your share trade making a 50%
> gain.  This gain 
> neatly cancels out five percent losses on all the
> nine other shares 
> leaving you with a small profit.  The purpose of
> this example is to 
> show that even with 90% of your trades being losing
> ones, you can 
> still come out ahead. 

Sell stops, buy stops, trailing or fixed are simple
forms of trade risk management.  The Turtles
demonstrated that it is mastery of this, and not of
technical analysis, that makes a person a successful
trader.
> 
> > 
> >It won't tell you when the trend ends, it will give
> a
> >late signal.  As useful as "that was the turnoff"
> from
> >one's wife while you've just sailed past it in the
> >right lane at 100kmh.
> >
> 
> Phaedrus has never claimed his system gives anything
> other than a 
> late signal.  A reactive system, such as he uses,
> will always give a 
> late signal.  The fact that by using a late signal
> means that he will 
> never buy at the absolute bottom of the market nor
> never sell at the 
> absolute top is a fact he has acknowledged here many
> times.  By the 
> same token a Buffett value investor cannot
> guaranteee that he will 
> sell at the top either.  But that fact doesn't
> invalidate the 
> usefulness of being a value investor.

Why not just cut out the whole technical analysis
thing then, given that it is generally agreed to have
little or no utility for forecasting and implement a
risk and trade management system like the ones the
Turtles use?

This thread started out with my pointing out that for
every example of support and resistance that works
there will be at least one that doesn't, implying that
buying or selling based on the concept is dodgy at
best.

It has migrated away from that into a discussion
singing the praises of loss minimisation via simple
stop loss techniques.  I've never claimed that those
don't work, only that TA is a waste of time.

Risk management isn't TA so anyone trying to argue
that TA works because risk management works is tying
themselves into logical knots.
> > 
> > 
> > You wouldn't maybe turn this around and determine
> a
> > fair price before the trend turns around, so
> you'll be
> > prepared?
> >
> >
> 
> Fundamental analysis, if it is useful, requires much
> more legwork 
> than T/A.  Ideally if you were using a mixed T/A and

Which is why TA is so popular.  Why bother doing
something properly when you could do it easily?

> fundamentals 
> approach  you would do your fundamentals research
> first.  However, 
> that would necessitate looking at a much narrower
> range of shares and 
> you might miss out on a few bargains you hadn't
> thought of, which a 
> wider T/A net might pick up.

Ok, why not widen our net as broadly as possible and
all buy market scanning tech software that will find
opportunities in the entire market?

I wish someone arguing the TA case would explain to me
why their TA techniques never work when examined by
impartial researchers or computers, yet the strategies
are supposedly successful enough that you can all make
money from them in real time??

 
> > 
> >
> >I can't name a successful investor with a public
> >record that has used TA with any success.  
> >
> >
> 
> Given that T/A people spend many hours in front of a
> computer each 
> day, and are happy to do so, they may not have the
> personality type 
> to loudly declare their success to the outside
> world.

Just one... the audited public records of dozens of
incredibly successful fundamental investors who've
beaten the market for many years are available and
there are now a number of academic studies that have
shown beyond reasonable doubt that many traditional
fundamental methods such as purchase of stocks with
low PEs, good cashflow, steady earnings etc do work.

On the negative side we've got a large body of
literature that has systematically demonstrated random
walk character in stock prices and the rapid breakdown
of trends followed by regression to the mean, we've
got behavioural finance researchers like Terry Odean
investigating the accounts of tens of thousands of
discount brokerage clients and thus demonstrating that
amateur investors underperform indexes by a large
margin and that turnover is directly correlated with
the degree of underperformance and that the only group
that outperformed the market was the group with the
lowest turnover (buy and hold for many years).  See
http://faculty.haas.berkeley.edu/odean/ 

We've got studies by Heironymous showing that
speculative futures traders lose their money very
quickly and that only a minute proportion ever do more
than break even over many years.  

There's the Johnson Report on day traders,
commissioned by North American Securities
Administrators Association (NASAA, a consumer
protection group trying to keep the futures industry
honest) that demonstrated that almost none of those
surveyed made money and the few that did so appeared
to take on so much risk that their risk of ruin was
calculated to be nearly 100%, implying that eventually
they were nearly guaranteed to lose.

Dalbar Inc's ongoing study "Quantitative Analysis of
Investor Behaviour" has tracked the performance of US
mutual fund investors since 1984 and found that
generally people use a trend following system, they
buy when results have been good, they sell when things
turn bad.  The result of this, according to the 2001
update was that from 1984 to 2000 the average US
mutual fund investor underperformed the index by over
10%pa.  

So widespread and conclusive data collected looking at
the results of many tens of thousands of investors has
shown that they do very poorly, there are no known TA
exponents that come anywhere near the calibre of a
Buffett or a Soros or even any known TA users running
a particularly successful managed fund.  But...
somewhere out there are these successful traders, they
jusy keep very quiet about themselves.

I've seen this sort of claim before, I used to debate
with astrologers back at the time when I was more
active as an astronomer.  They all agreed that the
majority of people passing themselves off as
astrologers were full of crap and that the horoscopes
from the paper were written in vague terms that could
be interpreted any way you like to interpret them...
but... they all claimed that they happened to know
this bloke that was a "real" astrologer, except these
guys naturally keep quiet about it.  Many others
claimed that while astrology didn't work for most
people, it worked for *them*, and no, they weren't
willing to allow anyone to verify this somehow. 

 
> >after the fact you mean, long after the purchase
> will
> >you have any idea if there was a valid buy signal. 
> >Or, if you wait until the "confirmation" what do
> you
> >then do if the stock falls back?
> >
> >
> 
> You sell, taking a loss.   Not a problem as long as
> your total return 
> from winning trades exceeds that of the losing
> trades

Sounds like risk management to me.  I'd put it to you
that the results would be pretty much identical if you
flipped a coin rather than drew support and resistance
lines on charts or plotted breakouts or topping
patterns etc.  As long as you have a trailing stop of
some sort you'd get the same results.
 
 

> >I know the answers to these, but i have great
> trouble
> >reconciling the whipsawing reality of technical
> >analysis in real time with your claim that turnover
> is
> >really low.


>>From what I have seen Phaedrus's turnover is low.  He
frequently holds shares for several months which is
probably about the same turnover rate as some of those
so called fundamentalist managed funds who
theoretically buy and sell only for the long term, yet
still manage to churn their portfoloio over several
times a year.

The mind boggles that someone would consider a person
that "frequently holds shares for several months" has
a low turnover.  Low turnover is when you hold each
share on average for several years.  Sounds like the
turnover you are claiming for Phaedrus is of the order
of several hundred percent pa, which is monstrously
high by any standard.

The fact that a number of very poorly managed managed
funds have a high turnover does not make this ok.

A survey by Morningstar USA found that funds with a
turnover above 100%pa underperformed funds with a
turnover below 20%pa by 1.58%pa.  This is what you'd
expect given estimates such as those by the Plexus
group that each side of a trade costs a managed fund
about 0.8%, which means 100% turnover costs 1.6%pa in
trading costs alone.

Many other surveys of funds timing prowess have
confirmed that mutual funds are rotten market timers
and that the higher a fund's turnover the worse the
performance. 

> >Value investors missed
> >out on all the profits from buying Dot Coms in the
> >late 90s.  They still have their capital though,
> >unlike many.
>

> This is why I don't trade.  It is all very well
having > a stop loss et 5% below today's price.  But
if the 
> share falls 40% in one day you are screwed.  The 
> problem is there is no guarantee you will be able 
> to implement your carefully crafted risk management
 > policy in times of sudden change.

this has nothing to do with trading.  The argument has
moved again.

First we were discussing the validity of TA concepts
such as support and resistance.  I argued that they
aren't real or useful because they have been
demonstrated to work no better than coin flipping. 
You countered with an argument that risk management
can save your skin, which is all rather surrealist
because TA and risk management are quite distinct
concepts.

In the previous paragraph I pointed out that buying
grossly overpriced stocks just because they are going
up is a dumb idea and you argued that this is why you
don't trade.  Different concepts again.  

"Greater fool" momentum based speculation, buying
something just because it is going up despite the
fairly obvious fact that the thing is insanely
overpriced and guaranteed to fall eventually isn't the
same as trading.  Trading can be done with any stock
or commodity including reasonably priced ones.  The
collapse of tech stocks doesn't disprove the trading
concept, it disproves the greater fool concept.

> >To believe that you are going to achieve superior
> >returns to the bulk of market participants by using
> >the same techniques is a bit crazy, to say the
least.

> Yes but all T/A is not the same, just like all F/A
is not the same.  
> If for example you picked shares only on P/E ratio
being high, thus 
> showing the company to be a 'good growth company'
you may come 
> unstuck.  You can trade on different time scales
with different risk 
> systems and in effect customize a T/A system unique
to yourself. 

That would be a very dumb system given the bounty of
evidence that one is better off avoiding high PE
stocks.

The process of fundamental analysis goes along these
lines:

* Identify what the company does and how much money it
makes.
* Identify via simple valuation techniques (including
PE ratios) how well respected the company is in the
market, ranging from glamour stock through to dog.
* Find out as much as you can about the stock to see
what news has contributed to this price, whether that
be high or low.
* Examine this news in detail to see if the market is
correct.
* Perform a proper valuation based on your assessment
of the news surrounding the company and the
fundamental economics of the business to determine a
fair value.

Fundamental analysis goes very far beyond simple
indicators.  If you wish to invest with passive
methods (focus on a single quant indicator such as PE
ratio or company size) to take advantage of general
market tendencies (such as the "value premium" or
company size effects) then you'll need to diversify
extensively.  Or you could just do the sensible thing
and buy an index fund with those characteristics.  A
good example of such a manager is Dimensional Fund
Advisors.  They have passive funds that buy the
cheapest 30% of stocks ranked by price to book ratio,
and the indexes followed by the Australian and MSCI
value funds have outperformed the ASX200 and the MSCI
respectively by well over 5%pa in the period 1980 to
present.

If you are buying a stock based on PE without either
diversifying or carrying out more thorough research
then you are just not doing your job correctly as an
analyst and deserve the inevitably bad results that
follow.

> >I'm quoting me... my calculation shows that after
tax
>and making reasonable (though perhaps too low)
>estimates of trading expenses including brokerage and
>bid/ask spreads that you'd need to achieve about 17%
>as an intrayear trader, before tax and costs, to get
>the same return as a 10% buy and hold investor with a
>5 year holding period.
>

> Tax systems in NZ are different to Australia and the
USA.  For a 
> start there is no capital gains tax.   In theory if
you are a short 
> term trader you can trade away without the inland
revenue department 
> getting a whiff of your profits.  You are meant to
declare such 
> trades as income.  But how many traders in New
Zealand do?  I wonder?
> If traders keep all of their capital gained in their
trades with no 
> tax, how does that affect your return analysis
Travis?

Ah, ok.  Doesn't affect the analysis done by Odean
showing that *before tax* the higher one's turnover
the worse one does nor the Dalbar survey of mutual
fund investors which also demonstrates the costs of
trading.

See 
http://faculty.haas.berkeley.edu/odean/

http://www.dalbarinc.com/content/showpage.asp?page=2001062100&r=/pressroom/default.asp&s=Return+To+Press+Releases

So NZ traders will underperform buy and hold indexes
by only a few percent pa instead of three times that
much after tax.

> > I call on evidence such as Terry Odean's studies
of
> > small investors at discount brokerages that show
that
> > the higher the turnover the worse you do, 
>
>

> Sure, but I don't think Phaedrus does do a high
turnover.  Perhaps 
> this is one of the reasons he has been trading for
several years and 
> its still here.

Please reconcile "holds many stocks for a few months"
with this last paragraph.  I'm totally mystified by
that.


Travis
www.travismorien.com

__________________________________________________
Do you Yahoo!?
Yahoo! Shopping - Send Flowers for Valentine's Day
http://shopping.yahoo.com

----------------------------------------------------------------------------
To remove yourself from this list, please use the form at
http://www.sharechat.co.nz/chat/forum/