Hey, Techies......Reality Check

[ducati998]

rogue

I let the phrase True Probability ride as I wanted your definition. Probability is a mathematical calculation, it is neither true nor false. It is probability, plain and simple.

That is true enough.
However, when you actually start to collate the results, you then start to see the actualised results..................the true probability and this is currently what is starting to be shown.

And that is, technical systems return a win ratio of approximately 50%.
Your own results are testimony to the fact.

However, I guarantee if you had asked someone before the fact, what is the % probability of your set-up, the answer would have come back at a figure other than 50%

You only need read through these boards and you find all sorts of rubbish being promulagated.

This statement is contradictory, particularly in light of your definition of probability. As you have pointed out, "his return on capital employed is still very low." and since he has both costs of execution and the occurrence of loss he must by definition return a much higher % of wins than losses. Therefore his actual win / loss ratio must be high and hence by your definition the production of "actual high win rate" denotes high probability.

Not really, as I maintain true scalping is only really performed by professionals who have through whatever arrangement, very, very low, or non-existant execution costs.
Some of the Prop.Shops provide near as damnit zero execution costs.
In that scenario, scalping is viable.

Therefore, the win loss ratio can remain at 50%, and still be profitable, assuming the average winner is 1 tick greater than the average loss.

Incorrect, if a trader has 5 trades in a day, and they lose, then he loses money, adding another 200 losing trades (driving it by volume) will not improve his results. His profitability is driven by a high win rate.

Not in the above scenario. It is then driven by the volume of trades for a $$ return.
The % return on capital will (unless using 100% of account) remain unchanged.

But returning to the estimation, or calculation of probabilities.
The accurate way to approach the problem is via the exit .

Many traders, myself included when I traded technically argued that the entry was all important And while it most certainly is, defining the exit is as important, as without an accurately defined exit, the probabilities are really back to square one.......50%

As an example.
If you knew the future, and you knew stock XYZ was going to trade at $100 in 6mths time.
Then logically, the lower the price today, the higher your return as measured by return on capital divided by time.

If you also knew that it would trade, in a tight range for 4mths, and explode upwards for 2mths, then unless it paid a dividend that exceeded the return elsewhere, you would utilise your money elsewhere until the last 2mths as the return is divided by the time held for an annulised return.

Of course, we cannot predict the future, therefore we have developed this concept of the probability of the set-up etc.

Taking the mechanical system as a first example.
The exit, is random.
It is defined as an exit when price closes below a 180ema
It's results are a picture perfect 50%

Method #1............2.90% return.
I will have to go back and count up all the trades
Ok............Total trades 70
Wins..............................43
Losses.........................27
%...................................63%
So that statistic will provide additional information immediately, and that is, the risk management on this system is weak, the losses in % terms must be nearly equivelent to the % returns on the winning trades, as in essence, this system is standing still.

Still consistent with all the others. Which comes right back to that circa 50%
Therefore system #1 also has a random probability.

The system returning 87%, was still a random probability, 50%.
The difference in % returns is the assumption of huge risk, viz. trading 100% of the account per trade, and wins in % terms larger by on average 1%.

Average win 1.5% return on capital
Average loss 0.5% loss on capital.
The probability.............................50%

You will find this time after time after time.

Where are the Fundamentals different?
I will come to that.

Ducati, I think we have encountered a problem here, which is not uncommon when people attempt to prove a specific point, belief, or opinion, that they already hold. It is the reason I don't simply take "an individuals word for it" People out to prove a point are missing an important component, objectivity. This is why you did not post your own results, as you said, " they were too good, and would not support the point" you were trying to make.

Have a read of those numbers one more time.
Cheers d998

 

[roguetrader]

However, when you actually start to collate the results, you then start to see the actualised results And that is, technical systems return a win ratio of approximately 50%.
Your own results are testimony to the fact.

By and large I would agree that the bulk of the technical strats out there return somewhere in a range of 40-60% win rate, and you could average that to 50% to avoid argument, I have no real issue with that.

However, I guarantee if you had asked someone before the fact, what is the % probability of your set-up, the answer would have come back at a figure other than 50%

That is because the term "probability" is also a word in common use in the English language. The term and the word commonly used have little in common. The term we have already somewhat defined as a mathematical calculation, however the common use of the word in the English language roughly translated means "Looks good to me." Not exactly mathematical.

Not really, as I maintain true scalping is only really performed by professionals who have through whatever arrangement, very, very low, or non-existant execution costs.
Some of the Prop.Shops provide near as damnit zero execution costs.
In that scenario, scalping is viable.
Therefore, the win loss ratio can remain at 50%, and still be profitable, assuming the average winner is 1 tick greater than the average loss.

That I would doubt, I doubt a scalper would make a great living with a 50% win rate but it's probably not worth arguing

Many traders, myself included when I traded technically argued that the entry was all important And while it most certainly is, defining the exit is as important

I would say all aspects of a trade require equal attention to trivialise any aspect invites trouble.

Where are the Fundamentals different?

Well since you chose to include your two trades I will make a start for you. The trades began in early April, circa 3 mnths ago, and returned 36% ish. That is 60 trading days ago approx so the return to date equates to 0.6% per day, regardless of the win rate. With much longer term trades you introduce a new factor into the profit equation, time, an element you have no control over.

 

[ducati998]

rogue

That is because the term "probability" is also a word in common use in the English language. The term and the word commonly used have little in common. The term we have already somewhat defined as a mathematical calculation, however the common use of the word in the English language roughly translated means "Looks good to me." Not exactly mathematical.

Agreed, an in addition, the number of trades that they have analyzed in many cases is so low, as to be totally misleading.

Well since you chose to include your two trades I will make a start for you. The trades began in early April, circa 3 mnths ago, and returned 36% ish. That is 60 trading days ago approx so the return to date equates to 0.6% per day, regardless of the win rate. With much longer term trades you introduce a new factor into the profit equation, time, an element you have no control over.

So based on that, we can see that at 0.6%, the per day return is lower than a technical system.
(Obviously the data is VERY low here, but just for hypothetical reasons)

However two points.
First point is that with the GM trade, a "random" occurance spoilt the party. Now this is part and parcel, but is a factor to consider. And that is, by introducing a factor to increase the return the opposite actually occured, and I was in essence "stopped out"

Second point being, the target was predefined, and the % available, gives an element of probability to the trade

Example.
TELOZ .............entry @ $6.52 target $10.44 = 60% profit margin
NEN .................entry@ $75.80 target $227.76 = 190% profit margin (excluding dividends)

The target values represent fair value and the larger the % available, the higher the probability of the trade.
This is the case as, calculating values is fraught with problems, as is any methodology.
Therefore, the larger the available profit margin, the greater the undervaluation, or the greater the market inefficiency The probabilities in the market always rise the greater the extreme or inefficiency, and reduce back towards random or 50% as the market approaches efficiency.

So while TELOZ was undervalued, it was not as undervalued as NEN currently is. What we are looking for, is not exactitude, but huge inefficiencies in the market that have a true, or higher probability of being corrected.

As regards the element of time, another subject.
cheers d998

 

[LION63]

Ducati & RogueTrader,

Just to add some more data to the ones already being discussed, here are some from a purely fundamental trader operating in short/medium time frames. These are results for the first half of the year 2005.

Total trades 164
Losing trades 25 (15.2%)
Winning trades 139 (84.8%)
Return on Capital employed 363.5%

Please remember that these have been achieved with massive leverage of 20:1.

Total trades 9
Winning trades 9 (100%)
Losing trades 0 (0%)
Return on Capital Employed 15.6%

These trades have been done without the benefit of gearing.

The two types of trades have been shown separately because of their distinct nature and the fact that they are initiated from different trading accounts.

 

[ducati998]

LION

These damned Fundies spoil all the fun........................I was going to get to leverage and expand the discussion as regards to risk assumption and reward calculations.............beaten to the punch

We'll have to have a recap with the results so far posted, and revisit some of the assumptions made to date

cheers d998

 

[LION63]

Ducati,

The problem was that it was said that your sample was not large enough to give a true picture of what might happen if more trades were executed. It has also been stated that we (Fundies) enter trades and sit back for things to happen. There are one or two people that have engaged in meaning full discussion, most notably RogueTrader and I am sure we have all learnt a thing or two along the way.

In order to show that Fundamental Analysis does not mean buy and hope as the misinformed tend to indicate, I have provided my results year to date.

If one were to average out the first batch of trades it is 27 a month and even a Technical Trader would struggle to match that. I will pre empt most of the nay sayers that will definitely say that any fool can do that in a trending or bull market, I must make it clear that of the 164 trades, only 15 were long positions. Of those that were outright stock purchases (the second batch) all 9 trades were long positions.

Surely this has got to be better than the average 50% strike rate that is achieved by TA (please note that I refer to the average as there are a few hot shots that achieve much more than that).

 

[roguetrader]

Ducati

So based on that, we can see that at 0.6%, the per day return is lower than a technical system.
(Obviously the data is VERY low here, but just for hypothetical reasons)

True I would no more critisize an approach based on a smaple that size, than endorse it. My point there was to introduce the time factor which is much more pronounced in any longer term strat. We started with a very respectable 36% return, but when one introduces time in to the equation it becomes very ordinary. As we both agree, the sample is too small and therefore it would be unfair to draw any further conclusions.

The target values represent fair value and the larger the % available, the higher the probability of the trade.

Sounds reasonable, assuming one can accurately value the companies, and assuming that value either holds steady or improves.

Lion

Total trades 164
Losing trades 25 (15.2%)
Winning trades 139 (84.8%)
Return on Capital employed 363.5%

Those are an impressive set of results for a purely fundamental approach, that is an average of 27 trades a month. I would not have expected that sort of turnover from a fundamental approach.

Total trades 9
Winning trades 9 (100%)
Losing trades 0 (0%)
Return on Capital Employed 15.6%

What was the time period of this sample?

 

[LION63]

RogueTrader,

My apologies for not stating that the latter sample is for the same period of January to June 2005.

 

[ducati998]

LION & Rogue

Where does the Fundamental approach deliver................?
In stock selection, time in the market and in profit margin, or return on capital.
The data is thin, as it takes a bit of time to build up the trades, but in essence, and you'll just have to take my word on this currently...................

Probability.............90%
Profitability............73%
Timeframe............Maximum 2yrs

Total trades 164
Losing trades 25 (15.2%)
Winning trades 139 (84.8%)
Return on Capital employed 363.5%

Total trades 9
Winning trades 9 (100%)
Return on Capital Employed 15.6%

Surely this has got to be better than the average 50% strike rate that is achieved by TA (please note that I refer to the average as there are a few hot shots that achieve much more than that).

True I would no more critisize an approach based on a smaple that size, than endorse it. My point there was to introduce the time factor which is much more pronounced in any longer term strat. We started with a very respectable 36% return, but when one introduces time in to the equation it becomes very ordinary. As we both agree, the sample is too small and therefore it would be unfair to draw any further conclusions.

But we can and should analyze the results, as this will provide at least a rational assessment of what is actually occuring in reality.

So far, we have provided evidence that supports the assertion that Technical analysis has an expectancy of 50%.

We would also seemingly be moving towards providing the evidence that Fundamentals provide a 90% expectancy

When we examine the unleveraged returns we see that;
Technicals;
1.........................2.90%..........annualised 5.80%................timeframe January 2005 - June 2005
2.........................16.81%.............................5.60%.....................................February 2002 - May 2005
3..........................87%...............................................................................March 2003 - January 2004
4..........................16.13%
Average.............30.71%...........................28.63%

Fundamentals
1........................36.5%.........annualised 73%......................................April 2005 - June 2005
2........................15.6%..............................31.20%................................January 2005 -June 2005
3.........................leveraged = 363.5% unleveraged = 18.18%....36.36%..Jan. 2005 -June 2005
Average.............23.43%.........................46.85%

So on aggregate returns, and having annualised the returns, always a bit of an assumption, we can see however that the Fundies are sneaking ahead.
Now obviously there are many statistical weaknesses, viz. #of trades, and disparate timeframes, however, it does allow a snapshot into the debate.

What we should start to see as time progresses, is, an increase in the lead by the fundies over the technicals. The reason being that with a much higher expectancy, (the # of wins will rise more quickly than the technicals,also assuming that the reward component remains a constant, and this will return a larger % on capital)

Which is the same way of saying that Fundamentals have less inherent risk, as a methodology, which can allow an intelligent use of leverage, as opposed to a dangerous or reckless use of leverage.

This assertion is also supported in the figures, if we calculate stability.

Technicals stability ratio = 20%
Fundamentals stability ratio = 67%

We can immediately see that the stability, or consistency of the Fundamental results are just over three times as high, again this correlates well with the expectancy ratio.

That is to say, the methodology is less dependant upon the skill ( or the intangibles of the trader ) of the practitioner, and more correlated with the methodology. Thus, the methodology is responsible for identfying market inefficiencies, that have high probabilities, or true probability, of resulting in profit.

From here we can start to move towards quantifying risk and reward, and applying leverage to rev up the $$ returns.

Cheers d998

 

[ducati998]

Have come across some additional live trades.
Here are the available stats.

6mths results from 23 / 10 / 2004 to 23 / 4 / 2005

Technical #1
dr (-3%)

Technical #2
15.5%

Value Fundamental #1
20.9%

Technical #3
0.5%

Value #2
22.4%

Unfortunately, the listed trades have been lost, so the analysis of wins against losses etc cannot be measured, but we can update some of the stats.
Incidentally, these results are from predominantly professional hedge fund managers.

When we examine the unleveraged returns we see that;
Technicals;
1.........................2.90%..........annualised 5.80%................timeframe January 2005 - June 2005
2.........................16.81%.............................5.60%.....................................February 2002 - May 2005
3..........................43.50%..................................87%.......................................March 2003 - Jan 2004
4..........................16.13%..............................32.26%
5.........................dr(-3%)............................dr(-6%)
6..........................15.5%.................................31.0%
7..........................0.5%.....................................1%
Average..............13.19..................................26.38%

Fundamentals
1........................36.5%.........annualised 73%......................................April 2005 - June 2005
2........................15.6%..............................31.20%................................January 2005 -June 2005
3.........................leveraged = 363.5% unleveraged = 18.18%....36.36%..Jan. 2005 -June 2005
4.........................20.9%..............................41.8%
5.........................22.4%..............................44.8%
Average............22.72................................45.43%

Technicals stability ratio = dr(-23%)
Fundamentals stability ratio = 69%

The results have changed quite dramatically as regards the stability ratio. This is significant, as we have added 2 random practitioners of Fundamental analysis, and have gained 2% in the stability ratio. This further supports the assertion of the methodology outperforming rather than the intangible component of individual genius. Whereas, the Technicals have just imploded, due to the losses sustained.

It is the overall consistency in returns, that should further underline the solidity of the approach.
But the 95% of traders, investors lose, statistic, is not being found currently to be accurate.
cheers d998

 

[danfreek]

Your figures mean nothing, anyone can write down unaudited figures on a thread to "prove" their point.

FA 1 day return = 0.001%

TA 1 day return = 10%

wow that really proves that TA must be better!!

Unless there is some fundamental reason why FA is better than TA then your arguement is meaningless.

 

[ducati998]

dan

Actually, if you had bothered to read the postings, the trades are live, posted in real time trades, validated by 3'rd parties, therefore absolutely valid

However, if this threatens your little world, just ignore the results.
cheers d998

 

[danfreek]

You may well say that ducatti, but there is no evidence there, we are just taking your word for it, and even if they are real, they only repersent a selection of trades picked specifically by you to prove your point. Your arguement reminds me of comical ali during the Iraq war "There are no Americans in Baghdad airport, I know, I've just been there" etc,

I actually used to use Fundamental analysis, I listened to the bias coming from the city, that technical analysis is about as useful as astronomy, but one day I decided to read reminicences of a stock operator and I began to think twice about technical analysis, at first it didn't do much good for me, afterall I was only learning, but after about a year of reading and trying techniques I began to be able to read price action etc, now I am making money from it, basically.

My point is that maybe for you fundamental anaysis is more profitable for you, but I would bet that that's because you don't really know how to use it properly, and clearly you're not interested in improving your skills, that's fine, if TA doesn't suit you, then you're better off without it, but that doesn't mean it's s**t it just mean that it's not best for you. FA doesn't make me as much money as TA, but again that doesn't mean I think that it's s**t, just that it's not the best method for me. Exactly the same as saying I don't use MACD, I don't think it's rubbish, it's just that I prefer other indicators.

 

[ducati998]

danfreek

You may well say that ducatti, but there is no evidence there, we are just taking your word for it, and even if they are real, they only repersent a selection of trades picked specifically by you to prove your point.

The evidence is right in front of your nose. They are not cherrypicked, they represent the "live" trades I have been able to find.

Strangely enough it is rather difficult to find large quantities of data on the results of trading methodologies, until you start going to professionals, who have to file their results with the SEC. But of course you run into the Buffetts etc.

These results represent your average retail trader or investor, that is with capital funds not exceeding $5M. Once you get past that sort of figure, the issue of diminishing returns due to the size of funds will start to slow down the hotshots.

My point is that maybe for you fundamental anaysis is more profitable for you, but I would bet that that's because you don't really know how to use it properly, and clearly you're not interested in improving your skills, that's fine, if TA doesn't suit you, then you're better off without it,

You have so missed the point.
We have a sample of 13 different traders, 7 utilising a technical approach, 5 a fundamental approach.

What the results are showing, are the following;
1.....Fundamentals, on aggregate, make larger ROCE than technicals
2.....Fundamentals are more consistent in ROCE than technicals
3.....Fundamentals are less sensitive to individual skill
4.....Fundamentals are more specific to methodology

This is acheived through the following;
1....Greater or higher expectancy
2....Greater or higher stability
3....Defined methodology for identifying true probabilities
5....Utilisation of the principals of selectivity & continuity

As regards the assertion, not being able to use it effectively, who's results do you think relate to the 87% return?

But as previously posted, if the numbers upset your little technical world, don't bother your head reading the thread, as it looks as if the trend may continue for the future.

For those interested in making money consistently, safely, and without being tied to a computer all day, caffeinated to the gills, then read on.
cheers d998

 

[ducati998]

Found some more results;

Total # Trades.....................23
Wins......................................13
Losses.................................9
ROCE...................................5.33%

When we examine the unleveraged returns we see that;
Technicals;
.............................6mths average...........................1year average
1.........................2.90%....................................................5.80%................
2.........................16.81%.................................................5.60%...................................
3..........................43.50%................................................87%......................................
4..........................16.13%...............................................32.26%
5.........................dr(-3%)................................................dr(-6%)
6..........................15.5%.................................................31.0%
7..........................0.5%.....................................................1%
Average.............13.19%..................................................26.38%

Fundamentals
............................6mths average...............................1year average
1........................36.5%........................................................73%
2........................15.6%........................................................31.20%
3........................18.18%.....................................................36.36%
4........................20.9%.......................................................41.8%
5........................22.4%.......................................................44.8%
6.........................5.33%......................................................10.66%
Average...........19.82%.....................................................39.64%

Technicals stability ratio...........................................dr(-23%)
Fundamentals stability ratio...........................................27%
Technical Expectancy ratio.............................................57.50%
Fundamental Expectancy ratio......................................85.45%

Very interesting result, add in a Fundamental strategy, but this time not a value strategy, but a Growth stock strategy and the results plummet.

Both stability & expectancy suffer from the addition of a growth stock strategy.
The technicals, are pretty much at that 50% expectancy.

This was my own personal opinion, and of course 1 result does not a pudding make, however, interesting none the less.

cheers d998

 

[roguetrader]

4..........................16.13%..............................32.26%

Just out of interest, what method of mathematics did you use to annualise that?
In fact as I can see how some of them have been annualised, I'm even more confused as to how the others have.

 

[ducati998]

roguetrader

Just out of interest, what method of mathematics did you use to annualise that?

A number of the sets of results have been for 6mths or less, and others have been for periods greater than 1yr.

So all I have done is to commonalise them all into an average annual figure.
So your result of 16.13% that was a 1month return has been assumed to be an average for 6mths, and then doubled to give an average 1yr return (ie. the returns will continue to approximate the average). This is just to overcome in the first instance the lack of historical data, as 1months results on their own are inconclusive. But as time passes, and the data takes on a nature of historical fact, we shall see if the assumptions are accurate, or not.

This method gives more conservative figures than a true annualisation of the figures.
For example taking 16.13% and annualising we get,

16.13% per month..........193.56% annualised
36.5% over 3mths...........146% annualised
38% per month................456% annualised

Both methods are inaccurate, and are only to give an idea or snapshot of what is occuring.
In the cases of results that are 1yr or greater, then they have been averaged into a 1yr return.

By averaging, we get an index that provides a basis from which to progress.
But I shall ammend the terminology.
Cheers d998

 

[ducati998]

roguetrader

I have now ammended the terminology, to one of "average" returns.

Hmmm, I see where you are questioning the validity, and that is where the actualised returns are annualised, due to enough data and where the returns are projected and averaged ................agreed, a bit of a problem, as it hurts the technicals unduely.

Any suggestions around the problem?
Cheers d998

 

[roguetrader]

16.13% per month..........193.56%

Ok I undertand what you have done though don't see the science behind it, and it is grossly misleading, though in fairness in my case multiplying the one monthe figure by 12 would have been equally grossly misleading, off-hand I think the total return for that year was 57% ish. Worked out in the high 5's as I don't trade for two months of the year.
Trouble is if your figures don't hold for even mathematical scrutiny, your doomed. the 87% is off as well as it looks to be a 10 mnth sample.

 

[ducati998]

roguetrader

See my amended post (previous)
Agreed, this is not, and was never supposed to win a Nobel prize in research, it is really only to start examining some of the truely wild claims made on behalf of the various methodologies.

The problem is that some of the data has been annualised, and makes a poor result, while some has been projected into the future, and looks better, however, there is not really a great way around the problem, until actualised results come in.

cheers d998