Dow Johnes theory

Dow Johnes theory - really a corner stone of the Forex trading analysis. It is one of the best and most popular methods of definition of the main trend of the Forex market. Its bases contain in Charles Dow's published by him during the period with 1900 for 1902 in daily newspaper The Wall Street Journal based by him the works. Dow Johnes theory has been modified in the first decades of 20 centuries by such researchers, as S.A.Nelson, William R.Gamilton and Robert Rea.

Seven postulates of the theory of Dow Johnes:

      1. All essential information on the market contains in the average prices (indexes), first of all in Dow - Johnes industrial and transport indexes. Indexes reflect the information on knowledge, moods, opinions and activity of all traders of the Forex market and, hence, give full representation about all processes influencing demand or the offer of Forex market.
      2. There are three basic trends of the Forex market. The long-term trend of the prices of shares refers to as a primary trend. However any trend is not capable to move long time in the chosen direction, therefore on change primary there comes the secondary trend opposite to it representing correction first, longer. And at last, small trends are the daily fluctuations of the price influencing only on activity of traders, working on small time horizons, and indifferent Dows for the theory.
      3. The primary growing trends known as well as the "bull" markets, will usually consist of three movements upwards. The first movement upwards - result of accumulation of shares the most acute traders during decline of economy at presence of positive forecasts. The second movement of the prices upwards is formed, when traders start to buy up ctions{shares}, learning{finding out} about increase of incomes of the company. And last wave of growth - result of purchases of shares wide layers of the population as consequence of a stream of favorable financial news. Usually at last stage of growth in the market speculators prevail.
      4. The primary falling trends known also under the name of the "bear" markets, will usually consist of three movements downwards. The first movement of shares downwards is caused by sales of the most acute traders which understand, that cost of shares is excessively high, and rates of growth of profits of the company cannot be kept for a long time at the given level. The second movement of the prices downwards - result of the panic arising when the sellers scared by absence of buyers, aspire to leave as soon as possible from the market. Last movement of the prices downwards is caused by a catastrophic wave of sales in connection with necessity for money resources.
      5. Movement of one of indexes should prove to be true movement of another. A signal about the beginning of the "bull" trend is rise of both indexes above maxima of their previous secondary growing reactions to the "bear" trend. A signal about the beginning of the "bear" trend is falling both industrial, and transport Dow - Johnes indexes up to a level is lower than minima of their previous secondary reactions to the "bull" trend. Movement to a new maximum or a minimum of one of indexes is not significant. We shall notice, however, that one of indexes quite often before another gives a signal about change of a trend. In Dow Johnes theory does not contain instructions on size of the period of time after which acknowledgement of a signal by other index becomes void.
      6. The prices of closing of indexes are taken into account only. Price movements within day are not considered.
      7. The primary trend is considered valid until the signal about its change will not be sent by both indexes.

The further development of the Dow Johnes theory

Overall objective of this classical theory - revealing of the basic movements of the Forex market. Similar movements are developed during long time and connected to appreciable changes of the prices. Dow Johnes theory does not specify concrete duration of a primary trend, however the latest researches have shown, that it lasts from one year about several years. The "Bull" market tends to keep more long time whereas "bear", being less long, shows sharper movement of the prices. Some fundamental concepts of the theory of Dow not so long ago have been specified by Victor Sperandeo (see his book " Trader Vick: methods of master Uoll-strit ", John Wiley and Sons, New York, 1991). He has found out, that 75 % of all primary trends of the "bear" markets show falling the prices in an interval from 20,4 % up to 47,1 %. The same of 75 % of the "bear" markets last from 0,8 up to 2,8 years. The "Bull" markets have the big duration: 67 % of such markets last from 1,8 up to 4,1 years.

The secondary trend is a reaction on primary or its correction having an opposite direction. This second wave lasts, as a rule, from 3 till 13 weeks. It cancels third, half or two thirds of changes of the price which were having place during a primary trend more often. Sperandeo has counted up, that while 65 % of secondary trends last from 3 weeks till 3 months, duration from 2 weeks till 8 months had 98 %. Besides 61 % of secondary trends cancelled from 30 % up to 70 % of changes of the price connected to a primary trend. Small trends last usually from 1 day till 3 weeks. Dow Johnes theory does not take into account them as insignificant handicapes. Sperandeo has found out, that 98,7 % of small trends proceeded less than 2 weeks.

It is accepted to name a line the narrow horizontal price range including 10 calendar days and more, - than more long, especially significant. Usually the range is established within the limits of 5 %, however William Gamilton, for example, suggests to count a line of value of 11 % not leaving for limits the prices observed from February till June, 1929. Indexes, as a rule, punch borders of a line in a direction of a primary trend. A similar sort breakdowns are rather reliable. Happens, that the line makes the way in a direction opposite to a primary trend, however these signals hardly probable can be considered quite reliable.

What strong was movement of one index, it is not enough of it for the indication on preparing change of a primary trend, to confirm which necessarily the second index should also. The nonvalidated signals (a situation when one index shows the price of closing exceeding a maximum of a previous secondary wave, and the second index does not reach a maximum) are only warnings that in the near future, probably, the significant signal will follow.

And last important remark: at comparison of the current price of closing of each of indexes with the extreme value shown by the price during the previous secondary wave, all shares down to a penny, that is up to 0,01 without a rounding off are taken into account.

The full cycle of change of the market " bull - bear " will consist of six phases: the scepticism, growing confidence, enthusiasm, blinding, a shock and disgust.

In the powerful "bull" market the first phase is accumulation of shares under the low prices "wise men" of the market - the traders possessing the greatest stock of the information and experience. During this period the general mood in the market is defined by mass scepticism and disbelief. Shares have fallen and can stand on a minimum still long time. The part of traders understands, however, that turn of the market is inevitable, even if now economic conditions are difficult for naming encouraging. Such acute traders start to buy up the unpopular shares offered under seductively low prices. The revolution of transaction, former extremely low, gradually increases, reflecting the introduction into the market of these far-sighted, patient participants.

The second phase begins, when the revolution of transactions more and more increases. In a society the confidence of tomorrow's day grows. The prices for shares raise. Work in the market at this time can appear extremely profitable.

The third phase of the "bull" market happens is marked by general enthusiasm and an abundance of gamble. Fundamental economic prospects look excellent. Everywhere conversations on " a new era " fast economic growth and never-ending well-being are audible. Mass media one by one present histories about huge conditions, earner in the market. Traders stay in high spirits, Shares actively bargain, revolutions reach record levels. However at the end of this third phase the revolution is gradually reduced, as the most active participants of the market by this time have already invested the capital entirely. Besides acute players who in due time had time to buy shares under below cost prices, gradually cease to buy. Moreover, they start to get rid of the purchases on the sly. " Wise men " sell especially willingly, than the crowd, in blinding buy shares under absurdly overestimated prices more recklessly operates.

At last, under an obvious "bull" environment of the market start to appear through while poorly appreciable "bear" tendencies. Analytics observe the Forex trading divergences connected to senseless purchases, made unsophisticated players, and simultaneous far-sighted sales of skilled participants of the market. Rates of growth of shares by this time start to be reduced. The first stage of the "bear" market is marked, first of all, by small reduction of prices which all over again is considered as insignificant market correction; the general moods at this time still extremely good. Meanwhile in a situation when all interested persons already have made purchase, to the prices does not remain anything else how to start to be reduced. When purchased ability weakens, sellers any more do not manage to get rid of shares under the current prices; thus, the prices become lower. The escalating number of shares falls, forming potentially "bear" tendency. However, even when at analysts any more does not remain doubts in the "bear" orientation of the market, the inhabitant stays in happy ignorance. And it is valid: the condition of economy still does not inspire fears, and brokers and dealers, no less than the television commentators paid by them, invariable advise to buy. Public trusts that scientists, strategists and economists Wall street this time "will not bring". Besides to public have inspired, that it is bought "for a long time", and for the long period of the price, certainly, will have time to grow. For this reason falling of the prices is regarded by the blinded traders as " annoying misunderstanding ". Mass purchases would proceed, do not exhaust by this time the inhabitant the capitals. But money is not present more. So, there are no also purchases. The second phase of the "bear" market begins with sudden change of moods: from optimism and hope for the best - to a shock and horror. In one fine day traders suddenly notice, that " king that naked ". Real economic conditions it is far from being so are good, as it would be desirable. To tell under the truth, before economy there is a lot of very serious problems. Public feels, that it is time to get rid of some shares, however buyers any more did not remain. The prices are steadily reduced. On change of greed quickly there comes fear. The market is overflowed with waves of a panic. The revolution grows under shouts of crowd: " Let out us from here! " The quickest professionals try to buy on floor prices as required to sell more expensively, however the market really does not grow: it is observed only " recoil a dead cat ", allowing to compensate only small part of losses. The third stage of the "bear" market are chaotic sales on a background of the common disgust for the market. The economy is in decline, the future is not clear. Movement of the prices downwards proceeds, the truth, in hardly lower rate as potential sellers already got rid of shares under the below cost prices. Even the best shares, till now " keeping on float ", concede to a pressure of "bears". The revolution, former extremely high during a general panic, decreases with the ending of liquidation. There comes the moment when everyone who could sell, have already sold; the "bear" market is exhausted. The crowd swears, that " more never will enter this game ". At last, scenery come back to a stage first act; the cycle is ready to repeat. In a situation when to sell there is nobody more, the prices need - to grow one.

The described phases of the market - a platitude of the economic theory. About them wrote Dow, and his followers already more than hundred years repeat these truisms. And still the crowd is not able to study on own mistakes. The gregarious instinct forces us again and again to give in to a wave of moods, to lose a head and to be carried together with crowd. The trader, not able to distinguish signals of trading indicators and not possessing sufficient discipline, most likely, will fail to operate independently, outside of "herd". We shall remind, however, that " to do as all " - means to lose in turning points of the market, to buy on a maximum, to sell on a minimum and constantly to receive from investments profit below middle market price.

To earn money and to change average market indexes it is possible if to operate in the consent with common sense. Dow Johnes theory gives the answer to questions on how to act in complex market situations.

Criticism of the Dow theory

Despite of the recognition, Dow Johnes theory has not avoided criticism. The basic reproach is reduced to the following: absence of strict definition of secondary reaction has brought in mess in timing of signals of the theory of Dow. Lack as it is seen, essential, however surmountable as we shall show further.

"Desirable never it is impossible to accept for valid, - Robert Pea in the reminds " the History of indexes " (1934). - Two adherents of the theory Dow hardly probable will come to the consent concerning classification and interpretation of many price figures submitted by Dow and Hamilton as indisputable; any attempt of work with them inevitably suffers subjectivity ". Pea warns of danger of thoughtless following behind the theory: " Criticism truly notice, that adherents of the theory Dow frequently are wise backdating... It is confident, that I shall express true respect for Dow Johnes theory if as the impartial observer, in process of forces on a course of the book I shall mark the moments in which indexes or do not warn of occurrence of new trends, or give incorrect data... Both 1917, and 1926, and 1930 show examples of the painful episodes connected to use of the theory of Dow, giving a false signal ".

In Normana G. Fosbeka's opinion, the author of " Logic of the market " (The Institute for Econometric Research Incorporated, 3471 North Federal Highway, Fort Lauderdale, Florida, 33306, 1976, page 9-12), Dow Johnes theory have absence specific criteria: " the Prices for shares, unfortunately, seldom move is ordered, forming strict cyclic figure, - for this reason specific criteria so are difficult for finding out. Followers of the theory Dow result the most different criteria determining signals to purchase and sale... Thus, and dates of the signals received by them, it is appreciable different ".

" The secondary trend - concept rather indistinct... Its true estimation... Puts before users many problems... ". Small trends - one more source of ambiguous interpretation: " Stratifications of these daily fluctuations mislead analysts... The Unequivocal decision of a question on submission of a signal during the critical moment of the market (which, unfortunately, quite often it it appears in a critical situation) is impossible, while researchers will not be left in opinion concerning interpretation of disputable places of the theory of Dow. Even the most skilled of analytics sometimes sharply change type of interpretation, seeing, that the initial purpose appeared unattainable in view of the events which have followed in the market ", - Robert D.Edward and John Medzhi, authors " the Forex trading analysis of trends in the Forex market " Qohn Magee, Inc., 103 State Street, Chicago, 1997, 624 pages) consider.

Dow Johnes theory " not always gives the true forecast, besides the essence of the forecast sometimes remains dark ", - Arthur A.Merrill in the " Behaviour of the prices on Wall street " (The Analysis Press, Chappaqua, NY, 1984,147 pages, page 81) asserts.

" The most difficult problem for Dow analytics, no less than for any trader using strategy of following behind a trend, - to distinguish normal secondary correction of an existing trend from the first phase of the new trend having an opposite direction. Analytics till now did not come to a common opinion concerning that, the signal of turn of the market " how looks, - John Dzh writes. Murphy in " the trading analysis of the financial markets " (New York Institute of Finance, New York, 1999, 542 pages, page 29).

" Estimating results of application of strategy, it is necessary to remember, that signals here - results of interpretation which on a number{line} of models is done backdating. Probably, not all adherents of the theory Dow will agree with my opinion ", - declares Martin J. Pring in " the trading analysis with explanations " (McGraw-Hill, New York, 1991, 521 pages, page 40).

" To understand, where there passes border between small trends and secondary correction, - a problem not always quickly feasible; this unique "dark" place of the Dow theory", - is noticed by Victor Sperandeo (see " Trader Vick: methods of master Wall-strit ", John Wiley and Sons, New York, 1991, page 46).

And still any critical remarks to address of Dow Johnes theory will not eclipse its unconditional importance for the trading analysis. Within hundred years Dow Johnes theory was the tool, irreplaceable at creation of numerous effective and durable strategy. Any future analyst will not seize wholly the skill, not having studied in full details Dow Johnes theory and the data concerning historical efficiency of its signals. Spent time and efforts, undoubtedly, will be paid back wholly.

New horizons of the Dow theory

Strangely enough, not possessing a little appreciable potential of development, Dow Johnes theory could go through successfully the century filled with rough historical events, and including two world wars, world economic depression and numerous revolutions in a science and technics equipments, inconceivable in days of Charles Dow. This phenomenon is represented to even more surprising if to recollect, that Dow considered, that creates the modest barometer forecasting the market for some years forward and that it worked with a small amount of badly organized data and, certainly, without any computer equipment. Be Charles Dou and his followers - S.A.Nelson, William Gamilton and Robert Pea - are alive today, they, undoubtedly, would continue the business with the help of modern programs and the data.

It is obvious, that is logical and precisely organized computer program will cope with processing of the complex data better, than the person. The computer will easily find out in the price data repeating formations there where only chaotic heap of a material will be presented a human eye. Not possessing ours emotionality, not having preferences in the world of hypotheses, the computer will not see a signal there where it is not present, and will not pass during that moment when it will appear. The quietest and disciplined analyst will not be compared to machine in skill impartially to make of the decision. The computer helps the trader to give strict definitions of rules of acceptance of trading decisions according to which the participant of the market then will carry out precisely designated actions. However, it is not necessary to forget: the computer is not able to argue and does not possess common sense, therefore reasonable restrictions should be imposed on its functions. Otherwise it becomes one more source of "noise", so preventing us in work.

New hypotheses of the theory of Dow for computer testing

Hypothesis the first: the objective and exact analysis of the data should be put in a basis of submission of a signal. As we see, the basic problem which till now has been not decided by the theory of Dow, - precise differentiation of a primary trend, secondary reaction and small trends; thus, this problem should be solved first of all with the help of computer programs.

In the general view, excepting all nuances and specifications, a principle of submission of signals, according to Dow Johnes theory, it is reduced to the following: purchase in case of growth up to the value exceeding a former maximum, and sale at falling below former minimum for both indexes. This elementary strategy reminds the technics equipment designated usually as a rule of breakdown of a trading range, known as well as a rule N periods Donchiane and one of turtle trading rules Richard Dennis (is more detailed see. Price channels). It is one of the oldest and elementary models of following behind a trend: we buy, when the day time price of closing rises to new for the last N the periods to a maximum; then we sell and we open a short position when the day time price of closing falls to new for the last N the periods to a minimum. This very exact and certain rule which is not leaving places for doubts and interpretation. To work with such model simply enough.

Hypothesis the second: the length of the period should vary depending on the type of a position assumed by a signal. The statistical data published by Robert Rhea in 1930th years and Victor Sperandeo in 1991, have shown, that the "bull" and "bear" markets appreciablly different on duration and extent. Thus, the length of the period for signals to purchase and sale also should differ. Moreover, for each of four possible actions (to open a long position to close a long position to open a short position to close a short position) the length of the period also can have the specificity. Hypothesis the third: the length of the period for each of two indexes can vary independently. As historical behaviour of each of Dow - Johnes two indexes - industrial and transport - variously (we shall remind, that they quite often miss in opposite directions), we shall establish special parameters for each of indexes.

With the help of three resulted hypotheses we have an opportunity to build full, covering all opportunities of the market, strategy. We shall choose till two lengths of the period for each of four probable actions (to open a long position to close a long position to open a short position to close a short position): one length for the prices of closing of an industrial index of Dow - Johnes, another - for the prices of closing of a transport index of Dow - Johnes.

Having four types of actions (to open a long position to close a long position, etc.) And two types of the prices of closing, we should test eight indicators for each model. We can vary number of lengths of the periods (the size of area of change of parameters) for each indicator. Luis Mendelson (see his "Creation and testing of trading systems: how to avoid mistakes which dearly will cost for you ", Mendelsohn Enterprises, 25941 Apple Blossom Lane, Wesley Chapel, FL 33544, has noticed, that the increase in area of change of parameters conducts to geometrical increase of quantity of models. For example, if for each of eight indicators we shall want to test till three lengths of the period, the amount of models will be equal to three in the eighth degree, that is 8 = 6561. If we shall want to test one more length of the period the amount of models will sharply increase - up to four in the eighth degree: 48 = 65536 models. Addition of this one parameter makes impossible backtesting for the software available in our disposal which in a condition to analyse for one reception of 32 thousand models. The software is obvious, that powerful, in comparison with accessible in the past, it appears insufficient for testing models of similar complexity.

Fortunately, we do not have need to limit itself, working as all with three parameters. It is much more convenient to divide research into two parts: only short and only long positions. Having four indicators, we can test 13 lengths of the periods as 13 in the fourth degree it is equal 28561. Having carried out the analysis of long and short positions separately, we shall unit result in one model. Improving it, we also shall achieve, at last, desirable result. We shall notice, however, that, segmenting testing, analyst risks to miss the best combination of parameters and to make the decision which is not being optimum.

After many iterations by us the following result has been received:

To open a long position (to buy), when Dow - Johnes industrial index reaches a new maximum for the last of 9 trading days, and Dow - Johnes transport index reaches a new maximum for last 39 days.

To close a long position (to sell), when Dow - Johnes industrial index reaches a new minimum for the last of 22 trading days, and Dow - Johnes transport index falls up to a new minimum for the last 166 trading days.

To open a short position (to sell shortly) when Dow - Johnes industrial index reaches a new minimum for the last of 22 trading days, and Dow - Johnes transport index falls up to a new minimum for the last 166 trading days.

To close a short position when Dow - Johnes industrial index reaches a new maximum for the last of 36 trading days, and Dow - Johnes transport index reaches a new maximum for the last of 32 trading days.

Dissymetric rules of acceptance of trading decisions means, that we not always have an open position. We shall notice, that we buy at very sensitive, short-term signal: the 9-day's new maximum of an industrial index confirmed with a 39-day's new maximum of a transport index. Thus, to receive a signal to purchase it appears simply enough. And on the contrary, the signal to sale and opening of a short position turns out hardly: it is necessary to wait, while the industrial index will update a minimum of last 22 days at presence of confirmation from the transport index fallen up to a new minimum for last 166 days. Hence, the mechanical trading model has distinguished dominating long-term character of the "bull" tendency in the market which more than time grows, than falls, and shows the greater final rise of the price, than its falling.

Analyzing the data from January, 2, till February, 16, 2000, it is possible to draw a conclusion, that above mentioned rules of acceptance of trading decisions in many cases allowed true signals of purchase and sale. Offered trading strategy does not leave doubts concerning recognition of the signal, and also concerning time of the transaction determined by it and a level of the price. Have the trader an opportunity to use the given trading strategy during all period in 101 year, it would receive result on 5637,10 % better in comparison with strategy " buy and hold ". The general net profit would make $1233454,40. Our more complex rule of acceptance of trading decisions has caused more active trade: one transaction on the average time in 290,83 days. From all 127 transactions 69, that is 54,84 %, appeared advantageous.

Construction of the strategy based on crossing by one exponential sliding average with both indexes of Dow - Johnes

Hypothesis the fourth: construction of the price channel and the analysis of its breakdowns - the technics equipment, allowing to define movement of an index inside horizontal frameworks; we shall remind, however, that the market quite often goes upwards or downwards in an inclined direction. In this case use of inclined lines will make process of submission of trading signals by more productive. Exponential moving average - the inclined line, which problem, with reference to both indexes, is reduced to confirmation of a trend or to submission of a signal about its change. Research of the historical data shows, that strategy on the basis of analysis Exponential moving average is the effective tool at work both on long, and on short time horizons, however especially successful results manage to achieve on short time horizons. All ranges in length of 100 days also have less shown results, the best, rather than strategy " buy and hold ". For traders with low charges on carrying out of transaction of the optimal would be Exponential moving average length about three days.

To open a long position (to buy) under the current day time price of closing of an industrial index of Dow - Johnes when this price of closing rises above 3-day's Exponential moving average the day time price of closing for a yesterday, and the price of closing of a transport index of Dow - Johnes also rises above 3-day's Exponential moving average the prices of closing for a yesterday.

To close a long position (to sell) under the current day time price of closing of an industrial index of Dow - Johnes when this price of closing falls below 3-day's Exponential moving average the day time price of closing for a yesterday, and the price of closing of a transport index of Dow - Johnes also falls below 3-day's Exponential moving average the prices of closing for a yesterday.

To open a short position (to sell shortly) under the current day time price of closing of an industrial index of Dow - Johnes when this price of closing falls below 3-day's Exponential moving average the day time price of closing for a yesterday, and the price of closing of a transport index of Dow - Johnes also falls below 3-day's Exponential moving average the prices of closing for a yesterday.

To close a short position under the current day time price of closing of an industrial index of Dow - Johnes when this price of closing rises above 3-day's Exponential moving average the day time price of closing for a yesterday, and the price of closing of a transport index of Dow - Johnes also rises above 3-day's Exponential moving average the prices of closing for a yesterday.

Dow Johnes theory: ways of evolution

The ideas schematically outlined here, everyone can use for own researches. The purposes and the most comprehensible means of their achievement for concrete traders should underlie processing of traditional strategy. The set of the indicators submitted in our book, in a condition to become additions of the theory of Dow, develop which on the basis of the careful analysis of the historical data everyone analyst can on own fear and risk.

Let's remind main principles of evolution of any theory. The researcher saves the data, forms a hypothesis, then the hypothesis is checked. The hypothesis can undergo repeated changes, which essence - to put the theory in conformity with really observable phenomena. The theory varies and when analytics the new data become known.

Simple consideration of the data without scientific testing - a way which inevitably results in creation of false hypotheses, to deducing of the incorrect conclusions and fulfilment of rash acts. The rule, faultless on a kind, during check on a real material can find out set of incongruities. Backesting is an excellent way of clearing of dark places. Without backtesting outside of a zone of our attention sometimes there are as hardly perceptible, but essential nuances, and the gradual changes of the data which were having place eventually. Self-confident analyst, not checking on the basis of the historical data the concepts, most likely, appears in authority of subjectivity and biases. Within long years Dow Johnes theory not time has been understood incorrectly. The matter is that - absence of accurate informations and constantly spent tests. Any concept, even so powerful classical theory as Dow Johnes theory, requires specifications. The testing spent by analysts, should be unbiassed, exact and objective. At all stages of check the researcher is obliged to realize precisely as as is going to undertake. Built models should differ severity; experience and skill to argue - the basic qualities, capable to provide success.

The best way of testing - modelling of trade on the historical data, allowing on a material of the last years in detail to study functioning rules of acceptance of the trading decisions giving in the past the best result. In our opinion, it not only the most effective, but also is unique an effective kind of check. Everyone theory remain unsteady until will be supported by strong bases of testing. The unique opportunity to create the strategy, capable to sustain a peripetia of the market, - to design artificial model where hypothetical efficiency will be shown on the real historical data.

Copyright 2007 Gordago. All rights reserved.
Copying of all making parts of a site in any form without the permission of holders of copyrights is forbidden.
EnglishEnglish RussianRussian