51. An example of good automated trading system: Spike Hunter

ANYONE WHO WILL TRY TO MAKE MONEY SOLELY BY TRADING AND SUCH SYSTEMS OF TRANSACTIONS SHOULD BE AWARE THAT THERE IS A VERY POWERFUL AND ALMOST UNBEATABLE COLLECTIVE WILL SO AS NOT TO SUCCEED!  NO-ONE WANTS  PEOPLE TO QUITE THEIR JOBS AND MAKE MONEY THIS WAY AS IT IS SOMEHOW PARASITIC. IT IS IN SOME SENSE  UNETHICAL AS A PRACTICE ENFORCEABLE  TO  THE MAJORITY. AND OF COURSE NEITHER THOSE WHO HAVE LARGE CAPITAL  WANT THAT A MAJORITY WILL MAKE MONEY THIS WAY, AS THEY WOULD PREFER THAT THEY WORK IN THEIR COMPANIES FOR THEM. ONLY IN SPECIAL CONTINGENCIES AND SITUATIONS SOMETHING LIKE THIS WOULD BE ETHICAL. AND IN PARTICULAR A HIGHER MORALITY THAT WOULD SUPPORT SUCH A PRACTICE, WOULD BE PROVABLE WITH COLLECTIVELY BENEVOLENT DEEDS FROM A POSSIBLE SURPLUS OF SUCH MONEY!

There are two mathematical principles on which systematic long run successful trading is based:

1) There are hidden almost periodicities on the volumes and volatility of the prices. The most characteristic are , the daily cycle , the monthly cycles and the annual (or quarter) cycle.

2) The duration of the trends is usually a multiple of the characteristic half-periods  as above, and they also follow approximately the distribution of the capitalization sizes of enterprises, in other words a Pareto or a Power distribution. 
3) The demand-supply  functions under the three coupling modes a) Domination b) Competition c) Cooperation, and as result of this the 3 basic price patterns of spike, trend, and flat channel, emerge. 

If a uniform randomness would rule the markets (efficient markets assumption) , the probability of long duration trends would be much less that what it is observed, and their distribution would be a geometric or exponential, and not Pareto or power. Thus the law of "universal attraction" in enterprise formation makes also the markets non-efficient and more predictable that the academic world believes.

Spike Hunter

Spike Hunter is a fully automated strategy. It opens positions whenever spikes appear.
In addition it trades at reaction spikes, of terminal spikes of trends.   Therefore there is the psychological and risk management advantage that we should not fear sudden crisis, unpredictable news based crisis etc. In all such crises it locks profits. The logic behind the algorithm applies to the world balance and competition of the two major currencies only, the US Dollar and the Euro (EUR/USD). US-dollar down-spikes  or EU-euro up-spikes are stronger and usually related to panic.

  Spike Hunter  avoids opening positions  at the particular time when all bank-sessions are closed (Asian, European, and American).  It takes advantage of the daily periodicity of the volumes of transactions and its dynamics to open positions. Trailing and pyramiding and group of open positions  take profit, are utilized.

Spike Hunter makes on average 0.8  trade per day but when it opens positions usually it is more than one in the same day. There are weeks that no position is opened.

 Usually the trades do not last more than a 24 hours. Approximately 91% of the trades are winning, and with a maximum annual draw down less than 10% it has an annual rate of return of about 20%-30%. If the profits are reinvested, and the positions size proportionally increased to the balance, the annual rate of return is even higher. About 2/3 of the months are profitable.
The income from forex automated trading is on the side of counter-balancing the over-debt that the privatised banking-monetary system creates in society.

 Some  good features of the spike_hunter are

0) The spike hunter as the title suggests, is based on the appearance of spikes at the hourly time frame. Therefore panic and extreme crisis in the markets is not something to fear in this system, but on the contrary it locks profits. The system performs exceptionally very well at the years of crisis 2007-2008 etc.
1) it makes about 1 trade per day (0.8 exactly) 
2) it has 91% trade success rate 
3) At 60%maxDD it has 100% monthly rate of return (=1mds=1monthly doubling speed). At 33% maximum Draw Down has an average monthly rate of return of about 40% (measured for the last 6 years:2006-2011).
4) Contrary to scalpers the robot avoids trading during  the hours that all banks are closed: sessionless time zone 
5) It utilizes the daily volumes periodicity to open trades on spikes
6) A further remark about why all trades are buy: There is no restriction for sells. But the filters of favorable spikes give the buy spikes so far. No guarantee that in the future there will not be  a sell spike. Panic spikes are by far stronger than optimistic spikes. And US behaves much more emotionally than EU. panic-spikes in US are buy-spikes in EURUSD

As seem from the performance below and the classification of commercial  forex algorithmic trading robots of the web (in http://www.bestforexrobots.net/ ) the Spike Hunter robot is among the best 10 robots of the web.

Comparisons with the CD-auto-fast (post 50)

Spike Hunter

Advantages: 90% succes rate in the trades

Disadvantages: Average profitable trade is 5 times smaller than the rare avearge loss trade

(this creates the effect of gaining, gianing, gaining then a larger loss , etc)
Notice that the slope of the diagram of rate of return versus maximum draw down in the spike-hunter is of significant larger slope from the corresponding of Carpe-Diem that is why at 30% maxDD it has almost double the performance of Carpe-Diem.

While CD-auto-fast:

Advantages: Average profit trade 1.3 times larger than average loss trade

Disadvantage: Low trades success rate: 49%

(this creates the effect losing, losing losing, then a larger gain etc)

Strategy Tester Report

Spin_navigation_weaver_9

AlpariUK-Micro-1 (Build 419)

Symbol EURUSD (Euro vs US Dollar) Period 15 Minutes (M15) 2007.01.01 00:00 - 2011.12.30 23:14 (2007.01.01 - 2011.12.31)

For this type of expert-advisors, the maximum floating drawdown (during the open positions) is calculated by the code and does not exceed the published Maximum Drawdown which is at the closing or opening of the trades.

Notice that the slope of the diagram of rate of return versus maximum draw down in the spike-hunter is of significant larger slope from the corresponding of Carpe-Diem that is why at 30% maxDD it has almost double the performance of Carpe-Diem.

For publication and detailed analysis of the above backtest by the MT4i

http://www.mt4i.com/users/spikehunter_backtest/stats

We may compare these results with the trading results of Larry Williams, which are close to 0.47MDS here
http://en.wikipedia.org/wiki/Larry_Williams_(a_publisher_and_promoter_of_trading_ideas)

Or with the trading results of Chuck Hughes that also about 0.4MDS
here  http://www.chuckhughesonline.com/

Or with the trading results of Dan Zanger that are about 0.42MDS here 

http://en.wikipedia.org/wiki/Dan_Zanger

50. An example of a good automated trading system: Carpe Diem

ANYONE WHO WILL TRY TO MAKE MONEY SOLELY BY TRADING AND SUCH SYSTEMS OF TRANSACTIONS SHOULD BE AWARE THAT THERE IS A VERY POWERFUL AND ALMOST UNBEATABLE COLLECTIVE WILL SO AS NOT TO SUCCEED!  NO-ONE WANTS  PEOPLE TO QUITE THEIR JOBS AND MAKE MONEY THIS WAY AS IT IS SOMEHOW PARASITIC. IT IS IN SOME SENSE  UNETHICAL AS A PRACTICE ENFORCEABLE  TO  THE MAJORITY. AND OF COURSE NEITHER THOSE WHO HAVE LARGE CAPITAL  WANT THAT A MAJORITY WILL MAKE MONEY THIS WAY, AS THEY WOULD PREFER THAT THEY WORK IN THEIR COMPANIES FOR THEM. ONLY IN SPECIAL CONTINGENCIES AND SITUATIONS SOMETHING LIKE THIS WOULD BE ETHICAL. AND IN PARTICULAR A HIGHER MORALITY THAT WOULD SUPPORT SUCH A PRACTICE, WOULD BE PROVABLE WITH COLLECTIVELY BENEVOLENT DEEDS FROM A POSSIBLE SURPLUS OF SUCH MONEY!

There are two mathematical principles on which systematic long run successful trading is based:

1) There are hidden almost periodicities on the volumes and volatility of the prices. The most characteristic are , the daily cycle , the monthly cycles and the annual (or quarter) cycle.

2) The duration of the trends is usually a multiple of the characteristic  half-periods as above, and they also follow approximately the distribution of the capitalization sizes of enterprises, in other words a Pareto or a Power distribution. 
3) The demand-supply  functions under the three coupling modes a) Domination b) Competition c) Cooperation, and as result of this the 3 basic price patterns of spike, trend, and flat channel, emerge. 

If a uniform randomness would rule the markets (efficient markets assumption) , the probability of long duration trends would be much less that what it is observed, and their distribution would be a geometric or exponential, and not Pareto or power. Thus the law of "universal attraction" in enterprise formation makes also the markets non-efficient and more predictable that the academic world believes.

Carpe Diem automated

Carpe Diem Auto is a fully automated strategy. The logic behind the algorithm applies to the world balance and competition of the two major currencies only, the US Dollar and the Euro (EUR/USD). It utilizes a coupling competition oscillation of Demand-Supply as created by the European trading session and then later, the American trading session. It is also based on the statistical fact that it is 4%-5% more probable each day to have a trend opposite to the previous day’s trend.

Carpe Diem Auto begins opening positions only at a particular time when all bank-sessions are closed (Asian, European, and American). It utilizes the daily periodicity of volumes to choose a time to measure the trend, which is optimal for forecasting the next day’s trend.  It takes advantage of the remarkably statistical stable behavior of this oscillation competition of Europe and USA markets. It measures carefully the deceleration (and acceleration) of the trend, which is always prior to reversals and news releases.

 Carpe Diem Auto makes on average one trade per day and it never risks more than 0.4% of account balance per trade, and not more than 1%-2% (floating) of all open trades simultaneously. It always has on each trade, a tight invisible stop loss from 10 pips of 50 pips and a visible stop loss of 50 pips (almost 1/3 of the average daily variance). Rarely the trades last more than a 24 hours. Approximately 77% of the trades are winning. With the conservative risk settings of a) 0.4% funds risked per trade and b) 1%-2%  on all open positions (floating draw down) it has a maximum sequentially draw dawn per 5 years, less than 5%, and average annual rate of return of at least 15%  which goes much higher ( 30%) if the profits are reinvested and the positions size are increased proportionally to the balance.  With higher risk settings a) and b) as above, all the previous figures scale up more than proportionally.
About 80% of the months are profitable.

The income from forex automated trading is on the side of counter-balancing the over-debt that the privatised banking-monetary system creates in society.

As seem from the performance below and the classification of commercial  forex algorithmic trading robots of the web (in http://www.bestforexrobots.net/ ) the Carpe Diem robot is among the best 15 robots of the web.

Comparisons with the Spike Hunter (post 51)

Spike Hunter

Advantages: 90% succes rate in the trades

Disadvantages: Average profitable trade is 5 times smaller than the rare avearge loss trade

(this creates the effect of gaining, gianing, gaining then a larger loss , etc)

While CD-auto-fast:

Advantages: Average profit trade 1.3 times larger than average loss trade

Disadvantage: Low trades success rate: 49%

(this creates the effect losing, losing losing, then a larger gain etc)

The next is a backtest since 2006 till 2012. It was run on hourly bars. The expertadvaisor is set to run only at the opens of the hourly bars. 

Strategy Tester Report

Spin_navigation_weaver_9

AlpariUK-Micro-1 (Build 419)

This is a table that shows how rate of return and maxDD depend on the lot size

For this type of expert-advisors, the maximum floating drawdown (during the open positions) is calculated by the code and does not exceed the published Maximum Drawdown which is at the closing or opening of the trades.

CarpeDiem Account_15,000         Test Horizon 2011          Trades success rate 75%

For a 5 years backtest since 2008 as analyzed by the MT4i see

http://www.mt4i.com/users/cd_auto_backtest/stats 

An alternative version of the system is the Carpe Diem fast

Carpe Diem Automated  Fast

Carpe Diem Auto Fast is a fully automated strategy. The logic behind the algorithm applies to the world balance and competition of the two major currencies only, the US Dollar and the Euro (EUR/USD). It utilizes an oscillation of the coupling of Demand-Supply, were each day it is 4%-5% more probable to have opposite trend from the trend of the previous day.

 Carpe Diem Auto Fast begins opening positions only at a particular time when all bank-sessions are closed (Asian, European, and American).  It takes advantage of the daily periodicity of the volumes of transactions to measure the trend for optimal forecasting. Trailing and pyramiding are utilized.

Carpe Diem  Fast makes on average 1.8  trade per day and  has never more than 2 simultaneously open positions

It always has a tight invisible stop loss from 10 pips of 50 pips and a visible stop loss of 50 pips (almost 1/3 of the average daily variance), on each trade. Usually the trades do not last more than a 24 hours. Approximately 50% of the trades are winning, and with a maximum annual draw down less than 10% it has an annual rate of return of about 30%. If the profits are reinvested, and the positions size proportionally increased to the balance, the annual rate of return is even higher.About 60% of the months are profitable.

The next is a backtest since 2006 till 2012. It was run on 15-minutes bars. The expertadvaisor is set to run only at the opens of the 15-minutes bars. 

For this type of expert-advisors, the maximum floating drawdown (during the open positions) is calculated by the code and does not exceed the published Maximum Drawdown which is at the closing or opening of the trades.

CarpeDiem_fast Account_20,000         Test Horizon 2006-2011                   Trades success rate 49%
Average_Monhly_rate_of_return

For a 5 years backtest since 2008 as analyzed by the MT4i see

http://www.mt4i.com/users/cd_autofast_backtest/stats

We may compare these results with the trading results of Larry Williams, which are close to 0.47MDS here
http://en.wikipedia.org/wiki/Larry_Williams_(a_publisher_and_promoter_of_trading_ideas)

Or with the trading results of Chuck Hughes that also about 0.4MDS
here  http://www.chuckhughesonline.com/

Or with the trading results of Dan Zanger that are about 0.42MDS here 

http://en.wikipedia.org/wiki/Dan_Zanger

Also other list and reviews and ranked by voting of EA are here

http://www.myfxbook.com/reviews/expert-advisors/10,1