Part 1
Groundwork
When we are developing a trading system, we first need to identify its basic elements:
1. Time period: What is the time frame for the system? Is this a system ford ay trading or intraday trading?
This is a trading system for intraday trading.
2. Type of system: Will the system be one of the two main types i.e. a
trend following system or a trading range system or will it have a
particular focus within one of these two types?
We would like to be active on the market for as much time as possible so
our system is designed to work well on both trending and ranging
markets.
3. Tools: What tools will the system use to generate signals? By tools I mean technical indicators (e.g. RSI or MACD); chart patterns (e.g. lines and levels or the classic Western patterns like head and shoulders and double bottoms) and Japanese candlestick patterns (1 candle patterns; 2candle patterns or 3 candle patterns)?
support and resistance levels are our primary tool. The signals that support and resistance levels give us are confirmed
by the signals of a number of other tools. We use a number of one and
two candlestick patterns because they signal quickly. We will also
employ some technical indicators namely RSI and Stochastic.
4. Construction: How will the system combine the different tools? How
will it prioritize signals? What will it do if the signals contradict
each other?What confirmation is needed for each signal?
We combine the signals by giving each a ‘weight’. The signal that
each tool gives has a point rating. We add these points together and we
only open a position if the total is above a certain amount.
This means that if the signals for the hour ending 11.00 add up to a
tradable position then the opening price should be the opening price of
the period ending 12.00. What tends to happen is that the opening price
of your position will most likely be different from the opening price of
the period because you will need some time to check the signals and
rate the trade.
5. Further considerations: What additional factors should the system take into account?
The existence of a trend; the day of the week and the unexpected.These
factors and their ratings will be dealt with in the rules of the trading
system.
Part 2
Rules of the system
Rule #1: Bouncing off a level
When price hits a level of support or resistance it can do one of two
things.It can either ‘bounce off’ from the level or break out through
it. We believe that it would be a pretty tall order to design a trading
system which could generate signals for both a bounce off (reversal) and
a breakout. We reckon that you would need two trading systems for this.
We are only designing one system here and we have decided to design one
that identifies reversals from a level. This is because reversal
happens much more often than break out.
First we need to define what we mean by a reversal:
- If price reaches a support level and then starts rising; or if it
breaks through this level only to rise back above it and keeps rising
after that.
- Likewise: If price reaches a resistance level and then falls back or
if it breaks through the level only to fall back below it and keeps
falling after that.
We want to stress
that support and resistance are in fact zones with a width of a number
of points rather than one-point wide levels. On intra-day charts we take
these zones to be 10 points wide. This means +/-5 points from the
precise level we chart. So, if price reverses within 5 points of a
level, then we take this as a reversal from that level.
Different levels have different strengths and the stronger the level
is the more likely that price will reverse from it. Here are the ratings
for the different levels:
- A level confirmed on the daily or weekly chart: 5 points.
- A level confirmed on 6-hrly charts: 4 points.
- A level confirmed on hourly charts: 3 points.
- A level confirmed on 30-minute charts: 2 points.
The first rule of our system is that we are only looking at reversals
that occur in our zones of support and resistance. These zones surround
a level which can have a maximum strength of5 points. We will only be
considering levels that have been retested in the last four full weeks.
Now we’re going to hand over to Peter Eriksson, one of our top
analysts and a man who likes to tell it as it is. This is his take on
those mystery men:the market makers.
Who are the market makers? They are the fat cats who employ analysts
and academics as consultants. I really don’t believe that any of the
‘market makers’ in question ask these nerds what Stochastic is doing
right now or whether the MACD histogram is signaling a good time to buy.
Who knows how they decide when and what to buy. It’s down to
guess-work. If you’re asking me, the only important factor is the bottom
line: the prices. What is the high and what is the low? Where can he
buy and sell and when? The price is the market maker’s
only reference point. Indicators and what-not are for the birds, using
them takes graft and skill: qualities of the poor and needy. The highs
and lows on the market are the only benchmarks. They are all you need.
But which highs and lows should we be looking at? Forget about
intraday charts: the market-makers don’t check them; their lives are
full of other distractions. Day charts are perhaps more telling, but big
players don’t place orders everyday and I use the highs and lows of the
week, at the very shortest, as my guide. A further refinement of my
system is that these weekly highs and lows are what Tomas Demark called
base points. Demark described a base point as being made up of three
candles with the centre candle higher (in the case of a High base point)
or lower (in the case of a Low base point) than the two flanking
candles. Take a look at the EURUSD chart:
- The May 27th candle is both a base point and the week high: it’s a significant high and we can chart a level from it.
- The May 13th candle is a base point because the lows of both the12th
and the 14th of May are higher than it. Unfortunately, it is not the
minimum for week 12th – 16th May and so no level is charted. However,
the May 16th candle is both the weekly low and a base point and we can
chart a level.
- The May 5th candle is the weekly low, but it is not a base point, so no level.
What Andrei is telling us here is that the highs and lows of the week
that are also base points give us significant levels of support and
resistance.
We also use shorter-period candles for less emphatic (but more numerous) signals. We monitor support and resistance on six-hourly candles for intra- day work.
Plotting support and resistance levels on a 6-hour chart
Take a look at the EUR (6-hourly) chart. The chart also features the
Price Channel indicator, with the time period set at 4 (4 x 6hr = 24hr)
giving us the high and low for the previous day.
We plot support and resistance levels when Price Channel is
horizontal.The arrows mark some of the horizontal sections we have used
to generate levels of support and resistance. We only chart the levels
we feel price can reach today. We should be flexible and ready to bring
in other levels as and when the situation changes and we limit the price
history that we use to the last six weeks.
When a number of lines are close together, we have a ready-made
support and resistance ‘zone’, with its midpoint running at the average
of the levels. In the chart we have a level at 1.1651 (1), another at
1.1654 (2)and a third at 1.1657 (3). We treat these levels as part of
the same support zone that has a ‘spine’ at 1.1654.
Notice how often the chart confirms certain levels as support/resistance.
One final point is that although this system may not identify support
and resistance exactly, our 10-point wide zone rule for intraday charts
means that any inaccuracies will be filtered out.
You may like to experiment with Rumus2 using hourly charts with your
Price Channel time period set at 24 (24 x 1hr). This will give you the
same levels of support and resistance but we think that you’ll agree
that the chart does not display the key info as emphatically and that
there is far less interaction between the candles and the channel
borders. Try it and see for yourself.
With a little practice you will soon be easily able to identify
important levels of support and resistance on a six-hour chart without
the assistance of Price Channel or any other indicator.
Rule #2: Candlestick patterns
This section identifies the candlestick patterns that our trading
system uses. You’ll notice that our patterns include some of the
classics (like the Dragonfly and Harami)
and some of our own (the Ice and the Umbrella). Each pattern is given a
rating. You can incorporate other candlestick patterns into your system
if you want. However, in our opinion the ones on our list will be
sufficient. In intra-day work speed is the important thing and many
classic patterns like Three Black Crows or Three White Soldiers can only
be positively identified once the third candle has closed when it maybe
way too late to open a profitable position based on the initial signal
of two hours before.
One tip for anticipating significant patterns on hourly candles is to
monitor30-minute candlesticks. If you get a clear pattern on these
candles, the chances are you will get a similar strength pattern on your
hourly sticks.
Our second rule is that candles confirm price bounce-off. If you
flick through 2.2.1, you’ll see that the maximum confirmation rating is 4
points.When two patterns appear at the same time (e.g. tweezers and the
ice) you should only rate the stronger pattern.
Bearish signals (confirmation to sell):
Bullish signals (confirmation to buy):
Rule #3: Stochastic reversal
Most experienced ‘techies’ are familiar with Stochastic and the usual
way to use it. We are going to ‘tweak’ the indicator a bit.
Stochastic settings for our system are (6, 2, 1). 6 is a cycle within the trading session. 2 smoothes the signal without too much lagging and 1removes the %D line which our system does not use.
Set the overbought and oversold levels at 70 and 30.
We focus on the reversal of the Stochastic line in the direction we
want rather than the more usual crossing of levels 70 and 30.
Here are the ratings for our signals:
- When Stochastic is in the overbought zone (70-100) and reverses down,it is a confirmation signal to sell with a rating of 2.
- When Stochastic is lower than 70 and reverses down, this confirmation signal to sell is rated 1.
- When Stochastic is in the oversold zone (0-30) and reverses up, it is a confirmation signal to buy with a rating of 2.
- When Stochastic is higher than30 and reverses up then it is a confirmation signal to buy and is rated 1.
- When there is no reversal of Stochastic, there is no signal.
To sum up then, rule #3 states that bounce off can be confirmed by a
reversal in Stochastic and that Stochastic reversal can give us a rating
of2 at the maximum.
Please note that Stochastic also features in rule numbers 5 and 6.
These rules are about divergences between the trading system’s
oscillators(Stochastic and RSI) and price. Divergence (rule 5) and
double divergence(rule 6) are more significant signals than Stochastic
or RSI reversal.
Rule #4: RSI reversal:
RSI is a popular oscillator and you’ll no doubt be pretty familiar
with how to use it. Once again we are going to ‘tweak’ it for our
trading system. Set the overbought and oversold levels to 60 and 40.
Our RSI time period setting is 9 and the averaging setting is 1.
As with rule 3 for Stochastic reversal, rule 4 focuses purely and simply on RSI reversal in the direction we want.
Here are the signals and ratings:
- When RSI is in the overbought zone (60-100) and reverses down, then this is a confirmation signal to sell with a rating of 2.
- When RSI is less than 60 when it reverses down then this confirmation signal to sell has a rating of 1.
- When RSI is in the oversold zone level (0-40) and reverses up, this is a confirmation signal to buy and has a rating of 2.
- When RSI is above 40 when it reverses up, then this confirmation signal to buy is rated 1.
Once again divergence is not covered in this rule. Divergence is covered in rules 5 and 6.
To sum up, rule #4 states that price bounce-off can be confirmed by
an RSI reversal. RSI reversal gives us a maximum rating of 2.
USDCHF hourly chart RSI and Stochastic reversal:
Rule #5: Simple divergence:
Simple divergence is a divergence in the direction that price and one
of our oscillators are moving. For example, a simple divergence is
signaled when price sets a new high which is higher than its previous
high while one of our oscillators fails to signal a corresponding new
higher high. At the other end of the market, a simple divergence occurs
when price reaches a new low which is lower than its previous low but
the corresponding low set by the indicator is not a new low. We cover
divergence in some depth in our film,The Essential Indicators. Our
trading system rates a simple divergence between price and either
oscillator (Stochastic or RSI) as follows:
- Divergence when price bounces off a resistance level:
When the first high of the oscillator is above the overbought level and
the second is below, the divergence is rated 4.When both oscillator
highs are above the overbought level or they are both below the oversold
level, the divergence is rated 3.
- Divergence when price bounces off the support level:
When the first low of the oscillator is below the oversold level and the
second is above, then this divergence is rated 4.When both lows are
below the oversold level or the both lows are above the oversold level,
the divergence is rated 3.
We will have some examples of simple divergence.
Rule #5 states: price bounce-off can be confirmed by a divergence of either oscillator. This rule has a maximum rating of 4.
Types of divergence simple
You can split simple divergences into two: total divergence and lateral
divergence. In total divergence one signal (either price or oscillator)
is making a move in the opposite direction to the other signal. Lateral
divergence takes place when one signal is moving up or down while the
other has leveled out. Let’s run through the different types of simple
divergence that our system recognizes.
Divergence at the top of the market
Total divergence at the top of the market:
Price plots a new local high which is higher than the preceding high.
Meanwhile, the oscillator’s new local high is lower than its preceding
high.
USDCHF hourly chart Total divergence (Stochastic) at the top of the market:
Lateral divergence at the top of the market (type 1):
Price plots a new local high which is higher than the preceding one. Meanwhile, the oscillator plots a double top:
CHF (hourly) Lateral divergence (Stochastic) at the top of the market (type 1):
Lateral divergence at the top of the market (type 2):
Price plots a double top. Meanwhile, the oscillator’s new local maximum is lower than the preceding one.
USDJPY hourly chart Lateral divergence (Stochastic and RSI) at the top of the market (type 2):
Divergence at the bottom of the market
Total divergence at the bottom of the market:
Price plots a new local low which is lower than the preceding
minimum. Meanwhile, the oscillator’s new local minimum is higher than
the preceding low.
USDCHF hourly chart Total divergence (Stochastic and RSI) at the bottom of the market:
Lateral divergence at the bottom of the market (Type 1):
Price plots a new local minimum which is lower than the preceding one. Meanwhile, the oscillator plots a double bottom.
GBPUSD hourly chart Lateral divergence (Stochastic) at the bottom of the market (Type 1):
Lateral divergence at the bottom of the market (Type 2):
Price plots a double bottom. Meanwhile, the oscillator’s new local minimum is higher than its preceding min.
GBPUSD hourly chart Lateral divergence (RSI) at the bottom of the market (Type 1):
Rule #6: Double divergence:
Double divergence takes place when price, for example, plots three
consecutive higher highs while the oscillator plots three consecutive lower highs or vice-versa. The different types of double divergence are covered in here below.
Here are the ratings for the double divergence of either oscillator:
- Double divergence when price has bounced off the resistance
level:When the oscillator’s first high is above the overbought level the
divergence is rated 5.When the first high is below the overbought level
the divergence is rated 4.
- Double divergence when price rebounds from the support level:When
the oscillator’s first low is below the oversold level, the divergence
israted 5.When the first low is above the oversold level, the divergence
is rated 4.
Rule #6 states that price bounce-off can be confirmed by the double
divergence of either oscillator. This rule can give us a maximum rating
of5. Double divergence is a strong signal.
Rule #7: the trend
Although support and resistance levels are more significant than the
trend for intraday trading, we do not completely discount the
significance of trend.
As you know there are three types of trend:
- An uptrend
- A downtrend
- A flat – an overall sideways movement.
Our intraday system considers the trend for two different time periods: daily charts and hourly charts.
We identify the trend on the day chart using Kagi.
- A thin Kagi line moving downwards indicates a down trend.
- A thick Kagi line moving upwards indicates an up trend.
- In every other case our system considers the market flat.
Kagi settings for the trade system should be:
- CHF: 65 points (0.0065)
- EUR: 50 points (0.0050)
- GBP: 50 points (0.0050)
- JPY: 60 points (0.60)
The system uses the Price Channel indicator on hourly charts. Price
Channel should be set at 12 to identify the trend as follows:
- Uptrend signal: The lower channel line starts rising and continues
to rise or levels out and the upper channel line is not falling.
- Downtrend signal: The upper channel line starts falling and
continues to fall or levels out and the lower channel line is not
rising.
- In every other scenario the system does not recognize a trend.
Trends are rated as follows:
- When the daily and hourly candles agree on trend direction, a long
position opened on an up trend and a short position on a down trend are
rated 2.
- A position in the direction of the trend identified on the day chart when no trend is signaled on the hour chart is rated 1.
- A position that is opened when no trend is signaled on the day
chart(regardless of what is happening on the hour chart) is rated 1.
- A position against the direction of the trend signaled by both day and hour charts, is rated 0.
If you wish you can choose other methods of signaling a trend and still work within the structure of the overall system.
Rule #7 states: price bounce-off can be confirmed when price bounces
off from the level into the direction of the trend. Rule #7 gives us a
maximum rating of 2.
The next two sections cover some examples of trends.
Kagi signals on day charts
Take a look at the below chart; it’s a EUR Day Kagi chart. The chart
ends in a rising thick line. Our intraday trading system reads this as
an uptrend. The arrow indicates where the flat market passed into the
uptrend.
EURUSD daily chart Kagi – uptrend:
In the below chart ends in a falling thin line. This is our system’s signal for a down trend.
USDJPY daily chart Kagi – downtrend:
In the below chart ends with a rising thin line. Our system signals no trend.
GBPUSD daily chart Kagi – a flat market:
In the below chart ends in a falling thick line, once again a signal for no trend.
USDCHF daily chart Kagi – a flat market:
An uptrend on an hour chart:
On the below chart, up to line 1, our system signals a down trend as
neither the upper channel line nor the lower channel line are rising.
Although we can see that price has started rising well before line1, our
system only identifies an uptrend when neither the upper nor the lower
line is falling. So,our system identifies an up trend from line 1 to
line 2. From 2 to 3 the upper channel line is falling or horizontal and
the lower channel line is rising or horizontal: no trend. A down trend
is signaled as beginning once a candle closes on a falling upper channel
line.
EURUSD hourly chart and Price Channel (12):
Rule #8: The day of the week:
Not all days of the week are the same. The most problematic day for
effective trading is Friday. There are often very significant moves on
Friday(especially during the US session).
However, it is much more difficult to receive a signal to open a
position in time than on any other day of the trading week. Many trades
are being opened and closed with the specific reason that the weekend
is around the corner. Some ‘heavy hitters’ decide to close their
positions before the weekend break whilst others may well decide to open
in anticipation of a price shift after the break. These types of
activity are known as the weekend effect and they can have a significant
impact on price.
Monday is also affected by the weekend. On Monday many players see
how the land lies after the weekend and try to identify the direction of
price over the next few days and make plans for the forthcoming week.
This leaves us with Tuesday, Wednesday and Thursday as the best days to trade.
So a position opened on a Friday is rated 0; Monday rates 1 and Tuesday,Wednesday and Thursday rate 2.
Rule #8 states that the day you are opening your position must betaken into account. This rule gives us a maximum rating of 2.
Rule #9: The time of the day:
The best signal for opening a position comes immediately before price
moves in the desired direction. The greater the lag between the signal
and the move, the more likely that something will happen on the market
which will prevent that move. Take a look at the below table. This gives
a rating for each and every hour for the four main currency pairs. The time given is Greenwich Mean Time.
Different ratings are given for summer and winter because of the
Daylight saving adjustment that most industrialized countries employ to
manage energy use. If you are wondering when exactly winter and summer
start and end, we would advise you to follow the US system because the
majority of trading activity happens in the States.
These ratings are based on hourly price data over a year. We looked
at the range (the difference between the high and the low) for each main
currency pair at each hour of the day. The percentage for each hour is
the percentage of candles that had a range of at least 50 points for
that hour.
The hours with the higher percentage are given a higher rating
because a significant move is more likely to happen at these times.
Rule #10: The Unexpected
Rule #10 covers any
news
releases that are unexpected and which could influence exchange rates.
Our system acknowledges that often news is expected, such as an
announcement from the Fed and the result of this news is expected and
has already been factored into the price.Unfortunately, at other times
something happens completely out of the blue.
The most rigorous analysis is unable to account for the unexpected.
No one could have been able to factor in the 9/11 terror attacks but the
seat tacks had a significant impact on the dollar. This is an extreme
example but the fact remains that things happen which have a fundamental
impact on prices which are completely unexpected.
Rule #10 states that prices are affected by the unexpected. If you
know the unexpected is going to happen you give it a rating of 5 and if
you don’t know the rating is 0. Realistically, most of us are neither
insiders nor clairvoyants and rule #10 exists to remind us that nothing
is 100% certain.
Rule #11: the stop-loss:
Generally speaking, you need to use a stop-loss to cover yourself. It’s your first step in risk management.
The trade system has two beautifully simple rules concerning the level of your stop-loss:
- When you are opening a long position you need to set your stop loss
three points below the nearest local minimum. If you are working by the
system then this minimum will be the low of either the current candle,
or the preceding one, because the system is designed to open a position
just after the price bounces off a level.
- When you are opening a short position, put the stop-loss 3 points +
the spread above the nearest local maximum. Once again if you are using
the trade system correctly then this maximum will be the high of the
current or the preceding candle.
If these rules mean that your stop-loss will be more than 35 points
then it isa signal not to open the position. You should either wait for a
correction or simply by-pass this price move.
Rule #12: Closing your position
It’s the big question. When do you close?
There are a number of strategies to consider.
- Resetting your orders after each hour candle:When you are long you
need to focus on the low of the last candle and set your stop to close
at three points below that low.When you are short you need to focus on
the high of the last candle and set your stop 3 points + the spread
above that high.
- You could also adjust your stop to close in relation to price and
maintain a fixed distance between the two. This distance should be no
greater than40 points and no less than 20 and is known as a trailing
stop.
- You could also set your take profit order at a predetermined level.
This level could be a multiple of your original stop loss (eg. 3 x 35 =
105 pts) or it could be the next significant level that you have
identified on the weekly chart or through chart analysis.
- You could go for a combination of the above.
None of the above will be the best option 100% of the time and you must choose depending on the specific circumstances.
That is the final rule of the trade system. Although this run-through
may seem a little long-winded, we can lay the system out as a table.
Once we have the table we can quickly add the necessary ratings and
measure the success rate of each position.
Part 3
Using the trade system
The Ratings table
Take a look at the completed ratings table below. The table contains
the rules and ratings and makes the situation on any position clear. Our
example has been completed using some hypothetical examples to
illustrate some points.
Notice that the values in the maximum rating column and the maximum
possible column are different. This is because if you’ve got a simple
divergence, then you’ll definitely have a reversal of RSI and Stochastic
so you shouldn’t include the reversals into your total rating.
Likewise, a double divergence implies that you have a single divergence
and therefore the rating for the single divergence should not be added
to your total.
The maximum possible total is 25 and the probability of a position’s
success is expressed as a percentage. So if your position is rated 15
(as in example 2) then the success rate is
15/25 x 100 = 60%
It might make your work with the system more efficient if you remind yourself that a rating of 1 equals a success rate of 4%.
Rules and ratings:
Working with the ratings table: an example
Let’s run through example 1 in the table. This position rated a cool
76%which is only one point of the maximum score. This sort of level
isn’t reached too often. However, they do happen: take a look at the 4
charts in this section.
We are on CHF, it is October 24th and the 14.00 candle has just closed on the hourly chart.
- Level: Price has been making contact with resistance at 1.5072 for a
few hours now. This resistance level was set with Price Channel (4) on
the six-hourly chart which makes it a daily resistance level(see
section2.1.1/2). Price opens above the level and then rises before
dropping below the level and closing at 1.5067, five points below our
resistance. This gives us a rating of 5 points.
- Candle: The hourly candle for 24th Oct 2002 14.00 GMT is an inverted black ice, giving us a rating of 4 points.
- The candle also completes a double divergence with RSI, that’s another 5 points.
- The Kagi chart signals no trend giving us another point.
- It’s a Thursday. That’s another 2 points.
- The signal to open was at 14.00 hours, 2 points more.
Day level set using Price channel (4) on the six-hourly chart:
Price bounces off level as an inverted black ice:
Double divergence with RSI:
Kagi signals no trend:
That’s 19 marks in total. Divide 19 by 25 and multiply the result by 100%. This gives us a success rate of 76%.
You can work out a success rate using the same method for any
position you plan to open. The two other examples shown in our table
have a rating of 60% and 52%. We advise only opening a position on
trades with a rating of 60% or more. One trader, who has used this
system for the last three years now, only trades 60+ signals. This
strategy has bought him an average 600% yearly gain.
An 80% success rating is a rare bird. One thing that often happens is
that the day trend and the hour trend do not coincide. The hour chart
usually needs a little more time to register the trend which will
develop on the chart a couple of candles later.
Working with the ratings table: a general approach
The general approach to rating a position goes like this:
- Plot the levels on the chart and decide if price is bouncing off or
breaking out of these levels. If there is no bounce-off, move on to the
next currency pair. If there is bounce off, record it on the ratings
table.
- Then rate the reversal candle and enter that rating into the table.
- Check for double divergence. If it’s there, perfect. We can record
it on tothe table and not even think about simple divergence or
oscillator reversal.
- If there is no double divergence, we check for simple divergence.
- If there is no simple divergence, we check for oscillator reversal.
- Identify trend direction and record the appropriate rating.
- Rate the day.
- Rate the time.
- Total the ratings and calculate the success rate. We open a position when the rate is 60+%.
- Decide on the criteria for closing according to the current market. Open the position and place your orders.
And that’s that.
Ratings table:
You might want to modify the system and we’d like to wind this section up with a few words on possible modifications.
You may prefer to shift the focus of your analysis from hourly to 30
minute candles. That’s reasonable, but we would not recommend using time
frames of less than 30 minutes.
You could also change the criteria for identifying levels. You may
also decide to use lines of support and resistance as well as levels.
Please remember that this system is tried and tested. It has proved
itself time and time again. If you follow the mantra “If it’s not broke,
don’t fix it”you’ll find that the system will pay back the time and
money that you have invested in it.
