15 July 2010: The bulls roared
And I missed it. Need to consider a hedge trade to open in the reverse direction after the stop is hit.
14 July 2010: Scalpers Day
Choppy trade – Scalpers session at best.
13 July 2010: Don’t get the bloody trend wrong !!!
Not so easy to detect the trend particularly when it changes. Thinking of slowing down my time frame and making use of a better strength indicator.
12 July 2010: On Mondays Trade The Asian Session
Consistently on Monday the Asian session has provided reasonable trends. London and New York have a reputation for being unpredictable and today was no exception. For some reason the Asian session seems OK. At least for the last few months that I have been watching it has more or less offered reasonable trade. Perhaps the best idea is to approach the Asian session with a view to trend and/or approach London with a view to scalp.
12 July 2010: Should have traded the session boundary.
EURUSD: Monday Trade down 700 points during the Asian session. 15 Minute stochastic showing a bounce off oversold even though overall trend was down. 2 Minute stochastic hovering near overbought. British GDP results expected an hour and a half after open. Confusion on whether to expect a retracement using the 15 minute as a guide, further falls using the 2 minute or a sideways slide during the London session going into New York while everything cools off and people digest the GDP results. Ended up being a bit of everything.
In Hindsight should have starting with a short and then scalped. The 2 minute stochastic flipped into overbought exactly on London open just when the price ran into the 200 tick MA.
Source of confusion? The 15 minute stochastic. The trading plan calls for me to trade in the direction suggested by the 15 minute stochastic. Also didn’t consider the brick wall offered by the 200 tick MA.
July 12 2010 | Trading | No Comments »
I wanted an indicator that provides a view of the strength of the market. In particular I wanted it to capture the spikes that occur prior to trend direction changes.
The basic idea is an RSI accept instead of capturing the relative size of up and down moves it is calculated on how strong the high (push up) and low (push down) are from open.
The other difference is two sets of horizontal lines set at 30, 40, 50, 60 and 70. The inner two at 40 and 60 provide a channel between which the indicator moves when range bound. When the Price is bullish it moves between 50 and 70 and when Bearish 30 and 50.
Overbought and oversold when it hits a line. If the range bound then that means 40 and 60. When trending choose either the 50 and 70 pair or 30 and 50 pair.
I have also added a 4 point EMA as a signal line.
It works pretty well for my purposes. It tracks oversold and overbought signals generated by the stochastic, but has additional information regarding trend strength.
UpTick=High-Open
DownTick=Open-Low
//N a 12 point EMA for data collection.
Av1=ExponentialAverage[N](UpTick)
Av2=ExponentialAverage[N](DownTick)
//M as 9 point EMA for smoothing
Strength=ExponentialAverage[M](100*Av1/(Av1+Av2))
return Strength as “Strength”,30 as “Lower”, 40 as “Lower Channel”,50 as “Mid”,60 as “Upper Channel”,70 as “Upper”
July 16 2010 | Trading and Uncategorized | No Comments »
The new plan is in essence based upon the old plan in that it attempts to capitalize upon the strength of the London session. Where it differs is that it attempts to create more trading opportunities. So in effect it moves the trading sessions into the background as context. In particular the session boundaries are places to expect significant trend changes. The focus is on entry, exit signals and trade direction.
Background Problem.
Having tried a few different ideas, I decided to trade a stochastic overbought and oversold strategy. It seemed to make sense. Nearly every move up or down was preceded by the stochastic being oversold or overbought. In practice I found myself stopping out all the time. So how could I have consistently been on the wrong side of the trade? What would have happened if I had made the exact opposite trade? What could I use to suggest a different trading direction?
The answer turns out to flow from another error with the previous plan. I mentioned in the update below that if you change time scales you will see that an entry signal on one time scale may well be an exit signal on another.
So how can that make any sense. Does it mean that technical indicators are simply a steaming pile of bullshit that provides a perfect 20/20 in hindsight view of the market?
What I was doing was trading on the 2 or 5 minute time scale, ignoring the context of the 15 minute time scale. I was for example entering long on a stochastic oversold that would in fact go up a little, but then turn around and come down further.
I also was holding past the end of the cycle. When the stochastic had moved from oversold to overbought I was holding through the next half of the cycle from overbought down to oversold “hoping” it would just be a retracement. In fact what I was trading was the retracement against a background opposite trend….
Solution: Entry and Exit.
To beat this problem use two charts. One on fifteen minutes and one on five or two minute.
The fifteen minute chart provides context. If the market on a fifteen minute time scale is moving from stochastic oversold to overbought then it is reasonable to expect the market to continue in that direction. Care as usual on major support and resistance levels indicated by moving averages, pivot levels etc. Care also on major market “mood” changes that occur on session boundaries.
The two or five minute chart provides your entry and exit signals. Trade in the direction suggested by the fifteen using the shorter time frame chart to signal your entry and exit. When the fifteen is overbought or oversold use the hourly to provide a bit of context as to whether it will keep trending or reverse.
It is important not to hold the trade past the half of the cycle of the short term time frame you are trading on. It might simply stall and continue if the trend is strong. It most likely will retrace at least a little which will provide you with a better entry. In any case you must make a decision to exit and reenter or hold it through that point. You can’t just ignore it and hope.
So far this is working. My success rate has turned around dramatically. But early days yet. We will see in a month how it pans out.
Indicators: I use a slower than normal stochastic set on 12 and 6 rather than 14 and 3. It is a a lot smoother but otherwise reliable. For context on the chart I use Pivot levels and 25, 50 , 100 and 200 moving averages to let me know if there are resistance/support levels the price may bounce or slow down on. I also have the background of the chart coloured differently for each of the Tokyo, Hong Kong, London, New York sessions and the period when all markets closed. In fact I change session colours about 15 minutes before the actual start as a heads up that a big change may be imminent.
Stop Loss: 10 pips when using a two minute chart. Limit initially at 20 pips and then move it to around a likely resistance level as required.
July 10 2010 | Trading | No Comments »
A few weeks into the new trading plan and things have changed to a point that an update and a new plan is needed.
First things first though a summary of the last month. The essence of the old plan was to capitalize upon changes in market strength that occur on session boundaries. Particularly London open.
Problems and Solutions
1) Too few entry opportunities.
This was a big one. The way the old plan was structured I had to get the entry at the start of the session. If I missed it, it was 24 hours before I could try again.
Solution: Change the Plan.
2) Scared to pull the trigger.
After months of sitting and watching I found myself too indecisive to enter the trade. The reason was that I had a lack of confidence in the entry signals. It could go one way, but then again could go another.
Solution: Spent a couple of weeks simply marking entries on the chart as they rolled across the screen. Then when I had that working moved to a demo account and tried it with orders and finally to a live chart.
3) Lack of a good entry strategy.
When I found myself on a live chart I found myself changing strategies mid stream. A sort of chaotic mix of Heiken Ashi trend change signals, Stochastic signals and varying time scales. In effect I was randomly clicking and hoping. That by the way is not trading, it is gambling.
Solution: Only trade using one strategy. Only trade on one time scale. If you change time scales you will find that what was an entry on one may well be an exit on another.
4) Staying alert: Not a problem if you are in the Northern Hemisphere, but from Oz the London session starts at 5:00 pm.
Solution: Roll my sleeping hours around – and lots of coffee and energy drinks.
July 10 2010 | Trading | No Comments »
Well after a few months of simply sitting and watching the Forex markets I think I am finally in a position to put together a trading plan. It is based upon the observation that the Forex markets have several sessions throughout the day that correspond to different groups of traders entering the market. Within a particular session the market direction tends to be constant. Note: there are obvious variances to this particularly during the New York session if major economic announcements are released. But I am asleep New York time, so these are less of a concern.
Specifically we have the Asian Session that itself breaks into the Sydney Session and the Hong Kong Session. These two are followed by the London Session and the New York Session and finally a brief period when no market is really within a business day.
During each session we have a different group of traders in the market who more or less are in the same frame of mind depending on what was in their morning Newspaper and what the charts are saying when they hit their desks. So they are either in a Bullish/Short Covering, Bearish/Profit Taking or Scalping/Neutral mood. This frame of mind tends to persist throughout the session. At the end of a session there may be some profit taking.
This is mainly written with a view of the EURUSD, GBPUSD, AUDUSD pairs, but should apply for most of the majors.
So how to trade this.
First go and find out the opening times of each session and come up with an indicator or alert that you can throw up on your chart so you know what is open and in particular where the session boundary is. (Don’t forget to convert this to the time zone you are trading in – and remember to include daylight saving changes…)
The session times correspond to normal exchange hours plus or minus a bit. In local times I have Sydney 9:00 AM, Hong Kong 10:00 AM, London 8:00 AM, New York 9:30 AM, Markets Closed 5:00 PM.
Remember that when a new session starts even if the previous session is still open you have a new group of traders about to start pumping orders into the market. So if the market is going to change direction, this is when it is most likely to occur.
Use a five to fifteen minute chart and wait a few candles for the market direction to become obvious, but don’t wait too long. Watch out for bounces off pivots or other major support lines. Once you have worked out the underlying pressure, enter in that direction with a reasonable stop loss. For a stop loss I am looking at 15 to 30 pips depending on volatility on the EURUSD pair. The AUDUSD and GBPUSD pair can be more volatile so may require a larger stop.
Keep an eye on it because scalping/choppy days look the same when they start as long and short days do. If the session ends up choppy either scalp the channel it is in – most likely between a pair of pivots or simply take the hit and walk away for the session if you are uncomfortable with scalping.
Use a five tick per candle or some suitably fine resolution to set your entry.
To exit that is up to you. Either close before the session ends or take the risk and carry it though exiting if the trend changes. You could take a fixed number of points or even use the direction to define a direction to scalp throughout that session if you are nimble.
A few observations to keep in mind
Sydney and Hong Kong are more likely to be neutral than London and New York. But there are enough opportunities here even so. London is normally the biggy. If London is flat New York probably won’t be. Hong Kong reasonably often will pull back to a pivot around 5 or 6 am London time prior to London Open so watch out for this. London Open can also be sloppy in the sense that a firm direction may not resolve until 10:00 am London time. It usually does – but don’t walk away from your screen assuming it will.
Remember that when the next session opens it is a new game. Everything resets to zero so you have to adapt your mindset.
I will let you know in a month or so if it works. Been watching so long I have to get back into a trigger pulling mind frame.
Good luck and happy trading 
June 03 2010 | Trading | No Comments »
After years of dabbling with the usual assortment of penny dreadful miners from Australia, occasionally scoring a win and often a loss I decided late last year to make a serious attempt at trading for a living. It has been an expensive hobby ever since the tech wreck, but being a maths grad the never ending stochastic sequence they call the market has almost limitless appeal. Anyway just looked that calendar and realized that three months in and I still haven’t made any more than a couple of trades – and they were all of the small dabble your toe in the water kind.
It is interesting and possibly instructive to think back over the process to date to understand just how hard this actually is. Perhaps my marathon will help someone else cut some corners.
Which Market?
I decide to trade SPI200 (Australian 200 Cash Index). Why: – dollars per point risk and reward appeals. Leverage is huge at 100 to 1 which means you don’t need to tie up any more capital than you are prepared to risk per trade. Having a 24 hour market means no nasty overnight gaps when Wall Street sneezes. And finally at a 1 point spread, costs are low.
Problem: I am not nocturnal.
After weeks of following my penny dreadful habits of getting up early, getting my head across the US market, looking at the commodities news etc for a couple of hours and then settling in for market open at 10:00 am, it became apparent that the bloody Australian Index doesn’t do a damn thing during Australian trading hours. It simply trades sideways – occasionally kicked along after lunch if the Hang Seng is having a good day. But generally all the action is after midnight my time when the US trades. A couple of weeks wasted here working this little home truth out.
China Maybe?
OK lets try the Hang Seng and bypass the middle man. Starts at lunch time by my clock and trades fairly smoothly. Mmmm better but the spread is much bigger and then the Chinese government chose that particular time to piss on the parade. This pretty well took the exuberance out of the party and the Hang Seng resorted to making its moves US time. Once again we are back with a flat market when I am awake. Another couple of weeks wasted.
Lets trade index futures nights then. Sounds good, but apart from the insurmountable problem of rotating my body clock to wake up at 10:00 pm, the spread goes up making each trade much more expensive.
How about increasing the stop loss distance and reducing by dollars per point at risk. Same at risk amount, but lower probability of being hit. That would allow a buffer for overnight adverse swings. Problem – my risk is lower, but now so is my reward based on the normal trading range. Putting these two together and we lose anther couple of weeks.
How About Currencies?
Aha! How about currencies and not Indexes. At least the spread costs are more consistent over a 24 hour period. And lo and behold! they have most of their trading activity UK hours rather than US hours. At least this means there are a couple of trading hours when the sun is more or less above the horizon.
Cool. That’s settled, we will trade Forex and to keep it simple make it the EURUSD pair. High liquidity, not too volatile and responds well to London open.
Blow fly in a bottle.
Problem: Have you ever seen a forex chart? I have been charting for years and thought I knew a Stochastic from a Standard Deviation. But these damned things are all over the place. The more indicators I have on the screen the worse it gets – arrrrgh!. Less is definitely more when it comes to these monster markets.
The realization hits that forex traders are all trading off different time scales. Now with good old daily charts everyone is getting more or less the same signal at the same time. Not so Forex. Not only are people’s days starting at different times, but they are all using different periods. 1 hour, 15 minutes, 5 minutes, single ticks. This means that when one person is getting a buy, someone else is seeing a sell. I have absolutely no hope at all of predicting a buy or sell. It is just absolute chaos.
Solution: Definitely use London Time for forex and get your brain around the fact that the London stock market starts at 8:00 am their time. So they are at their desks before that. Find indicators that are independent of time scale. In other words use Pivot Points. Throw a set of pivot points up on the screen and all of a sudden that mad blow fly in a bottle motion settles down to a more or less orderly bounce between predictable resistance and support levels. A couple of more weeks wasted working this out, but finally I can now look over the days data and see a buy or sell occurring where you would expect them to.
Chaos almost cracks.
Armed with this, methinks I have it cracked. Wait for the EURUSD to pull back to a support level and enter. So we wait, and wait and wait, nature calls, we get back and have missed it. Day after day the same thing. I either miss the signal or to make things worse, it bounces off what appears to be a hidden resistance level. Later I later work that out to be the 200, or 288 tick moving average. It varies methinks. Computers or at least their programmers would think about the average price over the last 24 hours – hence 288 on a 15 minute charts, but humans use a 200 no matter what chart because that is what they have been taught.
After a couple of weeks of this the realization sinks in that a one or two trading signal per day strategy simply will not work. A fifteen minute time scale also doesn’t help. It is amazing how much a currency can move in fifteen minutes. More time wasted.
Mmmm what does tick by tick data look like? Oh shit! there is a an awful lot of it, that locked up the lap top. After a restart we discover that a 5 tick per candle time scale works and doesn’t blow up my hardware. My stop loss distance is relatively large with respect to the per tick motion. It moves fast enough to maintain my focus, but slow enough to react. Feeling good here.
What Signal?
Problem. What to use for an entry signal? A stochastic, MACD or RSI on 5 tick data is just impossible to read. The Pivots are still good, but most of the time it isn’t on them and there are a lot of opportunities between support levels.
Solution: Good old resistance and support levels with a bit of help from a slow stochastic. I mean really slow. You know the normal 14, 5, 3 stochastic? Well mine is sitting on 200,100,50. My MACD is 400,200 and 50 instead of the usual 26,12,3. But hey it works. The really slow stochastic – remember I am on a 5 tick per bar time scale which is really fast – provides a good picture of the underlying trend. That means I can finally read this. Add in a 400 bar moving average along with the pivot points and at last after months I can finally call entry and exits as they scroll across the screen more or less consistently. A few false signals and a couple of wiggles beyond my stop loss, but that is par for the course. Lots more entry opportunities a day
The current problem.
Well at least its not a problem making sense of the market. The main problem is the original one. The optimal time to trade is the London session. For the same risk, the market moves a LOT more. Reward to Risk is about 3 to 1 during the Asian session and can be about 10 to 1 at London open. The key issue is being alert late in the afternoon. Not as easy as it sounds. And worse if you have been at the screen all day.
The second problem is choice of markets. Focusing on the EURUSD sliding sideways while the AUDUSD pair is flying is not good trading.
Not so easy.
The lesson here is that if you think you are going to get rich quick, have a rethink. If you can find a mentor to cut corners you may get a leg up. But my gut feel is that it will take months of hard and consistent work with little or no money coming in while you get yourself around the learning curve to the point that you are trading and not gambling. And that assumes you have a good understanding of charts, you understand margin, risk management and have good money management skills before you start.
You could be throwing money at the market while your “L” plates are on, but remember that this game has a 90% failure rate. Most of which I would think happens during the first few months.
You have to be in it to win it, but that doesn’t mean you have to act like you are at a casino. This game is a business and like any business it takes a lot of personal investment.
May 20 2010 | Trading | No Comments »
Rather than trying to delve my way through the myriad of files on the US Commodity Futures Trading Commission website every week to try to pick up a picture from the weekly Commitment of Traders reports I figured I may as well put together a simple database and data display page.
COT Report Generator.
It provides a one year view of the Open Interest and Non Commercial Long and Short futures positions across the major currencies, commodities and US stock indices. I have left commercial positions out of the graph as these usually are hedges against unexpected market reversals – and hence are long when the market is short and vice versa. Open interest does include the Commercial data. It does not include everything the CFTC tracks – which is about everything from feeder cattle and corn to currencies, energy and indices.
In theory it will update once a week early Saturday morning. We will see whether that works as intended.
Fairly easy to use, select a contract from the list and hit the button
May 18 2010 | Trading | No Comments »
At the risk of sounding like a ten year old, “the absolute bestest” news summary for forex news I have found is the market updates page from Forex.com. As each session rolls over it it gives a good concise recap of what is just happened. Just what the doctor ordered to start the day.
May 06 2010 | Trading | No Comments »
Had another day of watching the EURUSD wiggle around the 200 tick MA on a five minute chart and managed to convince myself that it would probably move up to the central pivot before moving down. Of course it did nothing of the kind – it simply dropped a solid 170 pips over the next few hours leaving me pissed off on the sidelines…. (late note: and then continued on to make it a whole 250 pip move – damn!)
After sulk session of growling at the screen and berating myself while trying to figure out my screw up, I have finally convinced myself that I have discovered what appears to be “the” rule on moving averages for 24 hours markets. In fact I have touched on it in an earlier post, but today nailed it.
The rule in words is to find the average price over the last 24 hours. As simple as that. If today is better than yesterday it will be above it and if worse below.
The important thing is that on the EURUSD this line swamps the central pivot point. Well at least it has over recent weeks. If your central pivot is above it, the market is bearish, forget any possible thought of it retracing through the days average unless the trend is changing. This thing is like a brick wall.
The trick is your time scale. On a 5 minute chart it is a 288 period simple moving average. On a 15 minute chart it is a 96 point moving average. NOT 200 and 90 period moving averages – they are close, but no Kewpie doll.
Work out how many of your ticks there are in your favorite time scale over a 24 hour period and set the moving average to that. As an example there are 12 individual 5 minute ticks an hour so a day will have 12 x 24 = 288 of them. At 15 minute ticks, there are 4 sets of 15 minute ticks an hour, so a day will have 24 x 4 = 96 of them.
You probably use 20/90/200 MA’s don’t you? Well rethink it and get anything off your screen that is not rock solid. It will just confuse you. Just remember that even random lines will show some apparent evidence of support and resistance at times.
(Late edit: put your 200 back – perhaps as a thin line, now I have taken it off, I am finding that sometimes it holds. The hypothesis is that when automatic programming that uses the last 24 hours average price dominates then the 288 on five minute dominates, but when humans are the main players, the 200 dominates.)
Easy huh. Just weird numbers is all.
P.S.
The reason 20 day, 90 day and 200 day MA’s work on daily charts is because the correspond to a month a quarter and a years trading days. And yes 200 ticks work on weekly and monthly – but that is because people use them rather than people asking the “Is today better than the average over the last month/quarter/year etc” question.
EURUSD May 5th 2010, 5 minute interval with the 288 tick MA in blue.
May 05 2010 | Trading | No Comments »
It took a bit but I think I have made my mind up about the Goldman Sachs fraud allegation. Essentially they were simply doing business and the fault actually lies with the people they were doing business with.
As a company they had a responsibility to reduce their risk exposure. They looked at the market, worked out the risk reward ratio of the CDO’s and determined that they would be better off without them than with them. Fair enough.
The counter parties who bought the CDO’s thought they were getting easy money and that Goldman were the losers who sold a good revenue stream.
The problem was these banks who bought the CDO’s didn’t do the most basic of research. The boom in the real estate market was bought about by the Baby Boomers and their offspring being in the market at the same time. When the after boomers had bought their first home the pressure came off and you ended up in a situation where the assumption that real estate would go up forever was no longer true and the sub prime mortgages were no longer covered by a gain in the value of the underlying asset that would have occurred should that demand have persisted.
In other words the buyers of the CDO’s needed to do some basic research about who was in the real estate market, what they were doing and how long would the market last. It would have been fairly simple to bring up a population by age curve and ask a couple of questions about what people in such and such and age group are likely to be doing. In other words marketing 101.
The really scary thing given the millions of dollars these Bank CEO’s learn is that they knew about the demographic distribution. They were in fact targeting the Baby Boomers with reverse mortgages. All the “Unlock The Equity In Your Home” products that abounded. At the same time they were targeting the surge of “First Homer Buyers”. They weren’t making a lot of effort to targeting the mid mortgage people.
The way it works is that any organization has a finite marketing and product development budget. Companies target products and marketing campaigns towards where they can get the most bang for the buck. They targeted these two markets – the “soon to be retired” and the “just married looking for a home to start a family” groups because that is where the population was.
So if we make the assumption that the CEO’s knew that the demographics of the market was not flat across all age groups, then we have a situation where they knew that demand was driven by a surge that would die away once all the first home buyers had filled their immediate demand.
Why then did these multimillion dollar CEO’s make the assumption that demand would grow at a constant rate for ever? Can’t they read a basic graph? Couldn’t they look at their own children to see what they were doing? I definitely remember all the “Home improvement shows” and talk about the difficulty of searching for a home that in time was followed by lots of pregnant bellies around the office, that in turn was followed by lots of prams and doe eyed mothers proudly coming back to work to display their latest creation.
These new mothers were not buying new houses – they had just bought one and were now out of the market. Even if they did have a desire to move to a larger property it requires selling the one they are in. No net gain in demand.
The question is why were Goldman able to work out that the risk in the mortgages was too high and why were the other bankers unable to? Did Goldman simply do better research?
The question of why anybody other than what you would normally call “sophisticated investors” were buying these things is another thing altogether.
May 03 2010 | Market Comment | No Comments »
Next »
