I experiment with crazy ideas sometimes because I almost always learn something. So here's the latest . . .
A person using a tick chart decides they want the data plotted differently but still retains the overall tick parameter of drawing something new every X ticks. So they decide a box will be drawn every 100 ticks and inside the box they want hash marks instead of bars. And the has marks are drawn once every 5 seconds. So whatever the Ask price is at each 5 second interval plots a has mark inside the box. Once the 100 tick limit has been reached a new box is formed.
It would look something like the image below. I left the original bar-type plot next to the boxes so we could compare the two styles. So the idea is to take a price sample every 5 seconds rather than to deal with price on a tick by tick basis. How you trade it? I don't know.
They just look like equivolume bars because the width is different on the 5 minute bars - similar concept but in this case it is because a different amount of ticks are in each 5 minutes. This is Using Ninja 6.5
Just trying to show that without time stamp data or even with different providers the "bars" can look different based on how you are dividing up the data.
Time Time-based charts are the only non-random charts from my perspective. I say that because time is universal and cannot be changed. From the smallest trader to the largest trader using a Cray super-computer, 5 minutes is always 5 minutes.
I had hoped that I was showing exactly the opposite -(Yes a 5 minute block is a 5 minute block) but even a few seconds in difference as to start time can dramatically change the way 5 minute bars look over identical tick data. If you are trading off of this look be careful. To assume everyone/broker/platform/exchange has a clock synced to the second is too much to ask. This may or may not matter to your trading style but you should be aware of it. When you put great stock in certain bar patterns vs following the general flow or wave of price action you can get in trouble. Of course if you put stock in hourly open close prices you have the same problem but it starts to fade in significance. I would say most clocks are within 5 minutes of each other and maybe even within an under 1 minute tolerance of each other. This is possibly why pivot points tend to be accurate as to areas rather than to the tick. (yes - to the tick sometimes occurs)
From above document -Indicators of
relatively short length are fast enough to give signals, but don't consider enough
data to even attempt to find a predictive pattern. Indicators of long length
give excellent summaries of the action, but are too slow to give entry signals.
Further, your brain is in the same predicament. Watch for simple patterns
(candlesticks, ross hooks, double-bottoms, etc.) on your chart and you will see
them everywhere. History shows you will not win by trading all of the localized
opportunities. Wait for too much confirmation and you will enter at the end of
trends.... Changing charts and time-horizons is not the
answer. Non-localized information is.
5 minute candles are an indicator - use them accordingly.
Essentially this is a big video game between you and your broker. Your broker supplies an random number sequence with a small component of trend (enough to keep you interested) and you get to build a video game around this in an attempt to capture small pieces of the feed where you were right in guessing the correct direction. You can use any tools you can come up with....colored lines, statistics, bars of various shapes, etc. Only look at the game every 5 minutes, 1 hour, etc Everyone has an opinion of the "best" video game skin or wants to sell you their version but ultimately the "game" you design will work best for you. IMHO. Good gaming.
Last edited by websouth; January 11th, 2010 at 10:18 AM.
The following user says Thank You to websouth for this post:
read your post with interest regarding different data set etc,
20 year ago, no tech, no opps,
today much tech much opps,
the problem is that people have to many choices,
stick to the basics price and volume, support and resistance, the market is about prices and volume, not indicators type of computer or datafeed, try pencil and paper for a while, noting the volume and price every five mins, you will get a real feel for the mktmaybe intraday between 9.30 and 11 and you may make money
20 years ago, no small retailers like you and I; ONLY Fundamentalists and Big players(e. g. banks, etc).
today much tech much opps
today many retailers, many Big players and many market Manipulators.
... noting the volume and price every five mins ...
... noting the volume and price every five mins still USING COMPUTER, and DATAFEED.
Tomorrow, the market scenario will change (just like it has changed from 20 years ago). So, the ONLY choice is to USE available TECHNOLOGY as best as you can efficiently and profitably do (NOT many choices, really).
I think its worth mentioning what charts actually are, not just a graph of prices but a chart is showing market participants feelings about price. A price chart is a chart of emotions and pschcology. Also give that human nature repeates itself the same things will show up regardless of timeframe or chart style.
The following user says Thank You to rassi for this post:
When I've attempted to explain trading to my friends I'll use the analogy of a psychiatrist trying to treat a patient with multiple personalities. Step 1 is to determine which personality you're dealing with at this particular moment.
So yesterday I was dealing with Larry, the meandering patient with no self confidence. I can tweak my charts to deal with Larry very well. But today, I've got to treat Susan who is conniving and deceptive and must use a different set of charts altogether to work with her. And that's the underlying problem for us traders, no single chart will work with all personalities.
For argument's sake, let's assume you've identified 7 distinct personalities and have the ability to spot each one as it appears. Here's the options:
1] Create 7 different chart setups and pull up the correct chart as the occasion dictates. This method allows you to trade all the time.
2] Decide you will only trade 1 of the personalities and optimize 1 chart for that distinct personality. Then you sit on the sidelines and wait for that personality to appear.
3] Learn to trade without a chart.
The following user says Thank You to hondo69 for this post:
While we are on analogies I have always like the idea of the chart as a map (with the continuous flow of price being the terrain). The map is not the terrain.
If you want to go 5 blocks you would probably pick a different scale map than if you wanted to travel the length of the country. Pick an appropriate map for the journey you want to make. If you start off not knowing where you want to go and with the wrong map chances are pretty good you will get lost.
Markets do have 'personalties' but they are not as many and varied as some would have you believe. Picking a robust 'model' for market behaviour will allow you to know at any time what you are dealing with in a quantifiable fashion. A couple that spring to mind are Wyckoff (supply and demand as manifested by accumulation, mark-up, distribution and mark-down ) or Market profile (moving from balance to RE as a result of initiative buying/selling and rebalancing again at new value). Or dead simple, trending or non-trending. Lots of ways to quantify that.
I use range bars because they have proven to work best in my money manegment and volume stop indicators but also incorporate the 5 minute chart for S/R. Learning to incorporate them together has helped me to read the direction of the price and volume with minimal indicators.