It all depends on how you use them (Range bars or any other type).
I agree that with Range charts it is very important to keep track of the time elapsed between bars. That makes them somewhat hard to use for discretionary trading, so time based charts are probably preferable in that scenario.
From a statistical standpoint and from an algorithmic trading standpoint, range bars are superior in a number of ways. They reduce the variance between bars which greatly simplifies some statistical analysis methods. Also, since more bars will be generated during a volatile event the program has more chances to analyze and potentially react at FirstTickOfBar, this is necessary for many techniques which are unstable when used with COBC = false.
FWIW I know of a number of institutional traders who use range bars or similar concepts. One example is Cynthia Kase, who used to be an institutional oil trader, in her 1996 book she even goes so far as to call them 'Kase Bars' since they were not available in any retail trading platform at that time.
Range bars do have issues however, the primary one being dependence on initial conditions. In other words, if you open up two range charts that start at different times, the charts will be very similar but not exactly the same, ie they may be 'out of phase' with each other. In order to account for this in an algorithmic strategy it is necessary to make use of conditions that are more general, or to look back in the recent bar history for some signal and then try to confirm it. If you do not do this then you can have the same strategy running on two machines with very different results due to minor differences in bar placement.
Why can my chart look different after reloading historical data from the server? As ticks come into NinjaTrader in real-time, they are time stamped based on your local PC time if they do not already have an associated time stamp that is provided from the real-time data source. The majority of our supported brokerage feeds DO NOT time stamp ticks where most of our supported market data vendor feeds do provide time stamped ticks. NinjaTrader then builds bars based on the time stamp of the incoming tick and displays these bars in your chart in real-time.
I'm really glad we're getting some good discussion here, it is helping me a great deal.
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.
As Websouth noted, this is dependent upon our computer clocks. Some computers are in sync with "atomic time" and should therefore have matching charts. It is true that various data feeds will cause a slight discrepancy, but as a general rule the charts should pretty much match. Any computer not in sync with "atomic time" will show different charts.
I feel all traders should use the free utility called SymmTime, which can be downloaded here. It syncs your computer to the atomic time clock.
Charts using volume bars are random. They start counting when you first turn on the chart. And since we all turn our charts on at different times they will show different bars. The same is true of tick charts.
I use Ninja and trade with a buddy using TradeStation. No matter how we try our chats never match exactly. If we both use a 5 min chart and add a SMA with a period of 50 (based on close), they are quite different. Now the SMA indicator is about as simple as indicators come so one would think the two charts would match (more or less). But the same is true whether we use SMA, pivots or what have you. The charts simply don't match and never will. This is unacceptable to me.
We pay a lot of money for our platforms and data feeds. We deserve better.
I believe Zoethecus is correct by saying we all are basically looking for specific patterns within the chaos of the market. Take the example of a trader that uses a MACD crossover as his entry signal. That signal is based upon the type of chart he has selected to use. But if I try to copy his method exactly we will enter at different points due to the fact that are charts will invariably be different. So now we're back to the randomness of the charts.
Now some will see this as nit picking to the nth degree. I do not. For the amount of money we pay to be in this business we should have accurate data that we can trust. Instead, we are provided sloppy data that merely gets us in the ballpark.
Some traders take the long view and plan to enter and hold their position for several hours. These traders have the advantage of negating the randomness of charts. They really don't care if their charts are a little "fuzzy" because a few ticks here or there do not affect their trading style. Heck, they don't even worry about 10-15 ticks for that matter.
Other traders try to pick up a couple of ticks on a mini-scalp trade. These traders are greatly affected by the randomness of charts. They are basically shooting in the dark.
An understandable point of view, but please allow me to play devil's Advocate for a moment. I will be playing Devil's Advocate, because I myself also trade a 5 min chart. Not because I believe anything about what other traders use, but simply because after much experimentation and losses, I find that that is the chart that I can make the most sense out of; and the chart that gives me targets and stop losses that fit my account size, trading style, patience limits and risk tolerance.
That having been said, I also watch a 1 point (4 tick on the /es), Range Chart, even if it is not the basis of my primary signal.
The reason is quite simple. IMHO, nothing shows price action better than a range chart. The underlying moves 1 point; a bar is complete, and that is it. It does not matter how long it takes; how much volume came in; how many PPT orders went out, yada, yada, yada. It simply says: "The price has increased 1 point, since the last time that I took a reading that completed a bar."
A rapid series of small moves show up as a rapid trace of alternate red and green bars, painting so fast that they freeze any decision, thus keeping me out of rapid chop aka false breakouts and fakeouts.
By the same token, in slow chop, one can clearly see a horizontal set of green and red bars, clearly delineating chop by showing that the price is going nowhere. Sometimes this allows me to enter into 3-tick scalper mode, fading each bar until I lose a trade.
A single bar that has not completed for a really long time, tells me that volatility is so dead, that my trading style cannot make any money, and it is time to take my ball, go home, and come back another day.
So for me, by removing the time element in range bars, I can see even more clearly if the market breaks out or is chopping. For me range bars are a good adjunct; their smoothness attribute really does not mean that much. I just find them useful as a filter of pure price movement.
The time chart is of course much more useful in telling me when I have worn out my welcome in a larger trade. For some reason, I just find it easier to let the market flow into my trade parameters on a 5 min chart.
It's a data sampling issue. Charts should not be random. All charts with the same parameters should look the same if you feed them the same data. With constant range. volume or tick you have to start plotting from the same tick... be it the first one after midnight, first one of the session, first one of the contract or the 26th one after Feb 29th. Also not all charting packages plot bars consistently. For example a tick of 120 contracts with 100 contract constant volume bar chart and TS always used to throw the 20 extra away rather than adding it to the next bar. Maybe its fixed now, certainly should be. In short absolutely agree about your observations under the section Platform. Hound them to sort stuff out.
Again ask yourself what you want to see. For example if you want to trade from the globex session into the open constant volume bars might be just what you need as the volume and volatility jumps. They are also great for trendlines and channels as they tend to be 'smoother' with less 'drift outs'. If you want to see price rejection the tails on constant time charts are hard to beat. Likewise if you want to see consolidations, time based chats are a good bet as price flatlining with no trades just wont show on constant range or constant volume charts. Yet again if you want a proxy for the tape to see buyers hitting a 'wall' of resting orders you might want a very fast constant tick chart this will clearly show all the 1 lots hitting an unyielding best ask. Different jobs, different tools.
If you feel that things look random I would try and find out whats going on. Something is not right in you or your friends setup. I would expect small differences rather than completely different.
As an aside I don't buy into the 'noise' argument though that is not one for this thread. If you are trading weeks to months you might see anything intraday as 'noise' that dosen't mean it is. A scalper nickel and dimeing round the bid and ask might see hourly charts as suitable for 'investors'. Put another way just because particular data is of no interest to trader Joe dosen't make that data noise.
One small side thought with constant time charts that have a volume histogram below you have an extra dimension of information (the volume histogram). For a 'fair' comparison one should put a histogram with the eliminated dimension along the bottom of other chart types. For example constant volume charts with a time histogram are quite interesting as you now have the time dimension back!
I recollect demo trading on HotSpotFx and Currenex platforms sometime ago where NO charts are available to trade with. I had to enquire from the HotSpot guys how on earth can I trade on their platform without a chart; and the response was that the platform has all the in-built features to trade without any need for a chart. I tried very hard to understand trading on these platforms. In any case, I gave up trading FX, so I did not sign-on for a LIVE account.
I want to believe that there are platforms for futures trading that do not have in-built chart features or that you do not really need to use their in-built charts (e. g. PhotonTrader). I'll prefer these "chart-less" platforms if I can understand and master trading with any of them.
As regard bad data-feed; I can also recollect that HotSpot has one of the best FX data-feeds in all the DEMOs I traded. I'll therefore discover and trade with a broker that has good data-feed.
So, my takes are;
1. Go for a "chart-less" platform and master how to trade with it
2. Go for a broker that has efficient data-feed (not all brokers have bad data-feed)
3. Ensure that you have efficient internet service
4. Have a good hardware (I've never had any issue with computer timing on my VMWare Fusion-based Windows XP running on my MacBook, maybe because the XP picks its time from my MacBook OS X).
Hi, The Wizard, have you had any computer timing issue on the Windows running on your Mac ? Please, let's know.
It would be interesting to learn how to trade without a chart. I did a quick search on the internet and didn't turn up anything of value, so I might have to try again another time. If anyone has a link for this idea I'd love to check it out.
When I trade I have 2 longer-term charts and one short term. Without a doubt, I trust the long charts more for several reasons. First, as mentioned above, it gives me a clearer view of the big picture. Secondly, I don't worry if the data is a little dirty. Though it may be off a bit it's not that far off.
My internet connection is pretty reliable and my operating system has been solid (XP Pro). I'm not too sure about my data provider but will definitely do some checking there.
So I'm trying to gather the general consensus here. One idea is to learn to trade without any chart at all. Another is to use no chart faster than a 5 min chart. I guess I'm OK with either idea and could adapt my methods accordingly. It just really stinks that here we are in 2010 and we're operating off data that seems like it comes from the 1980's.