Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
When I took my first trade 12 years ago, all I knew were charts, and the candles or bars that filled them. I quickly learned, as you all have done, to see the "story" associated with a series of candles/bars. I followed a guy who uses "price action trading" and who looks at bar-by-bar analysis, and it seemed to make sense to my 6-month-old trading brain.
However, as you begin to understand even the basics of market microstructure (bids, offers, etc.), you should begin to see the market as buyers and sellers who are matched with each other to produce a series of transactions. For a very active market, with tens or hundreds of thousands of transactions per day, we would be overwhelmed if we viewed each of these transactions in isolation. So, for the sake of our ability to grasp the bigger picture, we aggregate these transactions.
The most straightforward way to aggregate transactions is by time. You group all transactions that occur within a particular time interval (say, 5 minutes), and you choose to display statistics about them. Among those statistics could be the average, max, min, first, last, median, variance, and the list goes on. Your standard candle/bar chart chooses to show 4 pieces of data: the price that traded first (the open), the price that traded last (the close), the max traded price (the high), and the min traded price (the low). This is your OHLC candle/bar. So, over a given time period, you know the very first price that traded, the very last one, the highest one, and the lowest one. If you think about it, that's not much data. Out of potentially thousands of transactions during that time window, you know 4 of them.
Now let's talk about the closing price. Let's use a 5 minute bar, since that seems to be the favorite of most. When you view the market as a series of transactions, the close of a 5 minute bar is simply the last transaction that took place before 9:35, 9:40, 9:45, ... 16:00. I ask the question: what is special about those times? Is 10:40 meaningful, in some way? 14:15? The strategy that says "the bar must close below price X" places importance on a particular time of day. What if you begin your 5-minute chart at 9:31 instead of 9:30? All of a sudden, 9:36, 9:41, 9:46, ... become important, and the closing price at 9:35, 9:40, and 9:45 are completely hidden from your view, lost in the aggregation of the bar.
A 5-minute "reversal bar" that closes at 10:45 may be a "doji" on a 15 minute chart. When the period of aggregation is arbitrary (1 minute, 5 minutes, 15 minutes, 3 minutes...), the story that those aggregations tell is also arbitrary. What happens beginning at 10:45 is the same experience viewed from any angle, regardless of whether you have aggregated data from 10:30 to 10:45 over 3, 5, or 15 minutes.
It's arguably worse if you use a non-time-based bar to aggregate transactions. "The 233 tick candle must close lower to go short" is perhaps the height of what I can only term superstition. What is so special about the 233rd transaction that we must designate it as the close of a bar, and then place some meaning on it? "I'm going to wait for this 5 minute bar to close before I enter" may be even worse -- so, at 9:49 it's not a buy, but at 9:50, at the same price, it is? There's no logic there. Perhaps it all comes down to seeking confirmation--but confirmation comes from many factors, of which time is probably not the most important, particularly not an arbitrary points during the day.
The exception I make are daily bars, particularly in equities, since NYSE equity trading is closed from 20:00 to 4:00, and since the day is so well defined to be liquid from 9:30 to 16:00. Settlement of futures occurs at a particular time, accounts are marked to market, gains and losses are calculated, and margin calls are made. Trading is actually closed, for example, in the ES, from 17:00 to 18:00, and margin calls occur before this. However, since other markets may be open, even the daily close is not as important as the weekly close, which happens from US afternoon until Australia/Japan morning the following Monday--a true 48 hour break from any activity.
So, what's a better way to look at the market? Well, that's a whole other subject. But once you eliminate ideas that have no sound basis in logic, you are left with fewer to choose from, which is a good thing! Disagree, agree? Talk about it!
time is relevant at certain intervals.
the daily close is very important for obvious reasons.
the 5min close is important in that many traders/algo's seem to execute on the open of time intervals. simply pull up a 5-sec chart and you'll see volume spike up at a lot of 5min intervals.
Lately i've been trading with a timeframe that is around 1-2 minutes, and use only H/L bars for a clearer picture. But you do lose some context. With closes, you are able to see rejections quicker than if you, say, took entries at a break of a bar high/low. When you see a wick come down off a key area, it doesn't matter what interval you use, because regardless there was some sort of pushback in price - something that is important to see.
I totally agree with you....if one uses candles to trade the way you have described.
I do believe that they have a place though. Not everyone has the ability to read a DOM and time and sales for example and figure out where buyers and sellers are meeting and one taking over from the other one. For some people who are more visual interpreters a picture makes a lot more sense. Just from the attached screenshot (CL for today, 800T) you can for example see where selling took over from buying and vice versa (also popularly known as support and resistance). 800T is arbitrary and was not chosen to reflect any sort of magic number. I agree the market doesn't care about these magic time frames or ticks or whatever. It is more about what makes the picture make sense for you.
This gives the opportunity to people who does not get the ladder type trading to also participate in trading. We can debate how effective it is etc. but the fact is that it can be done fairly profitable if practiced enough and getting emotional state in check etc. etc. Just look around this forum and we will find many people who trade this way and make a fairly good income from it.
We are unfortunately in an industry where unscrupulous people have seen lots of opportunities to make a quick buck from people who are trying to learn to trade. A lot of these people have made trading in general (not just candles) into something that it was never intended to be.
You're going to have some difficulty convincing a number of people about this. (I do agree with you, by the way.)
There are a number of successful traders who think that, for example, the close of a 5-minute bar is very important, and that the choice of 5 minutes (or some multiple of it) is also very important. If a trader is profitable and has that belief, it is hard to convince him to change it. It appears to him that his success proves the correctness of his approach... and in a sense it does, of course, but it doesn't necessarily prove that nothing else can also be "correct" in the sense of "can be profitable," and it doesn't prove that his particular bar type is "the only right one" or that nothing else works.
But this also does not invalidate his using that type of bar in that way, if it works for him.
My view is it doesn't matter, you're just sampling the trading activity in particular span of time (or number of trades or contracts or range of price change or whatever), and that one sampling is as good as another. But I don't insist on it.
And whatever works, does work, after all.
So I do agree, but I don't think you'll make a great amount of difference for anyone who already has their mind set one way or the other.
But I like Cervantes, and tilting at windmills sometimes too.
Bob.
When one door closes, another opens.
-- Cervantes, Don Quixote
I lot of people do it. The chart below is the average volume from each second in the day from Sept-Oct 2015 - includes ON data. There is an average of 40% more volume that is transacted in the opening second of a min - i.e so trades initiated based on the close of a bar.
I use volume bars, with automated entries based off the SC spreadsheet. The backtests are based on closed bars - at which point the 'features' are hard coded. There is no practical way to test and trade 'intra bar' (at least in Excel), as the triggering feature may change back and forth multiple times during the bar. The amount of data required to test for that would cripple Excel.
I'd suggest for the majority of systematic traders, closed bars is the only practical method.
While I think you do have a very strong and compelling point, and I fully agree with, regarding the "superstition" aspect that's behind why many traders use many types of charts, there are some things I thought I could share a slightly different perspective. By the way, this is quite an interesting topic!
So, just enumerating some points to clearly separate them, as they are mainly independent.
First, one simply cannot mentally process the huge amount of raw data generated by the market - we need to somehow compress it. Maybe by summarizing, maybe by sampling. A candlestick is nothing more than one way to summarize the data. A line chart is nothing more than a way to sample the data.
Second, it's a reasonable assumption that people in general feel comfortable displaying timeseries data in graphs with time on the horizontal axis. That's what we see almost every day in news since we start being exposed to newspapers as kids. So why fight it? In particular, people are used to seeing data where the time axis is uniform, that is, each segment of the same size represents the same time interval.
Based on this 2 initial points, it's pretty much a logical conclusion that displaying market data as either candles/bars (summary) or lines (sampling) on regular time intervals (5min, or whatever it is) just makes sense - there's less cognitive load, so the brain has more resources left to actually process the summarized/sampled data and make trading decisions.
Third, there actually is something special about certain intervals. You said you see why the daily interval is important. But many others are important, too. Not only algorithms may be looking at those regular intervals, but a lot of information is actually generated or published at the boundaries of those intervals. For example, many reports are released at specific intervals, or at specific hours. In some markets, some kind of average of trades that occur during certain intervals are of extreme significance as they define settlement prices that will impact cash flows and options exercise. The activity around these regular events makes certain intervals particularly importants!
Fourth, the fact that the same value at an instant may represent different things at different time intervals (ie, your reversal bar vs doji example) doesn't invalidate those (possibly conflicting!) analyses. We simply can't know what exactly is driving the market at any given time; but if you're able to tell yourself a plausible, convincing story that accounts for the data you're seeing, it may be enough to justify entries and/or exits which, with proper risk management, can become a profitable system. Also, if you look at different summarizations of the data (eg, a 1 min chart plus a 30 min chart, or maybe a volume chart, or whatever), conflicting analyses may be an indication of weaker entries/exits, while if they agree then it might be an indication of stronger entries/exits.
Fifth, and last, but not least, regarding these odd choices of things like 233-tick, or a 750 delta-volume, or a 12-4 range reversal, or whatever weird method to define the end of the current aggregation and the start of the next, for the summarization process, I don't see it as "magical numbers", but simply values/settings that allow me to see information with a sufficient amount of details, while eliminating as much noise as possible. For example, one of the charts I like to look at when trading the ES is a delta-volume chart (ie, new candle when the number of buys exceeds the sells, or vice versa, by a certain amount). I may use 250dv, 500dv, 750dv, 1000dv, or anything, really. The numerical setting there isn't important - what matters is that, in a given market context, I want to see useful data while eliminating as much of the noise as possible. On days with heavier activity, I may choose higher numerical settings (750dv, 1000dv, or really any arbitrary number); lighter activity, lower number. What I want is to keep the "rhythm" of my chart approximately the same, regardless of the characteristics of the specific day.
Is one bar type, or the close of one bar type, important or necessary for trading? (A slight rewording of the original question.)
I didn't think initially that this would be such a great topic for discussion, but already I've been proven wrong. These responses are very interesting, and are somewhat nuanced as well.
Let me toss in one additional idea. (And I am not saying that anyone who does or thinks differently from me is wrong.)
If a particular bar type or bar time/tick/volume/range/whatever setting were the only "correct" one, then nothing else would work. But many traders use other types and settings of bars, sometimes very unusual ones too. One very interesting guy ( @lancelottrader in his "The Crude Dude Oil Trading System thread:
Hi everyone. I'm new on here, but I thought I would make a post about my Trading style, since I feel it's fairly different than what I usually see out there. I have been trading about 6 years, mostly Forex, but have transitioned to Futures a few …
) uses mainly a line-only 2-tick range chart (!), plus a few others for context. He does well with this, in CL and NQ, two very fast markets. There is no open, no close, and even no bar on these charts, just a rapidly-moving line. (His videos are interesting. I would never try it, though.)
You can certainly make the argument that a particular chart/bar type is watched and acted on by many or most traders, computer algos, or whatever, and this may or may not be the case. (How would anyone know?) I am sure that a lot of trading decisions are made on the basis of some bar closing, just because using a close is a convenient thing to do.
But, for example, I recently moved from ES to YM. I had long used volume bars on ES, and I had settled into a round-number for my bar size (it was based on 6000 contracts, I believe), and then I scaled up and down by just dividing the 6000 by 2 or 3 or something.... I don't exactly remember, and I don't think it's important. As an aside, a lot of volume-bar traders use something like 2,000 for ES, so there is probably some confluence of bar sizes here, but also there is a simple appeal to numbers ending in "000" .
To transition to YM, I figured the average volume in ES and in YM, calculated the percentage YM was of ES, divided my ES chart numbers by that percentage, adjusted the number so it would end in a convenient number of zeros, checked the resulting YM volume chart compared to my ES charts, fiddled with the numbers until the charts kind of looked the same most of the time, and arrived at a roughly equivalent volume bar number (remember the ES number was fairly arbitrary in the first place.) I then subdivided down by factors of 2 until I arrived at my usual trading chart, which is 188 volume bars (It's actually 187.5, rounded up.)
How many traders in the world are using 188 contract-bars on YM? Probably some, because the process of getting there was not totally arbitrary, but certainly not many.
Now, I am not a great trader, but I am a decent reader of charts (which is not really that useful in trading, as every trader eventually finds out), and the charts make the same kind of sense and the same things work. Trendlines, MA's, chart patterns, measured moves, and on and on, all that stuff works the same exact way. I even use candlesticks the same way, It all works the same. (Well, until I start trading, and then the nut in my head starts making the usual dumb moves, but I/he was making those same moves on a 5-minute chart too.)
I think this argues well for the "fractal" hypothesis of the market (which is that market movements can be subdivided or expanded and they will show the same patterns), although it doesn't necessarily "prove" it either. It's just consistent with it. My hypothesis is more radical: I think you can slice up market movements all sorts of ways, and they will still show the same patterns. This isn't proven by any of this, either, but it is consistent with it.
The hypothesis that some some periods/types of price representation are specially good is also argued for, persuasively for many traders, by the fact that trading with them can be profitably done. But this also doesn't necessarily "prove" it. It's just consistent with it.
Perhaps there is more than one workable picture of reality?
My bottom line is only that when something works (which has a very simple meaning: it makes you money), then feel free to use it. Why not?
Bob.
When one door closes, another opens.
-- Cervantes, Don Quixote
In the last few months, I have actually changed my setup and am using 5 range hollow candle bars and 1 minute equivolume candle bars. Think I am doing even better now..
