Volume bars, as opposed to equivolume bars, provide a unique perspective on order flow and price action. One volume bar represents X number of shares/contracts traded, where X can be any positive integer. The SierraCharts volume bar is one example of the volume bar being referred to herein; other platforms may have it too. Typical numbers might be 50, 100, 200, 1000, etc. Using this indexing method we now see price movement versus number of shares traded rather than time or range. A tick chart is similar to a volume bar chart except that instead of count the number of shares/contract traded, the number of trades are counted. So in a tick chart a 10 contract trade increments the count by only 1; whereas in a volume bar chart, a 10 contract trade increments the count in that bar by 10. One often hears of the old adage “volume is cause and price is effect”. In this model, price and volume are the two dimensions of order flow with volume on the x axis and price on the y axis. With this charting method the volume study, usually below the price chart, is unnecessary.
Order Flow the exchange of shares/contracts at bid or ask that results in price movement. The trader’s goal is to detect shifts in order flow as soon as possible so as to be able to enter and exit a trade with as much profitability as possible.
In the markets, there are 4 types of players: Bullish Offense (Market Buys), Bullish Defense (Bids), Bearish Offense (Market Sells), and Bearish Defense (Asks) . Two types of exchanges occur: between Bullish Offense and Bearish defense (Market Buy orders fill Asks) and between Bearish offense and Bullish Defense (market Sell orders fill Bids). Exchanges between Market Buys and Market Sells are infrequent and, therefore, are ignored in this model.
Volume, or total volume equals bid volume plus ask volume. In a completed volume bar total volume is always the same: the pre-defined value. However the bid and ask volumes will vary but their sum will always be the same. The percentage of bid and ask volumes in each bar can give the trader insight into order flow.
What we can see:
Traders can see two things: the depth of market (DOM) ladder and the actual traded bid and ask volumes. For this model, we do not care about resting Bids and Asks. They can be pulled at any time and do not equate well to actual traded bid and ask volumes. So all we are concerned with in this model are bid and ask volumes, as well as the resulting total volume and the price movement.
Price movement occurs when enough contracts are exchanged so that the quantity of Bids or Asks, the defensive orders, are used up at current price. For example: Suppose there are 250 asks at 50.25. They may not be visible in the DOM ladder but they are there nonetheless. Market buys (bullish offense) begin to come into the market and when 250 of them occur, enough to fill the 250 asks, then if there are still more of those market buy orders, price will go up to 50.26. If there are still sufficient market buy orders to fill all of the asks at 50.26 then price will rise again, and so on and so forth. Likewise, market buys could cease and market sell orders could come in. Now the price would go down to the closest bid price (the inside bid). Think of it as “Blinky” in Pacman eating his way through the cheeses – except that the number of cheeses at each location varies as well as the speed of Blinky’s chewing.
Using this type of charting, certain aspects of order flow pop out once you learn how to read it. The first thing you will get into your head is that each bar represents a fixed number of shares/contracts changing hands. So now when we look at the size of the bar (range or open-close difference) we are seeing how many ticks price has moved for a fixed number of shares/contracts traded. And if we see a series of bars moving in one direction we can see in an instant whether that move required a great deal or a small number of shares/contracts.
Offense and Defense
For example: suppose we see a 50 tick move up in CL (crude oil) over ten 200 volume bars. That means it took 2000 contracts to move price up 50 ticks. Then suppose price dropped back down 50 ticks in only 3 volume bars, meaning that only 600 contracts had to change hands to move 50 ticks. 2000 contracts up, 600 down. That is a significant difference. But what does it mean? Going back to our definitions, we know that on the way up market buy orders had to take out 2000 asks, so there were a lot of asks to take out. There was a lot of bearish defense that bullish offense had to overcome. There were a lot of cheeses Blinky had to eat. Note that I did not talk about the amount of time it took to accomplish this. But on the way back down it took only 30% as many bars to move 50 ticks. Bullish defense was much weaker on the way down than bearish defense was on the way up. In both cases though, the two offenses were up to the task. You might see it on the chart with 10 bars of 5 tick height (on average) while on the way down you see 3 bars of almost 17 tick height (on average). The down move was much steeper. The taller the candle, the steeper the move, the weaker the defense is (fewer bids or asks).
When we look at a normal time based chart (time increment is constant on the x axis) and see a steep move, we conclude that a strong move occurred. But in a volume chart it is just the opposite: a weak move occurred, defense faded out. Volume charts, to some extent, require that you invert your thinking. This is good because it opens up your mind to a different way of thinking about price action and order flow.
Another interesting thing you will see on a volume bar chart is a series of bars with zero or 1 tick height – no price movement. In CL, you might eight 50 volume bars in a row with no price change. That’s 400 contracts all exchanging hands at the same price. You might think there was no action going on. On the contrary, there was an epic battle going on - mano a mano and nobody was prevailing. Sooner or later though, someone will prevail of course; and sometimes it will result in a big move – a breakout. You will see price action move up to a new price and then stop and stay at the price as if it had hit a brick wall. Clearly bearish defense (asks) were piling on to prevent any move upward movement – and it may not have been apparent in the DOM ladder. You may not notice this with a typical time based chart unless you have a volume study and are trained well enough to see, out of the corner of your eye, volume rising up to a relatively high value. With volume bars, it’s right there on the main chart with more and more bars forming without any price movement.
Cumulative Delta and Ask/Bid Difference
Volume = Bid volume + ask volume. Ask/Bid Difference = Ask volume – Bid volume. Suppose in a 200 volume candle that moved up 4 ticks, 50 contracts traded at the bid and 150 traded at the ask. The ask/bid difference for that bar would be 100. If we started at zero then we now have a cumulative delta of +100. If the difference on the next bar is +100 then cumulative delta is now equal to +250. If the difference on the 3rd bar is -150 then cumulative delta is now +100, and so on and so forth. Cumulative delta is the integral or sum of the ask/bid difference.
Here is where things get a bit tricky. Suppose you have a 200 volume candle with the close 5 ticks above the open and an ask/bid difference of -50. How can that even be possible? It is possible because price action does not necessarily go from 0 to +5 ticks in a straight line. It can zig zag up and down in some torturous manner to get there. And when price zagged down perhaps there was a lot of bullish defense there to contain the downward movement. This is where we have to think in terms of a four-player game: not bulls vs. bears, but rather bullish offense vs. bearish defense and bullish defense vs. bearish offense. Clearly, on that candle the bulls won as price went up 5. But the ask/bid delta was -50 so there were 125 bid trades and 75 ask trades. Thus, bullish defense was doing a lot of work and bearish defense was slacking off. This is an aspect of order flow that may give you a leg up in this game.
With this knowledge we can look at a series of price moves that make, say, a V shape. Price goes down and then comes right back up to the initial price. Now look at the cumulative delta. Suppose it went down then came back up but to a lower level. What does that mean? It means that, overall, bid volume exceeded ask volume over the entire V shaped move. What does that mean in terms of the four players? It means that on the way down there was more bullish defense than bearish defense on the way back up. What if cumulative delta came back up but to a higher level than where it started? It means that on the way down there was less bullish defense than bearish defense on the way back up.
What if price is going nowhere in one of those epic mano a mano batlles, but cumulative delta is rising? That means bullish offense (market buys) are meeting significant bearish resistance (asks). One of two things will happen to resolve the battle allowing price to move. Either bearish resistance (asks) will finally dissipate allowing price to rise, or bullish offense (market buys) will whither and price will fall.
Note that I haven’t talked about the intensity of offense (market buys and sells). The intensity of these two players often affects the speed (how quickly) with which price action changes. You can get an extra sense of price action from this, but I’m not sure it tells you anything about order flow.
That is all I have right now.
The following user says Thank You to wmueller for this post:
Your understanding of the 4-way game is correct. However, there is a huge gap between this understanding and its successful application to live markets and I am sure you know that. Some of the things you mentioned apply more or less depending upon how thick or thin the market is but they become clear once you start specializing in a particular market.
Overall, I think you are on the right path. Keep plugging away, start a journal and see if that helps.
The following user says Thank You to Hulk for this post:
I have a very high regard for @DionysusToast and I value what he says a lot.
What he said, translated to your terms, means that you would want to trade your bearish and bullish defense only, but not the first of those defenses (those become the turns) but the ones that follow that first defense (now you are trading momentum). Ideally, the first defense occurs at your anticipated level and then you get in on 2nd defensive hold. If you are scalping, then you scalp those defensive holds until the reverse happens. If you are scaling in, then you add on the defensive holds etc. If there is no follow up defensive hold, then its just chop and you stay out. No, its not as simple as that but its a good way to start applying your model in the beginning.
1. What is the market signaling at this moment?
*Is PA convergent or divergent with Cumulative Delta (CD)? Is PA consolidating, expanding, ranging? Is defense or offense controlling?
2. Who is paying up to get in or out?
*Market Buys/sells, or bids/asks.
3. How much strength is there?
*Small range bar = lots of defense bids/asks. Quick action = lots of offense (market orders)
4. Is momentum building?
*Look at slope. Are the volume bars getting larger?
5. Can it be measured relative to something?
*Compare the size (range) of the volume bars.
6. What would have to happen to indicate the momentum is changing?
*PA – CD divergence, PA slope change
7. Is the trend weakening or is this a normal retracement?
*PA – CD divergence, PA slope change. Compare the size (range) of the volume bars.
8. What would show that? If the market has displayed a fairly symmetrical type of pattern and that pattern has been disturbed, then it is a good indication the balance of forces has shifted.
*Look slope changes, bar length changes and cumulative delta divergences.
9. Are there any places where one side will definitely gain dominance over the other? If that point is reached, it still may take sometime for the other side to be convinced they are losers. How long am I willing to give them to rush out of their positions?
*Look for support/resistance, swing highs/lows that show a top/bottom, supply/demand.
What happens with PA there? Sideways action – could go either way, must wait. Reversal, wait for the pullback. Ploughs through – wait for a pull back. Always look for a spot where you can define a low risk point where your expectation is wrong and you would exit the trade (a low cost fail point).
10. If they don’t rush out, what will that indicate?
* If PA doesn’t move and CD doesn’t show a divergence, then there is agreement and lack of activity.
11. What did traders have to believe to form the current pattern relative to the past? Remember that people’s beliefs don’t change unless they are extremely disappointed (by unfulfilled expectations).
*Thee is always disagreement in the market; otherwise prices don’t move. It takes a stronger belief on the part of the bulls (bears) for prices to go up (down). The move can brought about either by offense or defense.
12. What will disappoint those with the predominant force?
*Lack of strong enough PA in the desired direction or a move in the wrong direction.
13. What is the likelihood of that happening?
*Must have statistics on the PA and CD patterns you are seeing to know what the likelihood is.
14. What is the risk in finding out in trading?
* The risk is difference between entry and fail point.
15. Is there enough potential for movement to make the trade worth the risk?
* must look for a likely target based on previous edge levels and volume-at-price structures to come up with an educated guess. Bu remember it is only a guess. Anything can happen.
Any thoughts would be most welcome and appreciated.
Last edited by wmueller; April 22nd, 2015 at 12:25 AM.
Reason: forgot some definitions
The trouble with CD is that, like many other signs of possible divergence, it works sometimes, and sometimes does not (by "work, I mean, it does indicate a divergence with price action and reversal or/and divergent move about to happen).
We would all agree that one can not rely to take trades that sometimes work and sometimes don't, altough this is what happens effectively in the sense that we always have winning trades and loosing trades, but one can not establish a trading strategy based on something that is known in advance to work half of the time, or one might as well trade with a coin, and get similar results, whith heads or tails.
Part of the reason why it may or may not indicate a real divergence, is because market orders sometimes get absorbed by Limit orders, and sometimes do not. Another part of the reason is that it might "work" from a divergence perspective still, if there is some follow up to the CD, whether markets orders get absorbed or not, which in turn, very much depends on whether or not a wave of market orders is an intentional agression (human or algo based but this is not the issue here), or just simply passive and therefore not aggresive maket orders (stop order which became market orders and got triggered on given levels).
The same applies to volume imbalances, for the same reasons, which in itself, does not mean much, and to some other order flows concepts.
The real and only relationship is Offer vs demand, so a divergence in itself only means something if compared to the right thing, i.e, Market orders Vs Limit orders.
I have been trading the order flow for some years now, and this is the main relationship I care about. There is no systematic green/red light to trigger trades, and very much of a discretionary approach(altough statistically, some divergence or else might generate postive results, I don't know, we'd need to see what an algo would produce trading CD or VI).
A good DOM (I use Jigsaw and can not do without it), a good news platform (Metastock offers Thomson Reuters Eikon for $150/month for non professional investment companies), a good understanding of the market you trade, supply and demand laws, "patterns" on the DOM (algos leave their footprints), know the typical range of the instrument that you trade, its average true range, how it reacts to economic announcements, how it reacts to general news, is it volatile (is there liquidity, when, why), how does it behave through various sessions (Europe, US, Asia), etc, in short understand and know the contract(s) that you trade, a decent charting platform (I used Sierra charts) with basic footprint charts/candle charts with volume profile to do morning and end of the day homework with key levels, areas of interest that are likely to be revisited (again it comes back to understanding the mechanics of the markets and why we might come back to some levels), VWAP and its SDs (simply as 70% of algos are vwap based at some level or another), a decent account (capital) and this is pretty much all that is needed, along with years of practice. I also think, and this obviously goes against all marketing/commercial intents to sell the Retail Traders all kind of services and tools, that one should mentally reverse all promises made by Brokers offerings: How to make $10K with $500, becomes how to make $500 with $10K.
Only once I understood, accepted, and spent some time on the above, did I start to be a profitable and consistent Trader in the long run.
The following 6 users say Thank You to dom64 for this post:
Assuming your comment is meant to be saracastic, no this is not the Holy Grail of Trading, and not meant to be. This is what I use (I don't make any money by listing what I do, just meant to say that it doesn't require much on the contrary), and this works for me. If something else works for you or for somebody else, this is equally fine, I respect whatever works for others, and have nothing to prove to anyone. I just wanted to share what took some years to come with, as this is my full time occupation.
Please do not hesitate to enlighten us with equally as interesting and useful comments as the previous one, so we can all benefit from your experience and great wisdom too.