Lessons From The Multi-Pivots For June 20, 2011

[Billy]

8893

The roadmap outlook for Monday is very bullish for SPY and GDX, slightly bullish for IWM and extremely bearish for QQQ.

The QQQ is the real source of technical concern for this market, plunging below the 200-day moving average and with almost no support left below last week’s lows. It is now in its first week of weekly “early distribution” sub-stage which is the very last neutral to slightly bullish point in the intermediate term stages cycle.
The two support clusters are totaling a bleak strength of 7 versus strength of 26 for the two resistance clusters! This is an extremely bearish imbalance rarely seen for any index. A down trend day on Monday could go at full speed to WS2 (52.44). However, the bullish outlook for SPY and IWM is not warning of an overall market washout. Most important too, as mentioned in other posts, MS3 (53.90) has a 95% probability of marking the low for the month of June and many algorithms must be programmed to stop selling at the current level. If you actually adjust Friday’s ex-dividend close with the 12.1 cents dividend, the adjusted close was actually 53.91, just above MS3! The adjusted first support cluster would therefore become 7 instead of 4. This MS3 of QQQ will be the most important level to watch on Monday for gauging the chances of a successful bullish reversal in SPY and IWM. QQQ should really not try to run away decisively to the downside from here and any attempt to do so must quickly be bought heavily by large players to avoid cascading panic selling.

8894

The outlook for SPY is much more reassuring and the clusters’ support/resistance ratios are very bullish with 19:8 for the first clusters and a total of 22:11 for the two clusters taken together. Here also the MS3 (125.99) becomes important and is in perfect confluence with the 200-day moving average (126.07). I would almost bet the castle that it is where the selling programs will stop and reverse until the end of the month.

8891

IWM is slightly less bullish than SPY but will be mostly helped by its improving Relative Strength vs. SPY. The first support/resistance clusters strength is 14:12. In the hypothesis of a last capitulation selling flush early next week, the robot’s trailing stop at 76.97 is just under the confluent cushion of the 200-day moving average (77.26) and the new WS1 (77.08). If the stop is hit, the robot will exit the long position. Due to the very strong buy probabilities, it will likely look to re-enter a long position on the next day but at a new entry price that could differ greatly depending on Monday’s close. The robot executes its plan first and never tries to anticipate.

8896

Although the new long buy signal for GDX is rather weak in terms of LT/ST probabilities, I like the proposed setup a lot. An entry at a limit of 51.57 would be just above the first support cluster with a good strength of 11 while the initial stop at 48.09 will be below a potential down trending day range and further protected by the second support cluster also strong with strength of 10. The initial one-day move after entry should be easily driven to the upside thanks to the weak first resistance cluster’s strength of 5.
Billy

8895

[slgerritz]

Billy or Pascal,
The GDX robot gave a new signal to buy at $51.57. Is the robot telling us to wait for a drop to the $51.57 before buying? If so, the strong MF into the sector may not give us to opportunity to by at $51.57. If the train leaves the station should we chase the price or wait for robot entry price?
Steve


8893

The roadmap outlook for Monday is very bullish for SPY and GDX, slightly bullish for IWM and extremely bearish for QQQ.

The QQQ is the real source of technical concern for this market, plunging below the 200-day moving average and with almost no support left below last week’s lows. It is now in its first week of weekly “early distribution” sub-stage which is the very last neutral to slightly bullish point in the intermediate term stages cycle.
The two support clusters are totaling a bleak strength of 7 versus strength of 26 for the two resistance clusters! This is an extremely bearish imbalance rarely seen for any index. A down trend day on Monday could go at full speed to WS2 (52.44). However, the bullish outlook for SPY and IWM is not warning of an overall market washout. Most important too, as mentioned in other posts, MS3 (53.90) has a 95% probability of marking the low for the month of June and many algorithms must be programmed to stop selling at the current level. If you actually adjust Friday’s ex-dividend close with the 12.1 cents dividend, the adjusted close was actually 53.91, just above MS3! The adjusted first support cluster would therefore become 7 instead of 4. This MS3 of QQQ will be the most important level to watch on Monday for gauging the chances of a successful bullish reversal in SPY and IWM. QQQ should really not try to run away decisively to the downside from here and any attempt to do so must quickly be bought heavily by large players to avoid cascading panic selling.

8894

The outlook for SPY is much more reassuring and the clusters’ support/resistance ratios are very bullish with 19:8 for the first clusters and a total of 22:11 for the two clusters taken together. Here also the MS3 (125.99) becomes important and is in perfect confluence with the 200-day moving average (126.07). I would almost bet the castle that it is where the selling programs will stop and reverse until the end of the month.

8891

IWM is slightly less bullish than SPY but will be mostly helped by its improving Relative Strength vs. SPY. The first support/resistance clusters strength is 14:12. In the hypothesis of a last capitulation selling flush early next week, the robot’s trailing stop at 76.97 is just under the confluent cushion of the 200-day moving average (77.26) and the new WS1 (77.08). If the stop is hit, the robot will exit the long position. Due to the very strong buy probabilities, it will likely look to re-enter a long position on the next day but at a new entry price that could differ greatly depending on Monday’s close. The robot executes its plan first and never tries to anticipate.

8896

Although the new long buy signal for GDX is rather weak in terms of LT/ST probabilities, I like the proposed setup a lot. An entry at a limit of 51.57 would be just above the first support cluster with a good strength of 11 while the initial stop at 48.09 will be below a potential down trending day range and further protected by the second support cluster also strong with strength of 10. The initial one-day move after entry should be easily driven to the upside thanks to the weak first resistance cluster’s strength of 5.
Billy

8895

[Billy]

Billy or Pascal,
The GDX robot gave a new signal to buy at $51.57. Is the robot telling us to wait for a drop to the $51.57 before buying? If so, the strong MF into the sector may not give us to opportunity to by at $51.57. If the train leaves the station should we chase the price or wait for robot entry price?
Steve

Steve, if you want to take advantage of all the edges included in the robot design, you need to wait for an entry limit to be hit. A big chunk of the outperformance of the robots is stemming from waiting for the optimal entry points. The initial optimal stop is an important function of the limitentry price. No need to chase the train blindly, tomorrow may offer another optimal entry limit for the risk management system of the robot. If you buy at another point, the robot management becomes useless and it becomes pure discretionary trading.
Billy

[Pascal]

Billy or Pascal,
The GDX robot gave a new signal to buy at $51.57. Is the robot telling us to wait for a drop to the $51.57 before buying? If so, the strong MF into the sector may not give us to opportunity to by at $51.57. If the train leaves the station should we chase the price or wait for robot entry price?
Steve

Really, we cannot advise explicitly on this point.
In the long run, buying at the open for GDX is also working fine as the back-test has shown.
On the other hand, the GDX ATR is 2.82% right now. The buy point is 0.77% below last Friday's close. This means that statistically, the buy point is well within the usual day volatility.


Pascal

[slgerritz]

Pascal,
Where is the GDX ATR of 2.82 derived from? On freecharts.com is says it has a daily ATR 14 of 1.46.
Steve



Really, we cannot advise explicitly on this point.
In the long run, buying at the open for GDX is also working fine as the back-test has shown.
On the other hand, the GDX ATR is 2.82% right now. The buy point is 0.77% below last Friday's close. This means that statistically, the buy point is well within the usual day volatility.


Pascal

[Pascal]

Pascal,
Where is the GDX ATR of 2.82 derived from? On freecharts.com is says it has a daily ATR 14 of 1.46.
Steve

We are using a 20D ATR.

Pascal

[adam ali]

Billy,

Stockcharts.com have recently instituted pivots on their charts. In their blog, they list their standard pivot calculation as the following:

Pivot Point (P) = (High + Low + Close)/3

Support 1 (S1) = (P x 2) - High

Support 2 (S2) = P - (High - Low)

Resistance 1 (R1) = (P x 2) - Low

Resistance 2 (R2) = P + (High - Low)

This appears different than the calculation used on this site. If true, do you happen to know why the difference?

Adam

[Billy]

Billy,

Stockcharts.com have recently instituted pivots on their charts. In their blog, they list their standard pivot calculation as the following:

Pivot Point (P) = (High + Low + Close)/3

Support 1 (S1) = (P x 2) - High

Support 2 (S2) = P - (High - Low)

Resistance 1 (R1) = (P x 2) - Low

Resistance 2 (R2) = P + (High - Low)

This appears different than the calculation used on this site. If true, do you happen to know why the difference?

Adam

Adam,
I have no idea what they're trying to do. First, using only 2 levels of support and resistance is useless. Second, their computation, even with their formulas are wrong.
They probably are trying to align on freestockcharts because I saw that they added VWAP too.
I don't recommend relying on their pivots as they are displayed now.
Many sophisticated variations of the formulas have been introduced over the years and most were only marketing gimmicks. What I'm sure is that the classic formulas are working perfectly for setup planning purposes so I don't see any use deviating from them. Stockcharts is an excellent technical analysis resource, but not a very good trading resource.
Billy

[adam ali]

Thanks for the heads up on this. One final question and it may be in one of the (few) books/articles on pivots out there: how did this methodology come to be in terms of its statistical/theoretical foundation? I'm wondering how someone one day decided to take the high, low and close, divide by three, etc. and said "That works!"

[Billy]

Thanks for the heads up on this. One final question and it may be in one of the (few) books/articles on pivots out there: how did this methodology come to be in terms of its statistical/theoretical foundation? I'm wondering how someone one day decided to take the high, low and close, divide by three, etc. and said "That works!"

If anyone can find a reliable historical source for this I would be most grateful!
My hypothesis is that many decades ago, before the age of computers and light-speed trading, when most floor traders were lowly-educated people who never graduated in anything, they needed easy-to-compute formulas for defining intermediate levels for sizing in and out of their positions.
One day a statistician made a study and probably discovered the bell curve standard deviation structure of the pivot formulas. It started to work for one, then a few, then several, then many floor traders as the word spread around, further reinforcing the self-prophecy nature of the pivots.
I remember my first days as a market maker trainee, expecting complicated technical analysis and charting trading lessons. I never saw any chart or any technical analysis. Only spreadsheets with the house position's average price compared to the various floor levels. And it worked!
Billy

[ernsttanaka]

If anyone can find a reliable historical source for this I would be most grateful!
My hypothesis is that many decades ago, before the age of computers and light-speed trading, when most floor traders were lowly-educated people who never graduated in anything, they needed easy-to-compute formulas for defining intermediate levels for sizing in and out of their positions.
One day a statistician made a study and probably discovered the bell curve standard deviation structure of the pivot formulas. It started to work for one, then a few, then several, then many floor traders as the word spread around, further reinforcing the self-prophecy nature of the pivots.
I remember my first days as a market maker trainee, expecting complicated technical analysis and charting trading lessons. I never saw any chart or any technical analysis. Only spreadsheets with the house position's average price compared to the various floor levels. And it worked!
Billy

And is the same for most of the great mm's on the cboe. I met a few over the last years, most can't read a chart even if their lives depend on it.

[EB]

As to the initial formulas for daily pivots and support/resistance levels, I have vague recollections of old time commodity traders saying they existed as early as the 1970's. But, they could have existed long before and originated in other markets (perhaps even as a footnote in the Buttonwood Agreement).

The idea was apparently scaled up to multiple time frames, including weekly and monthly, by none other than John Person himself (whose book you often cite, Billy) along with some other CBOT traders in 1984. From his website:


Before going any further let me further explain who uses the pivot point numbers, mainly floor traders, hedge funds, prop traders and large speculators all use them. However, the popularity is growing as investors quest for education increases. I picked up on the method back in 1984 as a colleague was showing me how to day trade the foreign currencies, namely the Swiss Franc and the German Deutsche Mark. This was way before forex was popular. In fact trading currencies at the CME has been around since the 1970's. Pivot Point analysis and using a pivot point calculator worked fairly well in those days so I incorporated it for the market I had a passion for, which was Bonds. After using them for some time as a “crutch” I was losing interest as a day trader. I experimented with them using this concept of a longer time period namely on a weekly basis. Wow what a discovery! Then I came to the conclusion to try to use them on a Monthly basis after all I was a believer in Daily, Weekly and Monthly Charts. In fact the Chicago Board Of Trade use to give out a monthly Bond Chart and another friend of mine on the floor use to get together with me for “homework” sessions. We would do longer term chart studies using Moving Averages. I figured why not look at the monthly pivot point price targets too. WOW again! I discovered that I could look at a potential price range target in the future based on price action in the past. The longer the time frames the more power or market move would happen.

A more candid and colorful description of these after-market "homework" sessions is here (http://vodpod.com/watch/5789331-john-person-interview-recalling-his-floor-trader-days-from-the-80s), in which Person is not afraid to give Heineken its due credit.

From there, it's conceivable that the NASDAQ, perhaps with other exchanges, created the more formal system you were taught, with weightings and even higher time frames.

[slgerritz]

Pascal or Billy,

Have you created a table showing the results of the backtesting of the GDX Robot? If so can you share it?
Steve



As to the initial formulas for daily pivots and support/resistance levels, I have vague recollections of old time commodity traders saying they existed as early as the 1970's. But, they could have existed long before and originated in other markets (perhaps even as a footnote in the Buttonwood Agreement).

The idea was apparently scaled up to multiple time frames, including weekly and monthly, by none other than John Person himself (whose book you often cite, Billy) along with some other CBOT traders in 1984. From his website:



A more candid and colorful description of these after-market "homework" sessions is here (http://vodpod.com/watch/5789331-john-person-interview-recalling-his-floor-trader-days-from-the-80s), in which Person is not afraid to give Heineken its due credit.

From there, it's conceivable that the NASDAQ, perhaps with other exchanges, created the more formal system you were taught, with weightings and even higher time frames.

[adam ali]

Bob,

That's a funny interview with John Person. He seems like a guy you'd like to have a Heineken (or two) with. Sounds like he and his buddies had some great times back then while finding better ways to trade.

[Rembert]

Heineken bah. I wonder what pivot system he'd come up with if it was Duvel instead.

[Pascal]

Pascal or Billy,

Have you created a table showing the results of the backtesting of the GDX Robot? If so can you share it?
Steve

sure:

http://www.effectivevolume.eu/content/Reports/Trading_GDX.pdf

The corresponding IWM file can be found on the Robot main page.


Pascal

[Billy]

From there, it's conceivable that the NASDAQ, perhaps with other exchanges, created the more formal system you were taught, with weightings and even higher time frames.

Bob,

The NASDAQ did not create a more formal system with weightings and higher timeframes.

In the 1990's, senior mm's did manage the longer (yearly, semester, quarterly) timeframes positions and that's where the bulk of the profits originated from. Every new week and month they would charge their "mature" subordinates with the weekly and monthly sizing in and out volumes.
So, mature mm's were managing the weekly and monthly timeframes. Each day they would distribute the planned sizing-in or out volumes for the day to junior mm's.
So, daily pivots were left over to junior mm's.

The "formal" system with weightings is my own creation, and I've never seen it practiced by anyone else than by those who learned it from me. I also added the 50 dma and 200 dma to the system because these are usually the levels that institutions prefer. Most institutional decision-makers need to be protected by an umbrella to justify their decisions and these moving averges are the ideal visual and psychological umbrellas. My best institutional client was giving me the list of stocks and approximate volume he was interested in, telling me ; "Don't call me unless these stocks come near their 50 dma or 200 dma". He only gave the green light for building a position or making block trades when it was the case. I also created the cluster concept and practice.

John Person is a great guy, but his "multiple timeframes" pivot system he's talking about has nothing in common with mine. He trades each timeframe independently, while I am weighing confluences of market makers and institutional investors interests.
By the way, I've mostly stopped drinking any alcohol now, but at the time he's talking about, my favorite drink was another creation of mine that I named a Red Nose: Gin, Crème de Cassis and Ice Tea. Beware, it's very treacherous!
Billy

[Rembert]

my favorite drink was another creation of mine that I named a Red Nose: Gin, Crème de Cassis and Ice Tea. Beware, it's very treacherous!
Billy

That sounds disgusting Billy :)
But I'm not one for cocktails anyway ... except for the occasional Mojito in summer.
How about beer ? What kind do you like ?

[Billy]

That sounds disgusting Billy :)
But I'm not one for cocktails anyway ... except for the occasional Mojito in summer.
How about beer ? What kind do you like ?

The Red Nose Beer of course!

89198920

[Rembert]

Yeah that's a nice beer on occasion. But a little too sweet for my taste. I prefer more bitter beers, strong on hop content.

[EB]

I did note that Person did not talk of confluence. However, I was unaware until you wrote the below just how much of your system was original to you. Did the senior MM's follow quarterly, semester and annual pivots? Or, is that yours as well?

[Billy]

I did note that Person did not talk of confluence. However, I was unaware until you wrote the below just how much of your system was original to you. Did the senior MM's follow quarterly, semester and annual pivots? Or, is that yours as well?

Yes Bob, that's how it worked back then. The bosses are focusing on quarterly, semester and yearly pivot timeframes, distributing and spreading over time the buying and selling of the houses's long term positions down the hierarchy.
The junior guy doesn't care about any timeframe above daily, he simply receives his volume to buy or sell for the house each day and he tries to optimize the VWAP according to the liquidiy and daily pivots. New institutional intraday orders keep coming in, but they normally don't change much the house's plan of action for its own account.
I guess today this whole process is fully automated in algorithms including all timeframes levels for optimization.
The only thing original from me was simply to follow the logic of weights and cluster strengths within ATR ranges, while everybody else is focusing on individual timeframes. I am trying to trade the collectively most probable setups, not the individual ones.
Billy

[adam ali]

Billy,

Is it the case - now that market-making has now effectively been taking off the "floor" and dispersed widely - it's difficult to truly know how MM is done today? I realize that if your pivot system works, it effectively means you DO know how MMs work, but it's interesting (at least to me) that's it's a guess and not a for sure. Or is something else coming into play that makes confirming how MM works difficult?

[Billy]

Billy,

Is it the case - now that market-making has now effectively been taking off the "floor" and dispersed widely - it's difficult to truly know how MM is done today? I realize that if your pivot system works, it effectively means you DO know how MMs work, but it's interesting (at least to me) that's it's a guess and not a for sure. Or is something else coming into play that makes confirming how MM works difficult?

Adam,

MM is done today with HFT and algorithms programs. It is turning more and more probabilistic and quantitative in nature. The very stable probabilistic results of the pivots methodology comfort me in my "best guess" that they are still at the chore of the market makers' programs. The best proof or confirmation of the guess is the backtesting done by Pascal which gave a zero % long term return from the pivot methodology when used independently of any market direction bias. In other words, when blindly selling and buying (reversing) at each 3:1 reward-risk long or short setups, on average in the long term, you will not lose nor gain anything (except costs and commissions).
Once you introduce a performing probabilistic market direction model like the 20 DMF, the same 3:1 reward-risk setups entries at the model signals suddenly provide exceptional returns and reduces the risk once you start holding the position for the number of days until the next signal change or stop hit.
Billy

Billy

[adam ali]

Ok. I'm reading you to say that the fact that pivot methodology neither makes or loses money in the back tests is what is to be expected, i.e., market making is not supposed to a "profitable" enterprise (although it obviously is), it is there to facilitate the smooth functioning of markets, price discovery etc.

[Billy]

Ok. I'm reading you to say that the fact that pivot methodology neither makes or loses money in the back tests is what is to be expected, i.e., market making is not supposed to a "profitable" enterprise (although it obviously is), it is there to facilitate the smooth functioning of markets, price discovery etc.

Wrong! What you must correctly read is that the pivot methodology makes a lot of money only if you are trading in the same direction as the majority of large players, including market makers. Market makers are making a lot of profits because they are the ones who are leading the large players directional moves.
Until 2007 it was relatively easy with classic volume analysis at cluster breakouts. Today only the 20 DMF can help
you when trading broad-based indices with the pivot methodology.
With the pivot system, you are in the same risk management setup as the MMs.
Billy

[adam ali]

Fair enough. Would it also be fair to say that we're "seeing" market direction after the fact (analyzing executed volume) whereas HFT MM, thru their algos, try to divine market direction by "testing" for market demand and supply in real-time?

In years past, the guys down on the floor had the book, could see demand and supply, and act accordingly. That sort of look is no longer available - at least not nearly to the same degree - as institutions got smarter about how their traders placed orders.

[Billy]

Fair enough. Would it also be fair to say that we're "seeing" market direction after the fact (analyzing executed volume) whereas HFT MM, thru their algos, try to divine market direction by "testing" for market demand and supply in real-time?

In years past, the guys down on the floor had the book, could see demand and supply, and act accordingly. That sort of look is no longer available - at least not nearly to the same degree - as institutions got smarter about how their traders placed orders.

MMs, and especially the largest ones, have the best big picture about institutional investor’s intentions. They seldom divine and only act on certainty.
Again, you can only really understand this if you've worked in a MM's trading room.
I could give zillions of examples for this, but one will suffice for now. Let's suppose a very large fund wants to start accumulating big positions in many ETFs and stocks. The fund will first call around all MMs houses to ask if they can arrange block trades at a limit price. Let's say $50 for 1,000,000 shares of stock A currently trading at $49.50. If the MM is holding the position in inventory at a lower VWAP he will probably be the counterparty if it fits well with its outsizing plan. If they don't have the stock in inventory, or worst are short the stock, the MMs will instantly call their institutional clients and other MM's to search for a counterparty. If the majority of replies is "No, thank you, could you also arrange to buy a block for me?", you can be sure that the MMs will rush to their buy buttons and will launch their buy programs like mad. And because all MMs have been contacted it quickly turns into snowballing to the upside because they want to accumulate inventories as low as possible before the big flows of orders start coming in.
Billy

[Riskslayer]

Hi Billy,

I love the insider's view of the MM industry. I found the discussion around the Senior MM's focusing on the LT time frames (TF's) while the junior MM's focused on the daily TF's particularly interesting. I had a couple of Q's:

1. Back in the day, how were decisions made and coordinated between Senior MM's and Junior MM's at a firm to short a position nakedly such as to drive down prices and/or "fish for" for stops to takeout the retail holders?? Did individuals make these decisions and take action themselves without direction and coordination with others at the MM? I gather this is all done now via algo programs and HFT trading now, right?

2. How did the MM's make decisions about holding large inventories of a particular position for the LT time frames, monthly, quarterly, etc.? I don't suspect that a large client would leave their orders on the books for months, but perhaps I am wrong about that? In other words, weren't the Senior MM's making directional bets on particular positions without much clear visibility on the large institutional orders that would be needed to support them?

Thanks in advance,

Shawn

[Billy]

Hi Billy,

I love the insider's view of the MM industry. I found the discussion around the Senior MM's focusing on the LT time frames (TF's) while the junior MM's focused on the daily TF's particularly interesting. I had a couple of Q's:

1. Back in the day, how were decisions made and coordinated between Senior MM's and Junior MM's at a firm to short a position nakedly such as to drive down prices and/or "fish for" for stops to takeout the retail holders?? Did individuals make these decisions and take action themselves without direction and coordination with others at the MM? I gather this is all done now via algo programs and HFT trading now, right?

2. How did the MM's make decisions about holding large inventories of a particular position for the LT time frames, monthly, quarterly, etc.? I don't suspect that a large client would leave their orders on the books for months, but perhaps I am wrong about that? In other words, weren't the Senior MM's making directional bets on particular positions without much clear visibility on the large institutional orders that would be needed to support them?

Thanks in advance,

Shawn

Shawn,
1. We had a quick management morning briefing before the market opened. Each junior was allocated his daily quota of shares to sell- naked or not- or buy from/for the house inventory in the securities he was responsible for, usually only one or two maximum per junior. This daily quota was determined by the seniors out of their longer term timeframes position sizing strategies. Another daily quota was given for the accumulation/distribution programs of institutional investors clients. Each junior market maker had some leeway in his decision making for shorting nakedly intraday. But the rules were that a junior had to close the day with zero overnight risk exposure and had to fulfill totally his daily quota. The junior’s bonuses were proportional to their net VWAP performance on the day. Of course today much of this is automated and optimized by algorithms and HFT.

2. Most market makers were and are still a division of a big brokerage or bank. Think Goldman Sachs. They (GS) are managing or advising tens of thousands of funds and institutional investors’ strategies around the world. The priority is of course always given to the GS market making division for the planning and execution over time of these strategies. 90% of the planning goes beyond one month, so it is of strategic importance for the senior market makers to program all the requests beyond one month and to manage their own inventories accordingly for all securities. All the daily intraday gyrations you worry about as a retail investor really mean nothing in the big picture. And it is easy to understand from the above that the ones with the best market direction control and certainty are these senior market makers. That’s why GS has only one losing trading day on average per quarter.
Billy