PDA

View Full Version : Tutorial Intro: Path Of Least Resistance



Billy
05-26-2011, 05:38 AM
Path Of Least Resistance
I am still busy structuring my multi-pivots tutorials. It is a very difficult task because the methodology is mostly practice-based. One could write full books about how to make a golf swing or how to play a Chopin sonata, only hours, days and years of practice could actually make you a better player. Another difficulty for teaching is that most of what could be quantified and optimized, thanks to the initial help of Dr. K and Pascal, will be kept proprietary. The unfortunate experience in the VIT group that led to abusive external commercial use of my work without my consent forces me to do so. The quantified elements of the methodology are now an essential foundation for the risk management system of the robots setups. I will therefore keep the exclusivity of most tutorials for the early birds who will subscribe to the robots in mid-June.
In the meantime, I want to give you some introduction on how the multi-pivots can help you anticipate the path of least resistance day after day.
The formulas for calculating pivots, supports and resistances are as follows:
Resistance 3 = High + 2*(Pivot - Low)
Resistance 2 = Pivot + (R1 - S1)
Resistance 1 = 2 * Pivot - Low
Pivot Point = ( High + Close + Low )/3
Support 1 = 2 * Pivot - High
Support 2 = Pivot - (R1 - S1)
Support 3 = Low - 2*(High - Pivot)

We use six timeframes for the pivots and support/resistance levels:
Daily (D), Weekly (W), Monthly (M), Quarterly (Q), Semester (S) and Yearly (Y).
Additionally, we use two moving averages: the 50-day and the 200-day moving averages.

Each level is given a weight (or strength), proportional to the timeframe it covers, as follows:
Daily: 1
Weekly: 2
Monthly: 3
50 dma: 4
Quarterly: 5
Semester: 6
200 dma: 7
Yearly: 8
The next step is to locate clusters of support and resistance for the next trading day, within a realistic volatility multiple of the last close based on the last ATR% readings. (This is proprietary; please don’t ask for the details!). Once the clusters have been identified, we simply add up the weights of all levels within each cluster.
Below are today’s clusters results for IWM.
Cluster Strength
First Resistance Cluster: 82.606:83.213 11
Second Resistance Cluster: 84.086:85.313 8
First Support Cluster: 81.803:81.076 9
Second Support Cluster: 80.173:78.946 14

Our interest goes primarily to the first resistance and support clusters which are drawn as a white rectangle in the attached chart with their respective cluster resistance strength of 11 and cluster support strength of 9. The path of least resistance for the nearest clusters is slightly to the downside with a resistance-to-support strength of 11-to-9. This is not a big edge in front of a strong intraday money flow, but a neutral to negative intraday money flow will alert us of a potential reversal within the first resistance cluster.
At today’s open, IWM can move freely between the two clusters limits of 82.61 and 81.80 without testing the first significant resistance or support cluster areas. It will be day-neutral as long as it stays there.
Another part of the methodology will use these clusters to determine the optimal 3:1 reward-risk ratio buy and sell entries for the day, independently of market direction or money flow. These are the same as used by the robots. Today’s optimal buy point is 82.14 and the optimal short point is 82.82.
Based on the robot conclusions, I have included the “most probable” 3-day targets for long and short trades. The evidence that should jump at you all at once is that today’s recommended short entry limit at 82.82 is just below the “most probable” 3-day long target at 82.94, near the 50-day moving average (83.16) -so important to institutional investors- and most importantly, still well below the trailing stop (83.44) of the current robot position. In the absence of a very strong intraday 20 DMF, if IWM reaches that area, it will provide an optimal short entry as longs get exhausted within the strong resistance cluster. Only a sustained incoming buy order flow from institutions could prevent market makers from taking advantage of the weak buyers.
Billy

Riskslayer
05-26-2011, 01:47 PM
Hi Billy,

I'm considering a secondary short position on IWM this afternoon as IWM has cleared 82.82, and it looks like its stalling in the zone you mention in your note. Moreover, *my stop* on the trade would be the existing position stop and its pretty tight @ <0.8% (I would stop out at 83.44).

The "Day MF Strength" as of 1:12 pm ET is 109.5% is this very strong intraday 20DMF within the context of your note?

Thanks,

Shawn

Billy
05-26-2011, 02:07 PM
Hi Billy,

I'm considering a secondary short position on IWM this afternoon as IWM has cleared 82.82, and it looks like its stalling in the zone you mention in your note. Moreover, *my stop* on the trade would be the existing position stop and its pretty tight @ <0.8% (I would stop out at 83.44).

The "Day MF Strength" as of 1:12 pm ET is 109.5% is this very strong intraday 20DMF within the context of your note?

Thanks,

Shawn

Shawn,

Our observation is that the timeframe that really matters intraday for the 20 DMF is the last 30 minutes of trading. It can turn the day MF on a dime. So at this stage you must make a bet that the long exhaustion scenario will develop until the close and that buy programs will be switched off. The price of the bet is the stop you choose.
For full disclosure, I already made the same bet as you entering TZA at 35.66 with a tight stop at 34.87. Be careful at the open tomorrow, Markettells just issued statistics for the day prior to Memorial Day Holiday with a very high percentage of historical occurences closing in the direction of the open. Volume will be low and likely unable to make a big inflection in price after the open. So, respect your stop if it is hit and we open higher tomorrow.
Billy

davidallison@gmail.com
05-30-2011, 11:54 AM
Billy,

Thank you so much for taking the time and posting the tutorial “intro for the path of least resistance”. I’m very impressed with what appears to be optimal entries points for robot trades. I look forward to all of your posts as they really help me keep things in perspective. Please keep them coming!!

I have a question regarding after entering into a successful primary robot trade. One can never be faulted for taking money off the table along the way. It’s a personal preference, right? The recent trade on TZA, which was stopped out at close to a nice 5% profit, showed unrealized profits of over double that along the way. I believe this was true of a previous robot trade. It appears to me much as there are optimal entry points; there are also optimal exit points, to take at least some profits off the table. I understand the larger picture is to remain in the trade and the market could have just as easily roller over and we would be deep in the money at this point. I think you have made reference to intermediate exit points along the way giving probabilities, but I didn’t quite understand. Given how the robot seems so accomplished at picking an entry point, would it not very good at picking an intermediate exit point? For those of us that perhaps would like to reduce their position size and take some profits off the table. Maybe this clear to everyone given the LT ST information, but not to me.

Perhaps its old age, but I seem to be getting lost in the probabilities of all this. If you could just explain the edge in terms of probabilities for Tuesdays trade, that might help me understand. It’s a primary robot trade, but the short settings are neutral. The short settings are neutral at last close of 83.65 which would lead to a 3D loss? If the trade is triggered at 85.10, do we know from the past what the 3D gain or loss would be? If IWM rallies hard and approaches 86.86 stop level, do we know anything about the probabilities? Perhaps all of this could be explained in term of: If one had traded X dollars for the last robot primary trade, would you be allocating the same X dollars for this trade?

Thanks Billy, I hope I’m not exasperating you with my questions!!!

Dave

Billy
05-30-2011, 02:41 PM
Dave,

Q1: “I have a question regarding after entering into a successful primary robot trade. One can never be faulted for taking money off the table along the way. It’s a personal preference, right?”

A1: Absolutely. Statistically, scaling in on new signals is less optimal than scaling out. It’s a hard decision to make, especially by how much, 25%, 33%, 50%? And you may decide to re-enter the sold portion on new daily entry price limits. If I can understand the psychological comfort that such decisions can bring, you are bound for some good instant gratifications trades in a mean-reversion environment (40% of the time) and for many frustrating opportunity losses in a trending environment (60% of the time).

Q2: “The recent trade on TZA, which was stopped out at close to a nice 5% profit, showed unrealized profits of over double that along the way. I believe this was true of a previous robot trade. It appears to me much as there are optimal entry points; there are also optimal exit points, to take at least some profits off the table”.

A2: The last 2 trades were victims of mean-reversion and have been stopped out. Being stopped out, even with profits on a trailing stop, is the least optimal of all exits. But the one and only “optimal” exit point for the robot is to exit on a 20 DMF signal change after a lasting and trending trade. These are the ones we try to catch and ride fully invested because that’s where the bulk of the risk-adjusted performance is coming from. You may use robot statistics to evaluate “good” or “satisfactory” scaling-out levels, but they are far from “optimal”.

Q3: “I understand the larger picture is to remain in the trade and the market could have just as easily rolled over and we would be deep in the money at this point.”

A3: That’s exactly my thesis.

Q4: “I think you have made reference to intermediate exit points along the way giving probabilities, but I didn’t quite understand. Given how the robot seems so accomplished at picking an entry point, would it not be very good at picking an intermediate exit point? For those of us that perhaps would like to reduce their position size and take some profits off the table. Maybe this is clear to everyone given the LT ST information, but not to me.”

A4: I never made reference to intermediate exit points, but to “most probable 3-day targets” for monitoring the evolution of a trade. Such targets can be discretionarily used as exit points, but they are not exit targets in my view. Quite to the contrary, hitting these targets within 3 days tells you that the trade is evolving perfectly as expected. If probabilities remain positive from that point, you certainly don’t want to exit as a big trend-following run may be just around the corner.
The LT/ST information on the robot page is always based on projections from the last CLOSING price. For Tuesday, it says: In the past this combination led to a 3D short LOSS of -0.11% from the previous day's close. The trade became positive after three days in 53.1% of the cases. If you were in a winning trade with such statistics, would you really have an advantage taking some money off the table to avoid a 3-day loss of -0.11% while having more than 50% chance of increasing your profit in the same 3-day period?

Q5: “Perhaps its old age, but I seem to be getting lost in the probabilities of all this. If you could just explain the edge in terms of probabilities for Tuesdays trade, that might help me understand. It’s a primary robot trade, but the short settings are neutral. The short settings are neutral at last close of 83.65 which would lead to a 3D loss? If the trade is triggered at 85.10, do we know from the past what the 3D gain or loss would be? If IWM rallies hard and approaches 86.86 stop level, do we know anything about the probabilities? Perhaps all of this could be explained in term of: If one had traded X dollars for the last robot primary trade, would you be allocating the same X dollars for this trade?”

A5: You must absolutely distinguish between the ST/LT statistics from the previous close (83.65) and the expected gain and probabilities if entering a new position at the limit entry price (85.10).
For Tuesday, the “most probable” 3-day target from the CLOSE is a rise of 0.11% with only 46.9% chances (because there are 53.1% odds that the price will be lower) or :
83.65 + (83.65* 0.11% * 46.9%) = 83.693. This very neutral indeed.
But the “most probable” 3-day target if you have the opportunity to enter at the limit entry price of 85.10 is:
85.10 – (85.10 * 2.64%* 61.67%) = 83.71
This is pretty close to the previous target of 83.69. The main difference for the robot’s trading account is that it will not miss shorting a bounce to 85.10 in-between. If IWM never trades at or above 85.10 during open market hours, the robot will simply wait and stay in cash.

Here are the clusters and their strengths for Tuesday. Notice that the new monthly pivots starting on Wednesday will somehow improve the support cluster strengths if Tuesday’s close is higher or unchanged from Friday’s close.
8531
8533

Pierre Brodeur
05-30-2011, 04:27 PM
Dave,

Notice that the new monthly pivots starting on Wednesday will somehow improve the support cluster strengths if Tuesday’s close is higher or unchanged from Friday’s close.
8531
8533

Two quick questions please:
-1- Is the MS1 estimate currently calculated in the Support strenght of 13 or will it add to this number Wednesday.
-2- Does the fact that Res strenght = almost double support strenght has any impact on the robot wanting to short right now?

Billy
05-30-2011, 04:52 PM
Two quick questions please:
-1- Is the MS1 estimate currently calculated in the Support strenght of 13 or will it add to this number Wednesday.
-2- Does the fact that Res strenght = almost double support strenght has any impact on the robot wanting to short right now?

Pierre,

1. MS1 is not an estimate and is the actual first monthly support for May. It is part of the support cluster for Tuesday only and its weight is included in the cluster strength of 13.

2. No, the robot does not make its decisions based on cluster strengths. They are used for multi-pivot stand-alone decisions. But it is better to see them confirm the robot probabilities.

FWIW, here are the expected cluster strengths for Wednesday if IWM closes unchanged on Tuesday.
The new (June) monthly levels will make a big difference in favor of the bulls. Billy
8537

davidallison@gmail.com
05-30-2011, 11:47 PM
Billy,

Thank you for the great responce. This really helps with my understanding.

You mentioned:

But the “most probable” 3-day target if you have the opportunity to enter at the limit entry price of 85.10 is:
85.10 – (85.10 * 2.64%* 61.67%) = 83.71

Once again, it's my age! :) but where does the 2.64% and 61.67% come from?

Dave

Billy
05-31-2011, 01:50 AM
Billy,

Thank you for the great responce. This really helps with my understanding.

You mentioned:

But the “most probable” 3-day target if you have the opportunity to enter at the limit entry price of 85.10 is:
85.10 – (85.10 * 2.64%* 61.67%) = 83.71

Once again, it's my age! :) but where does the 2.64% and 61.67% come from?

Dave

These statistics are the ones from the new entry graphs on the robot page:

8538

asomani
05-31-2011, 02:59 AM
Billy, given where the S&P futures are currently trading, and where the 20DMF currently stands, is there a high chance the IWM Robot changes its bias to either neutral or long at tomorrow's close?

Would you take the current recommended entry like any other, or be more cautious and reduce position size - or avoid a trade altogether?

Thanks.

Billy
05-31-2011, 05:33 AM
Billy, given where the S&P futures are currently trading, and where the 20DMF currently stands, is there a high chance the IWM Robot changes its bias to either neutral or long at tomorrow's close?

Would you take the current recommended entry like any other, or be more cautious and reduce position size - or avoid a trade altogether?

Thanks.

Aly,

If you've read my last alphascanner blog update, you already know my discretionary feeling.
I think that the robot short entry at a minimum of 85.10 could still pay well on a 3-day horizon. But the signal is weak and we could indeed see the robot looking soon to enter a long position!
This could turn into a major follow-through upside day and you may want to see confirmation of a reversal before shorting. If one can afford to follow the market real-time and full-time, one can wait to observe how today's furious gap up will develop during the day if ever 85.10 is hit.
My inner bias is to stay in cash and planning for the nexr best long entry from the robot, but the robot has always beaten me up for ignoring its guidance! I sincerely hope that 85.10 will not be hit as it would satisfy most everybody to stay in cash ahead of a possible long entry from the robot.
Individual psychology and control of emotions are in play here, so anyone vulnerable or hesitant should better stick strictly to the robot guidance.
Billy

Pascal
05-31-2011, 10:43 AM
As the 20DMF is detecting some weak selling in the first hour of trading and since IWM has reversed from its resistance cluster, I believe that the robot will not reach its ideal short position.

So, we'll probably be in cash for the close of today. Without a bounce to the shorting zone, it will take a break down through the 50MA for the robot to find a new ideal set-up.



Pascal

Pierre Brodeur
05-31-2011, 10:51 AM
The new (June) monthly levels will make a big difference in favor of the bulls.


... because the weight of the support cluster is now greater than the weight of the resistance cluster!

Correct or Incorrect?

Billy
05-31-2011, 11:08 AM
... because the weight of the support cluster is now greater than the weight of the resistance cluster!

Correct or Incorrect?

Pierre, you are correct. But clusters strengths are only valid for one day at a time. The real major resistance remains Yearly R1 (85.68) and Semester R1 (85.46) and the most likely area for a new top. Hence, an entry around 85.10 is not so risky as it seems at first sight.
Billy

Pierre Brodeur
05-31-2011, 03:04 PM
Hence, an entry around 85.10 is not so risky as it seems at first sight.


You actually got my attention with this comment and for a good 10 minutes I have been staring at the chart you provided trying to coldly (froidement) evaluate the risk of that particular trade visually. It allowed me to appreciate the beauty of the cluster approach and make the hypothesis that the weights that you cleverly devised are really probabilities that the individual resistances will not be penetrated within a trader's investment horizon which I presume is relatively short.

When one focusses on the specified entry price, it seems to me that this number is probably calculated using a weighted sum of the resistance prices and weights (1,2 ... or 8) for the first resistance cluster as a whole or in other words penetrating WR1(2) and MPP(3) is more likely than SR1(6) and YR1(8) and therefore the ideal entry point must be closer to the weakest part of the cluster.

Whether this is true or not, the exercise has been fruitful for me as this adds another tool in my trading kit. Thank you Billy (and Pascal for the Stop part I presume) for bringing such a powerful tool to this community.

Billy
05-31-2011, 03:46 PM
You actually got my attention with this comment and for a good 10 minutes I have been staring at the chart you provided trying to coldly (froidement) evaluate the risk of that particular trade visually. It allowed me to appreciate the beauty of the cluster approach and make the hypothesis that the weights that you cleverly devised are really probabilities that the individual resistances will not be penetrated within a trader's investment horizon which I presume is relatively short.

When one focusses on the specified entry price, it seems to me that this number is probably calculated using a weighted sum of the resistance prices and weights (1,2 ... or 8) for the first resistance cluster as a whole or in other words penetrating WR1(2) and MPP(3) is more likely than SR1(6) and YR1(8) and therefore the ideal entry point must be closer to the weakest part of the cluster.

Whether this is true or not, the exercise has been fruitful for me as this adds another tool in my trading kit. Thank you Billy (and Pascal for the Stop part I presume) for bringing such a powerful tool to this community.

Pierre, thank you for the feedback.
In fact, the cluster parameters have been optimized for IWM and the daily buy and short entry prices are always providing a reward-risk ratio of 3:1 if the market trades sideways in the very short term (1 or 2 days) between the extreme limits of the first support and resistance clusters. Before the robot, I took my best market direction guess for choosing to go long or short at the buy/short entry prices. The added value from the robot is that it forces you to stay in synch with the 20 DMF and the highest probabilities beyond the initial 3:1 reward-risk ratio entry edge.
Billy

johnboy70
06-01-2011, 08:03 AM
Path Of Least Resistance
I am still busy structuring my multi-pivots tutorials. It is a very difficult task because the methodology is mostly practice-based. One could write full books about how to make a golf swing or how to play a Chopin sonata, only hours, days and years of practice could actually make you a better player. Another difficulty for teaching is that most of what could be quantified and optimized, thanks to the initial help of Dr. K and Pascal, will be kept proprietary. The unfortunate experience in the VIT group that led to abusive external commercial use of my work without my consent forces me to do so. The quantified elements of the methodology are now an essential foundation for the risk management system of the robots setups. I will therefore keep the exclusivity of most tutorials for the early birds who will subscribe to the robots in mid-June.
In the meantime, I want to give you some introduction on how the multi-pivots can help you anticipate the path of least resistance day after day.
The formulas for calculating pivots, supports and resistances are as follows:
Resistance 3 = High + 2*(Pivot - Low)
Resistance 2 = Pivot + (R1 - S1)
Resistance 1 = 2 * Pivot - Low
Pivot Point = ( High + Close + Low )/3
Support 1 = 2 * Pivot - High
Support 2 = Pivot - (R1 - S1)
Support 3 = Low - 2*(High - Pivot)

We use six timeframes for the pivots and support/resistance levels:
Daily (D), Weekly (W), Monthly (M), Quarterly (Q), Semester (S) and Yearly (Y).
Additionally, we use two moving averages: the 50-day and the 200-day moving averages.

Each level is given a weight (or strength), proportional to the timeframe it covers, as follows:
Daily: 1
Weekly: 2
Monthly: 3
50 dma: 4
Quarterly: 5
Semester: 6
200 dma: 7
Yearly: 8
The next step is to locate clusters of support and resistance for the next trading day, within a realistic volatility multiple of the last close based on the last ATR% readings. (This is proprietary; please don’t ask for the details!). Once the clusters have been identified, we simply add up the weights of all levels within each cluster.
Below are today’s clusters results for IWM.
Cluster Strength
First Resistance Cluster: 82.606:83.213 11
Second Resistance Cluster: 84.086:85.313 8
First Support Cluster: 81.803:81.076 9
Second Support Cluster: 80.173:78.946 14

Our interest goes primarily to the first resistance and support clusters which are drawn as a white rectangle in the attached chart with their respective cluster resistance strength of 11 and cluster support strength of 9. The path of least resistance for the nearest clusters is slightly to the downside with a resistance-to-support strength of 11-to-9. This is not a big edge in front of a strong intraday money flow, but a neutral to negative intraday money flow will alert us of a potential reversal within the first resistance cluster.
At today’s open, IWM can move freely between the two clusters limits of 82.61 and 81.80 without testing the first significant resistance or support cluster areas. It will be day-neutral as long as it stays there.
Another part of the methodology will use these clusters to determine the optimal 3:1 reward-risk ratio buy and sell entries for the day, independently of market direction or money flow. These are the same as used by the robots. Today’s optimal buy point is 82.14 and the optimal short point is 82.82.
Based on the robot conclusions, I have included the “most probable” 3-day targets for long and short trades. The evidence that should jump at you all at once is that today’s recommended short entry limit at 82.82 is just below the “most probable” 3-day long target at 82.94, near the 50-day moving average (83.16) -so important to institutional investors- and most importantly, still well below the trailing stop (83.44) of the current robot position. In the absence of a very strong intraday 20 DMF, if IWM reaches that area, it will provide an optimal short entry as longs get exhausted within the strong resistance cluster. Only a sustained incoming buy order flow from institutions could prevent market makers from taking advantage of the weak buyers.
Billy

Billy,

I've read this entire thread twice over the past several days (since you wrote it), to include all the comments, and I'm wondering, without disclosing how you calculate it (which you say is proprietary - and, I'm not asking for disclosure of any type of your techniques), is there any way you could possibly define the term "cluster", other than "bands" where one could expect the highest volume of trading activity to normally occur? Also, once you define them, can I presume that they must be redefined every day (which is what I presume you have to do)?

John

Billy
06-01-2011, 08:31 AM
Billy,

I've read this entire thread twice over the past several days (since you wrote it), to include all the comments, and I'm wondering, without disclosing how you calculate it (which you say is proprietary - and, I'm not asking for disclosure of any type of your techniques), is there any way you could possibly define the term "cluster", other than "bands" where one could expect the highest volume of trading activity to normally occur? Also, once you define them, can I presume that they must be redefined every day (which is what I presume you have to do)?

John

John,
We define a cluster of pivot levels as those levels that are close together. "Close together" is defined by all pivots within the Average True Range (ATR). The original basic cluster of the methodology is therefore the ATR. For the robots, we have optimized both the ATR and some multiple of it for the best correlation with the underlying ETF’s normal behavior.
This allows for measuring precisely the support and resistance pressures for the next trading day only within a “normal” daily trading range. So, yes, they must be redefined daily. A breakout above a stronger cluster resistance than support is the proof that large players are showing conviction in their accumulation and that market makers/ HFT are forced to follow their steps. The higher the volume and money flow when it happens, the better.
Billy