3
5

Does anyone have a scan for the "Pocket Pivot" that they would be willing to share?

asked Aug 22 '11 at 00:07

TicoBob's gravatar image

TicoBob
50 112

edited Nov 22 '11 at 03:34

markk's gravatar image

markk
3.2k 711


Note: Edited to incorporate Gord's improvement below.

This is an interesting one for sure.

The definition I used I pulled from here: Pocket Pivot

I guess this originally came from the book: Trade Like An O'Neil Disciple

The Stockcharts code for this would look like this


[close > yesterday's close] and

[yesterday's close < 2 days ago close] and

[2 days ago close < 3 days ago close] and

[3 days ago close < 4 days ago close] and

[4 days ago close < 5 days ago close] and

[5 days ago close < 6 days ago close] and

[6 days ago close < 7 days ago close] and

[7 days ago close < 8 days ago close] and

[8 days ago close < 9 days ago close] and

[9 days ago close < 10 days ago close] and

[ volume > yesterdays Max (10, volume)]


You can copy and paste the code below if need be.

[close > yesterday's close] and [yesterday's close < 2 days ago close] and [2 days ago close < 3 days ago close] and [3 days ago close < 4 days ago close] and [4 days ago close < 5 days ago close] and [5 days ago close < 6 days ago close] and [6 days ago close < 7 days ago close] and [7 days ago close < 8 days ago close] and [8 days ago close < 9 days ago close] and [9 days ago close < 10 days ago close] and [ volume > yesterdays Max (10, volume)]

This didn't return any results for me, but I tried it with shorter cycle (5 days) and it worked so I am pretty sure the code is good. I am guessing this scan is market condition specific. Looking at this makes me think "don't shoot the messenger"... I get feeling there is reason you have described this scan as infamous. Be careful out there. Good luck.

link

answered Aug 22 '11 at 19:30

AgnosticTrader's gravatar image

AgnosticTrader
8.7k 2636

edited Aug 22 '11 at 23:21

Agnostic Trader, great post and the coding looks fine.

Can't really simplify the close stuff, but you could simplify the volume section using the Min / Max function which is one of my favorites.

IE Volume today is > Maximum volume in any of the 10 previous days

and [ volume > yesterdays Max (10, volume)]

(Aug 22 '11 at 22:06) Gord Gord's gravatar image
1

No matter what the market is doing there won't be that many that meet the criteria from my experience...Doug

(Aug 23 '11 at 09:02) Sharptraders Sharptraders's gravatar image
2

My belated thanks for the prompt and thorough answers to my question. You guys are great to be associated with

TicoBob

(Aug 27 '11 at 10:37) TicoBob TicoBob's gravatar image

The definition from mypivots.com really doesn't match the definition used by Chris Kacher (who came up with the concept). See his book (cited by A-T) or his website: http://www.virtueofselfishinvesting.com/faqs/answer/Ten-Rules-for-Pocket-Pivots .

For one thing, the volume restriction looks at the largest from a down day:

"3.The day's volume should be larger than the highest down volume day over the prior 10 days."

Also, you're not looking for a series of lower closes, but penetration through the 10 or 50 d SMA.

There are some fundamentals and situational criteria that also need to be addressed.

This code will narrow the possibilities. It gets tedious to meet their Rule 3.

// Criteria to only scan leading stocks

[type = stock] and

[daily sma(40,daily volume) > 500000] and

[daily sma(40,daily close) > 10] and

[daily close > daily sma(200,daily close)] and

// Criteria to verify the day's volume should be larger than the highest down volume day over the prior 10 days

[daily volume > 1 days ago daily volume * 1 days ago ROC(1) /absval(1 days ago ROC(1))* -1] and

[daily volume > 2 days ago daily volume * 2 days ago ROC(1) /absval(2 days ago ROC(1))* -1] and

[daily volume > 3 days ago daily volume * 3 days ago ROC(1) /absval(3 days ago ROC(1))* -1] and

[daily volume > 4 days ago daily volume * 4 days ago ROC(1) /absval(4 days ago ROC(1))* -1] and

[daily volume > 5 days ago daily volume * 5 days ago ROC(1) /absval(5 days ago ROC(1))* -1] and

[daily volume > 6 days ago daily volume * 6 days ago ROC(1) /absval(6 days ago ROC(1))* -1] and

[daily volume > 7 days ago daily volume * 7 days ago ROC(1) /absval(7 days ago ROC(1))* -1] and

[daily volume > 8 days ago daily volume * 8 days ago ROC(1) /absval(8 days ago ROC(1))* -1] and

[daily volume > 9 days ago daily volume * 9 days ago ROC(1) /absval(9 days ago ROC(1))* -1] and

[daily volume > 10 days ago daily volume * 10 days ago ROC(1) /absval(10 days ago ROC(1))* -1] and

//Criteria to see jump through 10 or 50 d SMA

[[yesterday's daily close < yesterday's daily sma(10,daily close)] and [daily close > daily sma(10,daily close)]

or

[yesterday's daily close < yesterday's daily sma(50,daily close)] and [daily close > daily sma(50,daily close)]]

Once you've run this scan, then you need to visually inspect the candidates. You're looking for either an early buy point within a consolidation, or a consolidation buy point for a leading stock already firmly entrenched in a strong uptrend. (This is very easy to do visually, but would be very tough to program into the scan.) The scan found 6 symbols tonight, but none had charts that visually matched these criteria. Not surprising, since you don't expect to find pocket pivots very often.

link

answered Nov 19 '11 at 00:46

markk's gravatar image

markk
3.2k 711

edited Nov 25 '11 at 20:30

1

Hi markk, this is very interesting. I didn't know you could test for down day volume only. I'm not a math wiz (plus I missed a lot of high school). I wonder if you could explain how the ROC(1)lines are different from "volume > max(10, volume)". Thanks!

(Nov 22 '11 at 17:55) markd markd's gravatar image
3

If you look at the far right of the stack of ten comparisons, I multiply by "-1". The daily volume should always be greater than a negative number, so the comparison should always be true unless you do something. I then use the one day ROC divided by its absolute value to pick off its sign. It should be negative on a down day. -1 x -something is a positive number, so the daily volume is compared to a positive number. That way you can isolate the comparison to down days only. This is tedious, but it's the only way I could figure out how to honor their Rule 3.

(Nov 22 '11 at 19:15) markk markk's gravatar image
1

That's very clever. Thank you!

(Nov 22 '11 at 22:06) markd markd's gravatar image
1

Hi Markk, brilliant use of the scan parameters and functions to isolate the up and down volume days, I never would have thought of that one. thanks, one more trick for the toolbox.

(Nov 30 '11 at 13:35) Gord Gord's gravatar image

Any chance the "checkmark" can be removed from the answer that used the definition from mypivots.com?

That really isn't a correct answer.

link

answered Nov 19 '11 at 00:50

markk's gravatar image

markk
3.2k 711

edited Nov 29 '11 at 16:18

I wasn't familiar with pocket pivot before this, but I visited the site and read the pdf presentations with chart examples. It looks like no one scan will capture all PP instances, since they vary quite a bit. However, these seem to be the most common "scan-able" characteristics of the least risky and better performing examples:

a. in most cases the long term MA (say 200 day) is rising

b. in most cases the intermediate term (50 day) MA is above the 200 MA

c. in most cases the close is above the 200 MA

d. at some point prior to the pivot day, there has been a low below the short term (10 day) MA.

e. the range of the pivot day is large, it closes near the high and above the 10 MA

f. volume is huge relative to recent volume

So, although these scan lines do not exactly fit published criteria, they should capture valid examples (and others you may have to toss after visual inspection of the results).

// begin scan

// scope (edit per your requirements)

[group is sp600] and

// a. LT MA direction (higher than say a month ago)

[sma(200, close) > 20 days ago sma(200, close)] and

// b. MA relative positions (intermediate above long term)

[sma(50, close) > sma(200, close)] and

// c. close above 200 MA

[close > sma(200, close)] and

// d. recent price action - low below short term MA

[min(10, low) < min(10, sma(10, close))] and

// e. pivot day test 1 - wide range - note "range" probably doesn't account for gaps // but wide hi-lo range looks like a good filter with or without a gap

[range > ATR(14)] and

// e. pivot day test 2 - close near high (in top third, adjust .66 to your preference)

[close > low +[range * .66]] and

// e. pivot day test 3 - close above 10 MA

[close > sma(10, close)] and

// f. pivot day test 4 - high volume (choose your own MA length and multiplier)

[volume > sma(10, volume) * 3]

//end scan

Remarkably, this scan produces one hit today 11/21/2011, REGN. I also tried it against a couple of random dates in the past. There were hits, but the follow through price action wasn't great. To me, that means you should run your version of this scan against a favorites list of pre-selected stocks that meet the fundamental criteria proposed on the "virtues" site.

link

answered Nov 21 '11 at 18:27

markd's gravatar image

markd
14.5k 1523

1

Your volume criteria doesn't follow their Rule 3.

(Nov 22 '11 at 04:55) markk markk's gravatar image
1

You're right. Thanks for explaining your vote. Probably I should have been more clear why I did that. It was a judgement call whether to include the down volume test. My thought was, the best opportunities will be straightforward. The pivot day will have clearly higher volume than recent days. So the down volume test really only applies to marginal situations, which personally I would avoid. I thought of adding a max(x, volume)test for the pivot day, but prior up days with higher volume should be OK. A bigger volume down day will be apparent on the results chart, so you would toss it then.

(Nov 22 '11 at 17:28) markd markd's gravatar image
Your answer
toggle preview

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here

By RSS:

Answers

Answers and Comments

Markdown Basics

  • *italic* or __italic__
  • **bold** or __bold__
  • link:[text](http://url.com/ "title")
  • image?![alt text](/path/img.jpg "title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported

Tags:

×212
×2

Asked: Aug 22 '11 at 00:07

Seen: 6,056 times

Last updated: Nov 30 '11 at 13:35