Search | User Listing Vidsite | Calendars | Quotes
Home Page ->  Custom TradeTight Routines -> Tools In Team Development -> View ThreadLogon (or Register, or Join TradeTight)

You are logged in as a limited-access Guest.To join TradeTight, first read the info in the Organization & Content room, then click the link above. 

Psalm 127:1 (NKJV) ... Unless the Lord builds the house, they labor in vain who build it; Unless the Lord guards the city, the watchman stays awake in vain.


Sticky Cumulative Volume Indicator Distribution
Frozen
Jump to page : 1
Now viewing page 1 [50 msgs/pg]
Jump to room :
JimDean
Posted 1/3/2019 8:52 AM (#9175)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
This thread is set up to provide distribution files and docs during development of VOLeval (and possible later VOLpowr, VOLflex, VOLedge versions).

Use the separate thread for Q/A and discussion, here: http://tradetight.org/forums/thread-view.asp?tid=1432
Top of the page Bottom of the page
JimDean
Posted 1/3/2019 8:55 AM (#9176 - in reply to #9175)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
Here are the files for the first VOLeval beta release (Keys files are sent to you individually).

With OT off, copy the TradeTight-Keys.txt file into OT18's SDK\TT_Support-Files folder, then extract the VOLeval DLL from the zip into the SDK root folder, and UNBLOCK it (via right-click). If you don't already have it installed, copy the aJDctBasicBlack.otd file from the Zip into the OT18\ChartTemplates folder. If you haven't already, you'll need to update your OT2018 to pre-release 3A.

Start OT, open the Basic Black template, SaveAs aFLctVOLeval (or whatever name you want), then right click, Add Indicators, and select VOLeval from the "V" listing category. Voila! Your test template is set.

HERE'S WHAT I NEED YOU TO TEST: I modified this slightly after A2 was released

1. Always start with the "defaults" input pattern ... open the panel via double-click on the plotted indicator, and click on "Restore Defaults" (bottom left) to quickly get back to this point. You'll need at least 600 bars loaded, to provide for 126 output plotted bars and the max possible number of warmup bars.

2. Not much point in detail-testing the Return values, since the return-options will most likely be radically rearranged (per ToDo). I suggest that you create one test FL column, using Default settings, that has the last three inputs "0,10,-1". The upcoming revisions will use entirely different return-inputs to extract digits, etc, so there's no need to test or memorize the current approach. you can skip this step entirely

3. Lock in SpikeCap = 2, LastBar=0, BarsToCalc=126, MixWtd input=3, and MAvgSlope input=2 no, use MAvgSlope=3 for version A2. The calc's and impacts of those features are already fully tested and their impacts are predictable and not terribly significant. Don't change any of these values during the testing described below.

4. For now, leave the CumSlidWndo input=2. Later we will test it but since it has such a large and unpredictable impact, we will study the other things first. Step 8 is important

5. Using the Defaults settings, flip through a bunch of Nas-100 daily-bar symbols, and find a few (no more than 5) that have a fair sampling of tradeable Long, tradeable Short, and untradeable sections within the most recent 6 months (ie during the BarsToCalc=126 window). Each of you has chosen three ... all different from the others

6. General test process:
a. Select a CumMeth(VolObOg) input (4, then 6, then 2, then 8) and hold it constant for steps b-f
b. Select a MApdsFMS input (2, then 1, then 3) and hold it constant for steps c-f
c. Select a FuzzyEquiv input (2, then 3, then 4, then 1, then 0) and hold it constant for steps d-f
d. Select a TrdZonBgn input (6, then 7, then 5) and hold it constant for step e-f
e. Vary the TrdZonEnd input (2, then 1, then 3) and hold it constant for step f
f. Check out your selected Symbols, then and RECORD OBSERVATIONS for that a-f combo on the SS

7. From your observations in 6.a-e, write out a list of (at least) ten different input-patterns that you feel are *the* most useful. That is, each pattern will specify five numbers ... one from each of the a-e choices. Be sure to send that list to me. If you provide more than ten, great (using a spreadsheet to keep track during tests might be smart) ... but please try to "sort" the list (somehow) if >10 choices. Big change here ... the SS provides auto totals for the scores, and it can be sorted on that column ... after sorting you'll likely see an obvious group of high scores ... maybe 50 or 100 who knows (based on your avail time). Use that group, not just ten, and not seat of pants

8. Last step: for each of the top-ten input-pattern selections (lock in one at a time), vary the CumSlidWndo input (1, 2, 3, 4), checking your selected Symbols for each, and "grade" the results ... that is, whatever, metrics or holistic gut feel you used to determine the ten-best in step 7, use that same means to decide for each pattern which are the *two best and two worst* CumSlidWndo selections (not counting 0).
Note: this step is IMPORTANT, but requires the others to be completed first to make its results meaningful. The CumSlidWndo input has a big impact, but its current 1,2,3,4 settings are somewhat arbitrary ... I will use your feedback to change the window-sizes related to 1,2,3,4 inputs, to home in on the "ideal".Mark, you might want to create another SS just for this ... or an extra tab on that one which auto-extracts the highest-scoring patterns from the main sheet, and presents them for re-voting with the SlidWndo variants, re Lag Chop Dirn

GOAL: to figure out which settings are most/least useful, and to eliminate or tune the inputs in line with that. If I can kill off some option (such as one of the CumMeth's), due to all of you independently reporting negatively about it, that's great!

Please do NOT share your testing conclusions with the other guys, for now. I'd rather have your observations uncompromised by others' viewpoints. I'll share the results with all of you ... and if questions occur during testing, I'll pass on the Q & answer to everyone.

Please FIRST answer the ToDo email, THEN start the testing (so I can work on ToDo's while you're testing). Please plan to complete all the steps within a week ... ie by the end of Wednesday Jan 9th, at the latest. Earlier feedback will be even more helpful.

Note that this testing is what the $100 discount is for ... if you don't want to do the testing, then just send me an additional $100 for list price, and we'll be square (see above re Flex, etc later on). Alternatively, if you decide you don't want this tool at all, let me know and I'll either refund your money or keep it "on account" for other later work (up to you, tho of course I'd prefer the latter ;~). In order to allow you this flexibility, I'm setting the initial license (in the Keys file) to expire at the end of January ... if after that time you still want to keep it functional, I'll just email you a new Keys file that is for the lifetime use.

Thanks again!

LATER: extensive discussion and info about this was done on the DISCUSSION thread. Go there to get test spreadsheets and to review the scoring categories (Lag, Chop, Dirn)
Top of the page Bottom of the page
JimDean
Posted 1/3/2019 9:14 AM (#9177 - in reply to #9176)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
Here are THIRTY symbols that I've pre-vetted ... please select your test Symbols from the ones in this list. My gut-selection criteria was that the symbol appeared to be roughly swing or trend-tradeable during the past 6 months, and that the collection represents a big variety of patterns.

AAPL
ADP
ALXN
ATVI
CTAS
CTRP
DISCA
DLTR
EXPE
HAS
HOLX
HSIC
LILAK
MAR
MAT
MNST
MU
NCLH
NFLX
NVDA
PAYX
PCAR
QCOM
ROST
TSCO
ULTA
VIAB
VRTX
WBA
WDC
Top of the page Bottom of the page
JimDean
Posted 1/3/2019 9:40 AM (#9179 - in reply to #9177)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
Here are SIX symbols selected from the 30 prior ones ... the "pick of the litter". If you don't want to select your own from the 30, then these would be what I recommend.

Note: later this year, I'll be offering a "TradeTrak" tool and some other Scans that will greatly simplify and ideally select Symbols to use in Scans or Filters, which are particularly amenable to trading with a particular entry/exit strategy. The symbols below are ones that those tools would be likely to select automatically ...

AAPL
ATVI
HAS
MAT
PAYX
QCOM

... and of those, these are the "top 3" in my opinion ... good for testing and L+S swing/trend trading ...

HAS
PAYX
QCOM
Top of the page Bottom of the page
JimDean
Posted 1/5/2019 3:55 AM (#9203 - in reply to #9179)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
Hi, beta guys!

Hmm ... lessee now ... it's 3:15am. BUT - everything is done that I planned, including some extra polish of the Help Panels. The new DLL is compiled, and as far as I know, everything works as intended. The only "note" about upcoming changes is a comment at the bottom of Help Panel #1 saying that more info will appear when the Gurus are implemented. And on that front ... Salim and Lou have both told me that they see a strong positive benefit, for either their use or for marketing (or both), in creating at least two and maybe three "siblings" of VOLeval, when the Gurus are added ... VOLpowr, VOLflex, VOLedge ... or VOLxprt, VOLcore. I'll be working on that while y'all are testing. Please remember that the deadline is next Wed ... but hopefully you'll be done by Monday :~)

Your existing license Keys file remains the same for this release.

All you need to do is to:
0. if you created any chart templates or FL columns etc using the A1 version, open OT, delete them, and close OT
1. download the *two* Zip files from this post,
2. extract the "A2" DLL from the "A2 DLL (only)" zip, into the SDK folder,
3. UNBLOCK it,
4. delete the earlier "A1" version from the SDK (absolutely required!)
5. start OT18pr3A, and add the VOLeval to the Basic Black template for your testing
6. set your data periods to 600 and you'll have more than enough bars regardless of the param settings.

ALSO, I attached *another* Zip, with fully-updated snapshots ... and I created a handy 7-page PDF with all the snaps on it. CAUTION: the first 5 pages are readily printable, and I suggest you do that. The 6th and 7th pages are big black & colored charts ... most printers don't like using that much black ink on a page so I suggest you just look at those on the monitor.

OK ... that's it ... the full package including documentation. :D

Attachments
Attachments VOLeval A2 Params, Help & Charts.zip (5032KB - 0 downloads)
Attachments VOLeval A2 DLL (only).zip (38KB - 2 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/5/2019 10:19 AM (#9205 - in reply to #9203)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
OK apparently there are download problems. I've deleted the attachment from the prior post, and I'm attempting to attach two other tries to this post.

The "VOLeval A2.zip" simply contains the normal DLL file. Please try to download it.

The "VOLeval A2 as DOC.zip" contains the DLL file that has been renamed by replacing the .DLL extension at its end with .DOC ... after downloading, extract the file then rename and confirm.

I also tried sending each of these in separate emails to all of the addresses that I have for y'all. I already got reject notices for all three, from the gmail accounts.


Attachments
Attachments VOLeval A2.zip (38KB - 6 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/5/2019 10:56 AM (#9208 - in reply to #9205)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
OK.

I was able to download the zipped normal-DLL file, and after UNBLOCKing it, OT accepted it and it plots. NOTE: when I downloaded it (into chrome), good ole' Google gave me a warning message at the bottom of the browser window that the contents might be dangerous ... there was a DISCARD button next to the file in the bottom bar, AND an up-caret ... clicking the up-caret and then selecting "Keep" got rid of the message, and the download completed, and the resulting extracted and unblocked DLL *did* work with OT. (whew ... I hate google)

I was able to download the renamed file, and after renaming it and UNBLOCKing it, OT accepted it and it plots.

I was able to download the zipped renamed file, and after renaming it and UNBLOCKing it, OT accepted it and it plots.


When all is said and done, the UNBLOCKED file should have this name:
TTind_VOLeval.A2.O18pr3A.nsl.dll
... pay attention to where the four periods are ...


So ... the newly generated post just before this does have (three) viable attachments.

MY GUESS IS that the initial problem was caused due to the very very first zip (in the wee hours) actually having ANOTHER zip inside of it, which had the snapshots. So, I've tried one final time to attach the info to the original posting ... this time, the DLL (not renamed) is zipped separately in one attachment, and a second attachment has a zip of all the snaps and pdf. ... That worked, btw ... but with the warning + keep procedure, per the first paragraph.
Top of the page Bottom of the page
JimDean
Posted 1/23/2019 8:27 AM (#9247 - in reply to #9208)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
NEW "MOD 9" SPREADSHEET

This morning I found some oversights in the earlier designs and fixed them ... plus made it a bit smarter. I've deleted the old post from this thread. PLEASE USE THE ATTACHED TEMPLATE, not any earlier ones.

The snapshot has notes that tell you how to enter the Symbol Names that you are using (in A1) and explains that rows below 26 are all automatic, so should never be changed.

Before starting work on the sheet, spend some time playing with the variances of the light-green-background rows, so that you get a feel for the "range" of baadness(5) to greatness(1) scoring ... to "calibrate your eyeball".

NOTE re steps #7 & #8 ... send me the spreadsheet when you have completed step 6. I will do the copying and sorting, etc, and decide on which of the 540 test cases are "good enough" to warrant doing the #8 sliding-window testing. Then I will send you back a spreadsheet with just those cases, nicely organized for efficient review. You'll be testing the sliding window input for that situation, in a manner similar to what you just did.

THANKS!


(Extra Instructions for Spreadsheet.png)



Attachments
Attachments Extra Instructions for Spreadsheet.png (94KB - 1 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/23/2019 4:05 PM (#9251 - in reply to #9247)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
OK ... here is MOD-10 of the spreadsheet, which is needed for doing Step 8. You don't need this for Step 6, but IF YOU HAVEN'T STARTED STEP 6 YET, then use this instead of earlier mod's ... the first Tab is the same for both. (If you've started Step6 already using MOD-9, keep going ... I'll transfer your info to MOD-10 when you send your results in to me.

MOD-10 has a lot of new tabs, some of which are used for the sorting required by Step 7 (Jim does this). The sorting and copy-values process involves several sub-steps, but the final result is the new population of the Second (Step8Fours) and Third (Step8Threes) tabs ... and those are the ones you will be working on when I send it back to you after doing the sorts etc (which don't take much time but are too complex to reliably give instructions for).

The Step8Fours tab "looks" very much like the Step6Tests tab ... it has five extra rows waaay at the bottom but they are mainly informational. Note that cell A1 should be left as-is (important).

The most obvious difference is that the CumSlidWndoWyd2Thn input-row #9 now has a medium-green background ... that row is the focus of Step8. The values for it are all filled in ... firstly= 2 (the test scoress you've already completed are echoed for those columns).

You'll also note that the unlabelled row 17 has numbers in it that tell you what the "raw" test#'s were from Step 6, for each column in this tab. Those Test#'s are used since (for the Step8Fours tab), all of them had a final net net score of 4 from my sorting process (ie the best ones). This will be a lot less than the original 540 tests ... probably no more than 100 or so.

That list of "great" test#'s REPEATS as you move to the right ... and when it starts over, you'll see the Row 1 test-numbers restart at "-1" ... (neg indicates it's not the "2" set). The very first -1 column (in row 1) you find will also be the first column that NEEDS MORE TESTING.

Every time the cycle repeats and a new -1 appears, you'll also see that the newly-active CumSlidWndoWyd2Thn
input row's value changes (dark green background). The sequence may seem a bit odd, but there's method to the madness ...

After the recap of CumSlidWndoWyd2Thn=2 scores, your new tests will do CumSlidWndoWyd2Thn=4, then =1, then =3, then =0. So, if you have 90 cases that scored "4" from the step6 sorted analysis, then there will be 4*90=360 tests (for each of your three symbols) to complete this VERY IMPORTANT final round.

I did them in this order since your time constraints might not allow you to finish ... but please try really hard to. And please follow the order that I provided (long story as to why).

AFTER Step8Fours tab is done, repeat all of the above for the Step8Threes tab. It's arranged similarly re the new CumSlidWndoWyd2Thn variants ... but the input-patterns it uses are all different from those on the Fours tab ... these are patterns that scored "3" ... still purty darn good. I'm really, really hoping that you can complete this tab too. There will be a different number of test-patterns for Three's, but again, it's unlikely that it's over 100 (unless there were very few Fours).

OK ... that's the best explanation I can give in a post - call if you are confused. Thanks for all your help!
Top of the page Bottom of the page
JimDean
Posted 2/4/2019 4:25 PM (#9296 - in reply to #9251)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
Here is the "BgnEnd Testing" spreadsheet. It works like the others, but with a bunch of stuff stripped out or hidden ... and far fewer variants.

Use the VOLeval A2 DLL, as before.

I've attached a new chart template ... it has two instances of WaveTrader ZigZag loaded onto it ... I trust that all of you have a copy of this very useful plugin. If you don't, and don't want to buy it, I've attached three snapshots for the three Symbols being tested, so you can manually draw the pivot lines if you want to.

The template shows white for "perfect" pivot-based trend trade examples, and yellow for "swing" trade examples. Note that the very bottom has the ZigZag indicator pane that also may help you ID the pivot bars. I highly suggest you use my attached template!

The spreadsheet has 49 columns for the 7*7 variants of begin and end votes. Please do them all.
(type the Bgn/End 1-7 values directly into the inputs, since the sliders don't cover the range)

Set up the indicator: click Restore Defaults, then type 25, 130, 1 for the last 3 inputs (this leaves a gap on the RHS since Pivots can't form all the way to the HRE

Press the 6mo button on your display, and slide the chart to the right so that the calculated window is displayed. Don't change the Pivot entries.

Three "Scores" (1-5, where 1=great and 5=horrible) for each symbol for each column. You'll need to change the BgnVot input for each new column, and the EndVot input once every 7 columns. The score categories are:

Filter: Imagine that you have a totally independent trend or swing trading strategy, and that you are using this only as a FILTER ... that is, you want it to be green during the *likely* times that your strategy will be signalling a "good" bull entry, and red when your strat fires a "good" bear entry, and navy (or opp color of a flakey entry direction signal) during times when any bull/bear entry your strat fires should be *blocked*. Your strategy will NOT be "perfect" like the pivot-lines show ... it will most likely enter *later*, and exit differently (early or late) ... the pivot lines are there just to help you have a starting point for your eval. Score 1 if all the green/blue/red zones seem to be allowing the good stuff and blocking the bad stuff.

Entry: Now, imagine that this IS your Entry System (some other exit method ... presume the pivot lines are exits if you want), and that there are no other entries or filters in the strategy. The first bright lime bar in a wave (after navy or red-wave) is a long entry, and the first bright pink bar in a wave (after navy or green-wave) is a short entry. Maaaybe ... if the green/red wave is really long and if there are intermediate *dark* (not med) bars, then the first following bright lime/pink bar *could* be treated as a secondary entry or add, but focus mainly on the very start of the wave as described earlier. Score 1 if the Entries almost all look potentially profitable (presuming perfect exits from the pivot lines, for grins). Score 5 if most of them are in the wrong direction vs the pivot lines, &/or at the very very end of the pivot leg that is in their direction (ie way too late to profit from).

Exit: Now, imagine that this IS your Exit System (but the entries are totally separate ... presume entries occur at the pivots or thereabouts). Look for green waves within those up-leg pivots, and see if the first navy bar (or pink bar) at the end of the green wave offers a pretty decent exit ... it will NOT be perfect like the pivots which are based on crystal ball math. Same for red waves ending in navy (or lime) exits. Score 1 if the Exits almost all look potentially profitable, and Score 5 if most of them are way to late or way too early.

I thought about adding a fourth category for trade as a whole but I think that would be a lot harder to visually eval quickly ... and I can likely derive it from data mining, looking for cases where filter, entry and exit are all pretty good.

Please try *hard* to get this done by midnight tomorrow (ie Tuesday night). I suggest that you do all 49 columns for one symbol, then go back and do the next and the next.

Thanks!

(PAYX Pivot snap.png)



(AAPL Pivot snap.png)



(QCOM Pivot snap.png)



Attachments
Attachments aJDctVolEval.otd (288KB - 3 downloads)
Attachments PAYX Pivot snap.png (33KB - 0 downloads)
Attachments AAPL Pivot snap.png (33KB - 0 downloads)
Attachments QCOM Pivot snap.png (35KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 2/4/2019 9:35 PM (#9297 - in reply to #9296)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
Hi, guys

Sorry ... apparently when I fixed the labels on the SS, I didn't save it properly.
Lag => Filter
Chop => Entry
Dir => Exit

I corrected it, and the name (3 votes per, not 4) and attached it to the post just before this.

This post has a snap of how it's supposed to work.

Salim asked about the Bgn and End sliders ... no, they don't go 1-7 but you can always *type in* values 1-7 in each input. I double-checked the A2 code to make sure it works properly.

Thanks again for helping!

(49 test SS snap.png)



Attachments
Attachments 49 test SS snap.png (28KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 2/18/2019 2:16 PM (#9355 - in reply to #9297)
Subject: Cumulative Volume Indicator Distribution



Owner/Admin

Posts: 3798
20001000500100100252525
Location:
USA: GA, Lawrenceville
YET ANOTHER IMPROVED VERSION ... WITH VIDEO LINKS

I cleaned up this thread a bit ... removed the silly and early variants of the DLL ... the VOLeval A2 version with its orig docs are in the downloadable zip, in a prior post.

I removed all the early copies of the spreadsheet too. It's expanded a lot, and I polished it to death, not only to make it pretty and to milk all the possible use from it, but also to make it as generic as possible, for use with other testing (different param's, symbols, tests) ... but using the same methodology.

I'm attaching the the improved empty "templates" ... version "MOD 14", for future use & reference ... with signficantly improved label control, flexible dynamic sums from separate scoring, based on which scoring types should be totalled vs maintained separately. This need for separation - such as keeping Filter scoring separate from totals related to Entry and Exit scores - also called for four more tabs in the "net" sheet, to recap each top-thirds independently. In future, it's unlikely that the "Step6&8" version will be needed, since it layered the "fours and threes" tests on the initial "sixes" tests, requiring many extra tabs. For future studies, ideally simple variants of the "Step6 BE" template will be used. (replace older templates)

VIDEO NOTES:
1. the video was from a demo of Mod 13 ... so it will be missing some things in the paragraph above ... notably the way that the Step6 BE Filters scores should be sorted and binned separately
2. the audio is poor ... done at a starbucks while skype was running ... you'll need to turn the volume UP (as the session progresses), and maybe wear headphones ... sorry!)

Here is a link to the video-training for the "BE" (49-column simple) spreadsheet (Mod13):
https://vimeo.com/317090050/48981ccd66

Here is a link to the video-training for the "68" (540-column 2-phase) spreadsheet (Mod13):
https://vimeo.com/317087504/0d42405031

Wise to download these for your personal future use ... about five full days of development and tweaking time here)


Attachments
Attachments Mod14 SBAS Step6 BE - template.xlsx (746KB - 1 downloads)
Attachments Mod14 SBAS Net6 BE - template.xlsx (366KB - 1 downloads)
Attachments Mod14 SBAS Step6&8 - template.xlsx (1453KB - 1 downloads)
Attachments Mod14 SBAS Net6&8 - template.xlsx (383KB - 1 downloads)
Top of the page Bottom of the page
Jump to page : 1
Now viewing page 1 [50 msgs/pg]
( E-mail a Link | Printer Version | Search Room )
Frozen

Owner of site: Jim Dean -- Forum content is confidential, and may not be distributed without written permission.