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. 

1 Tim 6:18-19 (ESV) … (the rich) are to do good, to be rich in good works, to be generous and ready to share, thus storing up treasure for themselves as a good foundation for the future, so that they may take hold of that which is truly life.


CVW Archived Devel Discussion
Frozen
Jump to page : 1 2 3 4 5
Now viewing page 1 [50 msgs/pg]
Jump to room :
JimDean
Posted 1/3/2019 9:17 AM (#9178)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
This thread has the bulk of the CVW-development discussion, prior to adding several new Team members in June. It grew VERY long (5 pages), so I split it up into three pieces:

1. A thread with the initial VolEval Beta and SBAS testing (frozen now ... no more posts): http://tradetight.org/forums/thread-view.asp?tid=1431

2. A thread with "reference" posts such as help screens that will be kept up to date as development progresses (not open for user-posts): http://tradetight.org/forums/thread-view.asp?tid=1445

3. A thread with "ongoing" Discussion, for those who joined in later ... this is where comments and questions should be posted: http://tradetight.org/forums/thread-view.asp?tid=1444

This thread is now "frozen" (no more user posts) ... and may have some "gaps" from where I moved posts to the "VolEval Beta & SBAS" and the "Reference Info" thread.

If you have questions about anything here that is still relevant, post them in the DISCUSSION thread, with a link to the post in this thread that you're asking about ... a link is *required* if you'd like an answer ;~)Click here for steps to create that URL link.
Top of the page Bottom of the page
JimDean
Posted 1/23/2019 7:29 PM (#9223 - in reply to #9178)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
Approx starting point of post-beta "VolEval" routing ... eventually renamed as "CVWxprt" etc ...

Starting to ruminate about what the inputs would look like, if I added two Gurus. Arbitrary limitation (related to OT param panel popup) ... max height of 15 lines (param's plus "section-separators").

This would be slightly-rearranged "VOLxprt" version ... same pattern-inputs as now, but with Cap input deleted, and LastBar+Bars2Calc combined ... per the "nifty" explanation at the end of this post. Removal of those two not only reduces what user needs to fuss with, but also frees up two "slots" for the new gurus (so as not to force a second column on the Parameters Pane).

My prelim thinking re Guru mapping ... this may well change once I am able to review your spreadsheets ...
TimeFrame Guru provides automated settings for the following three expert sliders, and contributes to the automated setting of the fourth (MAvgSlopePdsFst2Slo).
Frequency Guru provides automated settings for the following three expert sliders, and contributes to the automated setting of the one just above it (MAvgSlopePdsFst2Slo).

VolCumulMethod expert slider is independent of the two Guru's since it's the core focus of the Indicator.

NOTE: if it turns out that CumSlidWndoWyd2Thn does not "mesh" well with the TimeFrame Guru ... I suspect that it will not ... then in that case I would move it down just below the VolCumMethod slider (making it independent of the Gurus), and would fully map the control of MAvgSlopePdsFst2Slo to the TimeFrame Guru (rather than also mapping it partially to the Frequency Guru). Again ... these mappings are just prelim guesses ... may well change after I've data-mined your spreadsheets.

... Layout Section ...
TIMEFRAME_GURU ... Short2Long ... controls next 3 and influences MAvgSlope
MApdsFMSorSm2Lrg
MixWtdHweExpHesSim
CumSlidWndoWyd2Thn ... if VolObOg=0, then CumSlid input sets RawV SpikeCap
MAvgSlopePdsFst2Slo

... Voting Section ...
FREQUENCY_GURU ... Many2Few ... controls next 3 and influences MAvgSlope
FuzzyBandNone2Large
TrdZonBgnCmbNetVote
TrdZonEndSepOppVote

... Overall Control ...
VOL_CUMUL_METHOD ... VolObOgVpVgAdAtNiNt
BarsToBeCalculated ... combines LastBarToCalcB4HRE & NumberOfBarsToCalc
OutputReturnHelpPlot


===============

QUESTION: if the "full" version (VOLxprt) has that set of 12 inputs (+3 headers), where nonzero values of the non-Guru inputs serve to override the "auto" settings implied by their related Guru (if any) ... THEN, would there be a "marketing" need for an "easy" version (VOLeasy or VOLcore) that has only these 5/6 inputs ... ?

TIMEFRAME_GURU ... Short2Long

FREQUENCY_GURU ... Many2Few

VOL_CUMUL_METH ... VolObOgVpVgAdAtNiNt

(maybe also CUM_SLIDING_WNDO, if one of its settings isn't clearly "the best")

BarsToBeCalculated ... Output window

OutputReturnHelpPlot ... Output format


=================

NOTE: In order to replace LastBar and BarsToCalc inputs with just one, I apply some logical reasoning. The BarsToBeCalculated input is an integer slider from 0 to 260, but allows neg# manual input. Its input value is interpreted differently, depending on whether a Plot output or a Return output is desired.
=== Return outputs usually just need 10 bars calc'd, so that the various "duration digits" can have their full 1-9 bar range ... the main thing re Return is to be able to specify the *final* bar, which is being displayed in FL or used for a Scan ... so, when Return is spec'd, the integer in BarsToBeCalculated *means* "#bars before HRE for last calc'd bar". (manual negative# input spec's a particular YYMMDD bar-date)
=== Plot outputs usually go all the way to the HRE, but sometimes only the most recent few months or couple of years are needed ... bars before that just waste calc time ... so, when Plot is spec'd, the integer in BarsToBeCalculated *means* "#bars to plot, ending at the HRE". (manual neg# input means plot all bars to the left of a gap on the right - input defines gap size)


==================

In the A3 version, I've modified five inputs so their ranges start with 1 instead of 0 (and of course the default and max inputs increase by one) ... all the logic remains the same. PURPOSE ... once Gurus are added, the Expert sliders that a Guru rules over will be ignored (ie Guru-gen'd value for them will be used), for every such slider that has a ZERO input. The inputs affected by this are: SpikeCapOffTyt2Wyd, MixWtdHweExpHesSim, VolObOgVpVgAdAtNiNt, CumSlidWndoWyd2Thn, FuzzyBandNone2Large

==================

The snapshot below has notes that might help explain how the Gurus consolidate existing inputs ... this results in the 15-line "full" list of inputs above ...

(Guru Param Ideas.png)



Attachments
Attachments Guru Param Ideas.png (277KB - 1 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/25/2019 12:34 PM (#9256 - in reply to #9223)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
(third reposting of this info ... had to clean up the snaps and some text ... sorry for repetition)

Informational Update re Progress ... no need for y'all to do anything ... most of these things have been mentioned in prior posts, and this post just announces their completion and implementation ... note that this implements the new Guru's (not yet functional), and rearranges inputs, etc as described in post just before this ...

Here's what I've done so far (broad strokes) for the A3 version (y'all have A2):

• dual-strip middle in Plot#2 (was #1), and big changes to Plot#3 (was #2)
• created minimal plot-1 with TZ band and 3 strips
• allow CumTyp=0 to have +incV /-decV TZ's (not bull/bear)
• Slope pds FMS input => 1-3 (2, 4, 6) with opt neg actual pds 2-10
• reverse Rtn minus-sign flag => NOT in active TZ
• Colors: Bgn7: 7, 6-4, 3-0; Bgn6: 7-6, 5-3, 2-0; Bgn5: 7-5, 4-3, 2-0
• ReturnT: 9 8 7 9 8 7 9 8 7
• TZ colors = grn/red, brighter for TZ votes closer to TrdZonBgn, new 1-9 map:
..... Bull TZ: 9= maxTrdZonBgn, 8= medTZvotes, 7=fewTZvotes ... 7= "warning"
..... TZ Off: 6= almostBullTZ, 5= no clear bias, 4= almostBearTZ ... 4 & 6 = "alerts"
..... Bear TZ: 1= maxTrdZonBgn, 2= medTZvotes, 3=fewTZvotes ... 3= "warning"
• neg FuzInp sets CntBig => counts track +1/0/-1 sign, vs std 1-9 detail levels
• shift param ranges >0 (prep for later Gurus): SpikeCapOffTyt2Wyd,
..... MixWtdHweExpHesSim, VolObOgVpVgAdAtNiNt, CumSlidWndoWyd2Thn, FuzzyBandNone2Large
• merge LastBarB4HRE & NumberOfBarsToCalc to BarWndoToCalculate (0=all):
..... Rtn: assume 10 calc bars, input= LastBarB4HRE; neg= ALL bars preceding LastB4HRE
..... Plt: assume thru HRE, input= FirstBarB4HRE; neg= plot all to the left of that gap
• add Rtn's 31-38 for B.bUuHh; 39=cappedV; 40=CumV; shift T.tPpKkB
• remove SpikeCap input (for any Cum calc assume inp=3, else SlidWndo def's Cap)
• revise gray param section headers and add two (nonfunctional) GURU param's
• updated first and second Help panels for all prior changes

So, the existing "expert" inputs are now pretty much ready for the two new Guru inputs to made functional, once I get the Step8 info back from Salim and Ken ... and the Plot revisions and Return expansions are completed ... and I've gotten a start on the Help panel revisions. Here are snaps of the new Param's panel (with two nonfunctional Guru inputs, and a two of the Expert inputs moved) ... and the first two Help panels ... note especially the major changes to the summary of Return-value options shown on panel#2 ...

(VOLeval A3 Parameters.png)



(VOLeval A3 HelpPanel -1.png)



(VOLeval A3 HelpPanel -2.png)



Attachments
Attachments VOLeval A3 Parameters.png (173KB - 0 downloads)
Attachments VOLeval A3 HelpPanel -1.png (302KB - 0 downloads)
Attachments VOLeval A3 HelpPanel -2.png (275KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/25/2019 4:23 PM (#9257 - in reply to #9256)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
The Gurus are now functional! ... and their 5*5=25 variations (seems about right) make useful sense ... ie they logically vary the patterns in a smooth and explainable manner.

In order to make this work consistently and logically, I needed to "reverse" the slider for the required (TrdZon aka Wave) End votes ... that is, when we require 7 votes to BEGIN a Wave, we are making it harder for one to begin. This will reduce the number of bars that are in red/grn waves. So, it makes more sense to have the End vote requirement slider have a similar effect ... as it's moved to the right, it should also reduce the number of bars that are part of red/grn waves. Therefore, what used to be an End vote slider that went from 1-3 votes, the new one now goes from 3-1 votes.

To help users grasp this, and not get too focused on actual vote numbers (which are after all "internal logic"), I changed the names of those two sliders ... and when I did, I removed the "TrdZon" from the picture since I'm using "wave" now to describe the overall green / red regions. The new names are:
BgnNetRqVotFew2Mny = 1-3
EndOppRqVotMny2Few = 1-3
... this makes the left-to-right slider impact for those two to have a similar visual effect as sliding the FuzzyBand from right to left.
... THAT makes it logical and easy to understand how the new "Guru_WavesBig2Small" input relates to those three subsidiary "expert" inputs ... when the Guru slides from left to right, the Waves go from including most if not nearly all the bars, to a relatively small percentage of the bars being in waves.
... We will need to remember, when examining the spreadsheet "scores", that I've reversed the End input

The Guru for the Trading Time Frame, which has the MApds, SlopePds, and MAcalcMeth as it's subservient expert inputs also maps to them logically ... right to left on the Guru has the same "kind" of effect that moving any of them individually right to left would have. The only "tweak" I did was to move the "Mix" option from the first to the last of the list ... the new MAcalcMeth input is:
WtdHweExpHesSimMix = 1-5
... I deliberately restricted the slider to max=5 ... if someone wants the Mix, they manually type in 6. Reason for this is that none of the five slider options in the Guru_TimeFrmFst2Slo input "map" to that "Mix=6" input.
... We may need to remember, when examining the spreadsheet "scores", the old "0" method is now "6"

As mentioned a couple of posts above, when I initially laid out the theory of the Xprt and Core versions, that the Cumul-Vol method input would be Guru-less ... and that since the Sliding Window is (still) as yet uncertain as to its range, interactions and usefulness, I've kept that out of the Gurus. Also, I mentioned earlier that the Cap limit would be handled in a more hidden fashion (see above). Finally, I also mentioned earlier that I was combining the 2nd and 3rd to last "output window" controls into a single one, that offers the same degree of useful control.

The NET RESULT of all this is what I think is a very orderly, understandable input collection. The attached snaps are updated from the prior post ...




(VOLeval A3 Parameters.png)



(VOLeval A3 HelpPanel -1.png)



(VOLeval A3 HelpPanel -2.png)



Attachments
Attachments VOLeval A3 Parameters.png (186KB - 0 downloads)
Attachments VOLeval A3 HelpPanel -1.png (303KB - 0 downloads)
Attachments VOLeval A3 HelpPanel -2.png (275KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/26/2019 2:53 PM (#9258 - in reply to #9257)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
and noowww ... here's CVWxprt and CVWcore!

Today I built in the code to allow the Xprt and Core variants of the routine to be gen'd from the same code (therefore keeping everything consistent). Also, I got the Guru defaulting and Expert overrides in place and tested, using the "shifted" param-lists (VOLeval A2 param's that could be zero now start with one ... except the CumVol Method and SlidingWindow inputs, since they aren't controlled by a Guru).

I spent much of the day creating the "promised" bottom half of the first Help-pane, which has a lot of useful info:
1. tells you the specific values that were input (and whether they "override" the Guru-default)
2. reports the "derivative" internal-calc metrics-of-interest (MA pds, Methods, Vote thresholds, etc)

Two trios of snaps are shown below, one for CVWxprt and the other for CVWcore. Outputs are the same of course.


(CVWxprt A1 Parameters.png)



(CVWxprt A1 HelpPanel -1.png)



(CVWxprt A1 HelpPanel -2.png)



(CVWcore A1 Parameters.png)



(CVWcore A1 HelpPanel -1.png)



(CVWcore A1 HelpPanel -2.png)



Attachments
Attachments CVWxprt A1 Parameters.png (173KB - 0 downloads)
Attachments CVWxprt A1 HelpPanel -1.png (428KB - 0 downloads)
Attachments CVWxprt A1 HelpPanel -2.png (276KB - 0 downloads)
Attachments CVWcore A1 Parameters.png (120KB - 0 downloads)
Attachments CVWcore A1 HelpPanel -1.png (328KB - 0 downloads)
Attachments CVWcore A1 HelpPanel -2.png (273KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/28/2019 7:38 AM (#9259 - in reply to #9258)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
Hi, guys

To clarify ... this is the "last chance" to comment on the Xprt/Core input arrangements I've posted previously. I don't plan on sending out another "intermediate" DLL if I can help it ... creating the DLL's is too time-consuming to do on a whim, and introducing "corrections" to a DLL's source is very awkward.

So, please look over the snaps, and give me your reactions as best you can. I'm mainly asking about "how it looks" re conceptual arrangement of things ... and also the extra detail at the bottom of the first help panel.

ALSO ... please let me know when you expect to complete the testing. Thanks!
Top of the page Bottom of the page
JimDean
Posted 1/28/2019 10:31 AM (#9261 - in reply to #9259)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
NOTE: (probably too late) ... ATVI, HAS, and PAYX all start off (usually) with a bullish green wave, even though when viewing the 126 bars I spec'd, it seems like the price action is anything BUT bullish. However, if you flip to a 9-month view, it's apparent that all of those had a fairly well established, clear uptrend prior to the arbitrary 6-month boundary.

QUESTION: did you notice this (ie the prior runup) during your testing, and interpret your results appropriately, or did you treat those lime band indications as "really bad direction" rather than "a continuation with too much lag"?

If you treated it as "really bad direction", to the best of recollection, did that mismatch heavily influence your voting for those symbols, or did the considerations of the middle and right side of the chart mainly rule the day?

Thanks
Top of the page Bottom of the page
JimDean
Posted 1/28/2019 10:56 AM (#9262 - in reply to #9261)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
I posted about this earlier ... this is the corrected version ... SIMPLER conclusions ... please review

Here is a new mapping I'm using that goes from -7 to +7 votes, to -4 to +4 return "T" values and three tricolor groups for the Wave Band.

Greenish/Pinkish tricolor Bull/Bear sets:
bright = ENTRY = enough net votes to begin a wave (can occur during the wave too)
medium = not quite enough to begin, but not enough opposite-votes (if any), to end
dark = WARNING = weak support for that direction, but not enough opposite-votes (if any), to end

Not-in-Wave tricolor sets:
dark-navy-blue = indeterminate case that doesn't seem to be leaning clearly towards bullish or bearish
blue = net votes are getting close to total required to begin a new Bull Wave (this used to be dark-green)
purple = net votes are getting close to total required to begin a new Bear Wave.

Of course, any time the Net votes are >= the Bgn threshold (and the #opp votes don't cancel it), the color is bright, and Wave begins if it's not already active.

PREVIOUSLY, I logic'd my way into finding TWO thresholds, one for weak and the other for warn ... in addition to the simple threshold for "begin". This was FLAWED, in that there are only three colors, so I can only use TWO thresholds ... and the begin thresh is one of them. SO ... the italicized comments below that develop two thresholds for warn and weak have been mushed together at the end for a single useful threshold between weak and warn ...

Coming up with "warning" thresholds is tricky since Bgn = net votes (sum of pos, neg & zero) while End = opposite voles (regardless of what other votes are). There are a variety of combinations, given that Bgn can be 5, 6, 7 and End can be 1, 2, 3 ... the table below maps it out:

Bgn=5,End=3 Bull warn upper thrsh: -2 & +5 => +3; lower thrsh: -2 & +0 => -2 ... so <= +0/1
Bgn=5,End=2 Bull warn upper thrsh: -1 & +5 => +4; lower thrsh: -1 & +0 => -1 ... so <= +1/2
Bgn=5,End=1 Bull warn upper thrsh: -0 & +4 => +4; lower thrsh: -0 & +0 => -0 ... so <= +2
Bgn=6,End=3 Bull warn upper thrsh: -2 & +5 => +3; lower thrsh: -2 & +0 => -2 ... so <= +0/1
Bgn=6,End=2 Bull warn upper thrsh: -1 & +6 => +5; lower thrsh: -1 & +0 => -1 ... so <= +2
Bgn=6,End=1 Bull warn upper thrsh: -0 & +5 => +5; lower thrsh: -0 & +0 => -0 ... so <= +2/3
Bgn=7,End=3 Bull warn upper thrsh: -2 & +5 => +3; lower thrsh: -2 & +0 => -2 ... so <= +0/1
Bgn=7,End=2 Bull warn upper thrsh: -1 & +7 => +6; lower thrsh: -1 & +0 => -1 ... so <= +2/3
Bgn=7,End=1 Bull warn upper thrsh: -0 & +6 => +6; lower thrsh: -0 & +0 => -0 ... so <= +3
... so, warnings for End input= 3,2,1 & Bgn input=5 => 0,1,2; Bgn=6 => 1,2,2; Bgn=7 => 1,3,3
... where 0,1,2,3 warn-thresholds are pos during Bull and neg during Bear waves
RESULT: for "most" waves, "warning" bar(s) will appear (unless price-reversal on mod-high volume)

The intermediate color is less important from my POV (please comment if you differ) ... it shows a weakening of the wave-strength, but not severe enough to be "questionable". So, it's somewhere between the Bgn (net) threshold input, and the Warning (net) threshold from the table above ...
Bgn=net5, End=opp3, Wrn=0 Bull weak upper thrsh: avg(5,0) = 2/3
Bgn=net5, End=opp2, Wrn=1 Bull weak upper thrsh: avg(5,1) = 3
Bgn=net5, End=opp1, Wrn=2 Bull weak upper thrsh: avg(5,2) = 3/4
Bgn=net6, End=opp3, Wrn=1 Bull weak upper thrsh: avg(6,1) = 3/4
Bgn=net6, End=opp2, Wrn=2 Bull weak upper thrsh: avg(6,2) = 4
Bgn=net6, End=opp1, Wrn=2 Bull weak upper thrsh: avg(6,2) = 4
Bgn=net7, End=opp3, Wrn=1 Bull weak upper thrsh: avg(7,1) = 4
Bgn=net7, End=opp2, Wrn=3 Bull weak upper thrsh: avg(7,3) = 5
Bgn=net7, End=opp1, Wrn=3 Bull weak upper thrsh: avg(7,3) = 5
... so, weak-Wave for End input= 3,2,1 & Bgn input=5 => 2,3,3; Bgn=6 => 3,4,4; Bgn=7 => 4,5,5
... where 2,3,4,5 weak-thresholds are pos during Bull and neg during Bear waves
RESULT: for "most" waves, some "weak" bars will appear (unless sharp price reversal with high volume)



As noted above, I can't implement these TWO thresholds for "weak" and "warn" colors ... I need to mush them together for a single threshold between those two colors (the thresh between weak and strong is the BgnVot input). Therefore, averaging the two above into an "orderly" pattern results in ...

Threshold for EndVote input= 3,2,1 & BgnVote input=5 => 1,2,2; Bgn=6 => 2,3,3; Bgn=7 => 3,4,4
... that is, if the net votes equal or better those values, it's "weak" ... if < those values, it's "warn"



I realize this may be sort of difficult to follow, but please try to ... it's hard to come up with reasonable rules for things like this, and I appreciate thoughtful feedback. This affects both the Return and Plot outputs, and could be the basis for strategy entries and exits.
Top of the page Bottom of the page
JimDean
Posted 1/28/2019 1:15 PM (#9264 - in reply to #9262)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
additional weak/warn vote threshold tweaks, after testing ...

I noticed when testing the arrangement above that the distribution of warn-weak wasn't bad, but could be improved. Also, it makes some sense to have the thresholds differ for each slider-increment, rather than being the same for two different settings (ie 2,2 or 3,3 or 4,4).

Additionally, since bearish/short runs tend to be quicker to end, and have more volatile retracements (ie dead cat bounces), I introduced a bull/bear distinction for the weak-warn thresholds so that warnings appear more readily for bearish waves ... the new thresholds are:

Bullish Waves: for EndVot=1,2,3 & BgnVot=5 => Thrsh=0,1,2; Bgn=6 =>1,2,3; Bgn=7 =>2,3,4;
Bearish Waves=more warns: BgnVot=5 =>Thrsh=-1,-2,-3; Bgn=6 =>-2,-3,-4; Bgn=7 =>-3,-4,-5
... that is, the farther from zero the threshold is, the more warnings will appear (vs being labelled as weak)


(note that all of these are within the ranges determined two posts earlier, but have additional "smarts")
(also - none of these color-threshold tweaks impact the overall duration of the green or red waves you tested)

similarly ... make thresholds for Alerts between Waves to be tougher for Bearish than Bullish
Top of the page Bottom of the page
SalimHira
Posted 1/28/2019 10:25 PM (#9265 - in reply to #9178)
Subject: CVW Archived Devel Discussion



Veteran

Posts: 174
1002525
Location:
USA: MD, Columbia
Hi Jim:

Here are early observations coming to mind:

1. On Help-Panel #2 that relates to Formula Builder pane of any OmniScript - after we Copy / Paste the omniscript, do we need to manually change the last two parameter BarsB4Hre & ReturnType ? Or, it stays AS IS wherever we plug that omniscript.

2. should this NOT be BEAR WAVE ? Your have BULL WAVE twice.
(#9262 - in reply to #9261)
"purple = net votes are getting close to total required to begin a new Bull Wave."

3. Could you please define the following in layman's words please ? How should it be interpreted in a real trading instance ? It's used quite extensively of its importance. i.e., does it define, maybe losing momentum, ROC reversing, volume itself significantly changing, etc.... #9262 - in reply to #9261

"SO ... the italicized comments below that develop two thresholds for warn and weak have been mushed together at the end for a single useful threshold between weak and warn ... "

IMHO, we should than be able to follow the rules as it states accordingly, thereafter.

4. One minor thought - on the bright colors

" Greenish/Pinkish tricolor Bull/Bear sets:
bright = ENTRY = enough net votes to begin a wave (can occur during the wave too) "

My thoughts here is that many times, the trend has already ended before it starts (confirmation bias), what kind of color(s) or Return values should we be observing that may be the case... I do understand the Slopes and Stacks and Bars horizontal bar will provide early clues.... Just searching for additional help clues on your side to observe while trading.
'''''''''''''''''''''''
Afterall, it all does look pretty good overall extensively written coverage in how it is explained in so much detailed, I'd never have thought of learning sooo much about VOLUME & PRICE itself.

Thank You.

Top of the page Bottom of the page
JimDean
Posted 1/29/2019 6:05 AM (#9266 - in reply to #9265)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
Hi Salim - thanks for your thoughtful reply

1. The top of the Help Panel instructs the user to replace the two “variables” in the function call with values, and indicates the range of those values. It says to copy the formula then paste it into the omniscript field. That seems to me to be fairly clear. What didn’t you understand?

2. Thanks. Yes. BearWave. Fixed. (A copy-paste oversight).

3. A Warning might be used for an early exit, especially if you have a way to re-enter when/if the wave resumes with a bright color, later. The exit could be partial or full. For very short term grabs, Weakness could be used the same way. Either could be used for scaling out gradually (with subsequent bright restarts used for scaling back in). I will build a System and Stop that allow the user to specify in the Params what to do when a Warning or Weakness is encountered.

4. The earliest bright color in a Wave is often right at the start of the Trend - that is, in many cases it’s a leading indicator (due to Volume component). However it’s not a crystal ball - it can’t know how long a move will last, and therefore you sometimes see a bright color just before a reversal. But it does change hues (green-Red/blue) pretty quickly compared to some other approaches. I think the answer to your Q is more related to Elliott wave theory - that is, crystal ball stuff. My recommendation is to tighten stops gradually as a trend progresses - ie as a function of time.

Hope this helps!
Top of the page Bottom of the page
JimDean
Posted 1/29/2019 6:28 AM (#9267 - in reply to #9258)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
Ref the snaps in message 9258 …

A. How do you feel about help panel #1? …
1. Are the summary one-line descriptions helpful? (Later panels describe more fully in most cases).
2. Do you “like the appearance”?
3. Is the new bottom half “Actial Inputs and Derivatives” understandable? Can you make reasonable guesses as to what the derivative values on the R half of the lines are referring to?
4. Are the relationships between the Gurus and their related Expert inputs fairly clear? (No example shown with asterisk - if an expert nonzero override is used, a “*” appears between the equals sign and the value.)

B. Most importantly- now that you’ve seen the two versions, do you think two (Xprt and Core) are “needed” for effective marketing?

C. Finally - I could allow negative inputs for the two Gurus, which would “map” to the “best” combinations of the various Experts, Willy-nilly, based on your Excel results. There could be maybe 20-50 such negative manual entries possible. (Or even via slider). I could do them separately (derived from subset sorts of your SS info), so the two negative Guru categories interact somehow - not necessarily in perfect ways though - OR I could set up one of them as a Master slider which controls every input except the last two. Doing any of this would have a LOT of sbas, and those “master” patterns wouldn’t necessarily “flow” rationally (ie slow to fast etc). But I am considering doing this if y’all think it will be helpful and not confusing.
(I’m not keen on setting up a new slider since that puts us over 15,into a second panel column - nor am I excited about repeating the whole user-editable Text file thing that MTV uses. But it’s possible for a later release if folks request it).
Top of the page Bottom of the page
KenWilsdon
Posted 1/29/2019 11:27 AM (#9268 - in reply to #9178)
Subject: CVW Archived Devel Discussion



Friend

Posts: 45
25
Location:
: ,
Hi guys,

Here are some answers to your questions Jim. *** is prior post questions or statements.

***To clarify ... this is the "last chance" to comment on the Xprt/Core input arrangements I've posted previously. I don't plan on sending out another "intermediate" DLL if I can help it ... creating the DLL's is too time-consuming to do on a whim, and introducing "corrections" to a DLL's source is very awkward.
***So, please look over the snaps, and give me your reactions as best you can. I'm mainly asking about "how it looks" re conceptual arrangement of things ... and also the extra detail at the bottom of the first help panel.
***ALSO ... please let me know when you expect to complete the testing. Thanks!

Answer: Looks like it is well organized, and fairly clear, end of first panel looks good. It provides a quick reminder of values.

Suggestion: One suggestion related to your response to Salim's question:

***1. The top of the Help Panel instructs the user to replace the two “variables” in the function call with values, and indicates the range of those values. It says to copy the formula then paste it into the omniscript field. That seems to me to be fairly clear. What didn’t you understand?

Suggestion: In your MARKETING of this product would be to give some examples of how the return values relate to a chart with the indicator on it. e.g., show where on a chart -1 to -10 appears (maybe by circling a -2 or -8 on two charts, along with the setting to get the chart), -11 to -20, etc. Many of us are visual learners (part of being in the YouTube generation), and having something like that would bring a new user up to speed faster, as they could compare a sample chart and settings to what they see on their chart. ("I like what I see, and I know I can get this as a return value, but which one is it that I am looking at right now on the chart"? - Answer that question, and things will be a lot clearer for end users). I realize some returns are for different MA's or calcs, but they should be shown as well, or a short demo video tying charts to both inputs and return values would be helpful. This is because the sliders go from 1 to 5, but the returns go from -1 to -40. The natural question is, how do the 2 relate? And how does it relate to what I am seeing? This will be a confusing part to many potential buyers. Most are not strong programmers.

Answer: I expect to finish today.

?***NOTE: (probably too late) ... ATVI, HAS, and PAYX all start off (usually) with a bullish green wave, even though when viewing the 126 bars I spec'd, it seems like the price action is anything BUT bullish. However, if you flip to a 9-month view, it's apparent that all of those had a fairly well established, clear uptrend prior to the arbitrary 6-month boundary.

***QUESTION: did you notice this (ie the prior runup) during your testing, and interpret your results appropriately, or did you treat those lime band indications as "really bad direction" rather than "a continuation with too much lag"?
?
Answer: Answered above?.
?
***If you treated it as "really bad direction", to the best of recollection, did that mismatch heavily influence your voting for those symbols, or did the considerations of the middle and right side of the chart mainly rule the day? ?

Answer: HAS was a problem? one way, as it looks like it had an earnings announcement where the gappy spike occurs at the beginning of the 6 month period. In the early days of the testing, this did influence me in regard to direction, but I normally needed at least 2 trades to go in the wrong direction for direction to be voted very bad( that being one of them). My reasoning is that the system should compensate for unusual market volume relatively quickly. As I moved along in testing, this became less of a consideration. It did affect lag as well.? As I was getting familiar with the process of scoring, I realized that it should go more toward lag than direction, and weighted it more heavily that way.

Answer: As for PAYX, I treated it as more lag than direction, unless it was getting out at the bottom of the down wave at the -6 month point (about July 28-30), then I penalized it for direction as well.

***Ref the snaps in message 9258 …
A. How do you feel about help panel #1? …

Answer: It looks pretty good to my eyes.

***1. Are the summary one-line descriptions helpful? (Later panels describe more fully in most cases).

Answer: Yes they are helpful. Again a demo video giving what the one liners mean would be helpful.

***2. Do you “like the appearance”?

Answer: Yes

***3. Is the new bottom half “Actial Inputs and Derivatives” understandable? Can you make reasonable guesses as to what the derivative values on the R half of the lines are referring to?

Answer: These would be guesses only. I would not be certain of the return values, and would probably have to do quite a bit of testing to figure them out. See my comments above.

***4. Are the relationships between the Gurus and their related Expert inputs fairly clear? (No example shown with asterisk - if an expert nonzero override is used, a “*” appears between the equals sign and the value.)

Answer: I think they are pretty clear.

***B. Most importantly- now that you’ve seen the two versions, do you think two (Xprt and Core) are “needed” for effective marketing?

Answer: Yes. The Core version is all that many will want to tackle with 6 sliders.

***C. Finally - I could allow negative inputs for the two Gurus, which would “map” to the “best” combinations of the various Experts, Willy-nilly, based on your Excel results. There could be maybe 20-50 such negative manual entries possible. (Or even via slider). I could do them separately (derived from subset sorts of your SS info), so the two negative Guru categories interact somehow - not necessarily in perfect ways though - OR I could set up one of them as a Master slider which controls every input except the last two. Doing any of this would have a LOT of sbas, and those “master” patterns wouldn’t necessarily “flow” rationally (ie slow to fast etc). But I am considering doing this if y’all think it will be helpful and not confusing.
(I’m not keen on setting up a new slider since that puts us over 15,into a second panel column - nor am I excited about repeating the whole user-editable Text file thing that MTV uses. But it’s possible for a later release if folks request it).

Answer: I personally don't think that is needed.
Top of the page Bottom of the page
JimDean
Posted 1/29/2019 11:43 AM (#9269 - in reply to #9268)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
Thanks Ken.

Re the return values - see earlier post #9199 posted on 1/4. Of course it’s noy totally up to date but I hope that the help panels about returns (albeit dense) and the big sidewise table snapshot help clarify things.

Returns for slider input -1 to -40 are very “orderly” once you understand the pattern - each “-x1” value is a meld of several (up to 7) piece of info, and the nine cases after that in sequence provide individual detail of each item. Help panel #2 is a concise summary.

Not all return items have corresponding plot colors. Many are durations (ie counts) of a given color. But nearly all of the capital-letter returns have color related plots, where 1=very red/bear and 9=very bull/green.

The new help panels will explain and will be more extensive. Problem with providing detail info in videos is that it can’t be quickly ref’d later. But it does make sense to do a video walkthrough to familiarize users.

Thanks again.
Top of the page Bottom of the page
SalimHira
Posted 1/29/2019 4:25 PM (#9270 - in reply to #9178)
Subject: CVW Archived Devel Discussion



Veteran

Posts: 174
1002525
Location:
USA: MD, Columbia
In response to post #9267 questions, I concur with Ken's comments while not to repeat myself either, and Jim's clarifications clears any doubts, with high emphasis on "C" - it is not needed, imho. Thanks.

p.s. I am trying to complete the SS by tomorrow late evening, if all goes well with my schedule - but trying sooner.
Top of the page Bottom of the page
JimDean
Posted 1/29/2019 4:56 PM (#9271 - in reply to #9258)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
Thanks to both of you.

I’m a bit confused about Kens response re the bottom half of Help pane 1 with the input values and the derivative info. Specific questions:

1. Is it clear that the first value in each of the first eight lines is the input value (if a guru or an asterisk), or if no asterisk, it’s the expert-input’s Guru-assigned value?

2. Is it clear that the other values on the line after that are:
A. Not answers but rather internal derivatives of the inputs
B. For which of those labeled values can you understand the labels. Which not?
C. Do you think it’s helpful to users for me to provide that info?

====

New question:
Since the Core version does not provide “fine” control over the six subsidiary expert inputs that the Gurus define, I’m thinking that it does not make a lot of sense to provide as-extensive output detail for that version (since user can’t really “tweak” things in detail). So, I’m thinking about limiting the Core Plot output to just the simplest “Wave Band plus three strips” option, plus the “explanatory-but-not-very-useful” plot #3 that has the various raw components all stacked on top of one another (Cumvol, vol, colored ma’s, colored dots, etc). Also, the Core Return would be limited to the first T.tPpKkB output and its breakout options, plus the CumVal.
Of course the Core DLL is less expensive - not sure if the subscription would be or not.
What do you think?
Top of the page Bottom of the page
KenWilsdon
Posted 1/29/2019 10:05 PM (#9272 - in reply to #9178)
Subject: CVW Archived Devel Discussion



Friend

Posts: 45
25
Location:
: ,
Jim,

I’m a bit confused about Kens response re the bottom half of Help pane 1 with the input values and the derivative info. Specific questions:

1. Is it clear that the first value in each of the first eight lines is the input value (if a guru or an asterisk), or if no asterisk, it’s the expert-input’s Guru-assigned value?

- - - I assume you are referring to the CVWxprt Help panel, specifically those in the Trading Timeframe Inputs and the Voting Sensitivity Inputs. I assume the Guru sets the values of each of the next 3 lines to the value after the = and before the ";"

2. Is it clear that the other values on the line after that are:
A. Not answers but rather internal derivatives of the inputs

-------Not totally clear what an internal derivative of inputs means. I understand derivatives from calculus, but you are using it in a different sense. I assume you mean what is after the ";" in each line. That is clearly the description of what the slider does.

--------Does the Guru number in first line in section set the value to what follows = and before the ";" for each of the next 3 lines? That is what I assume is the case. I hope that is clear.

B. For which of those labeled values can you understand the labels. Which not?

--------I dont understand WtdHweExpHesSimMix; Wtd is obviously weighted, but the rest is cryptic. bEMA is also one that I am not familiar with. The rest of the first 8 lines I can figure out.

C. Do you think it’s helpful to users for me to provide that info?

--------Yes, it helps them to understand what the control on the slider does.


What you suggest for the Core version seems reasonable for the stripped down version.
Top of the page Bottom of the page
SalimHira
Posted 1/30/2019 9:46 AM (#9275 - in reply to #9178)
Subject: CVW Archived Devel Discussion



Veteran

Posts: 174
1002525
Location:
USA: MD, Columbia
I will attempt to summarize to your questions. I pretty much understand but some concerns too - but herewith, is what may need clarifications in addition:

What are WavesBig2Small ? ... folks may need to reflect upon that.

How the CumSlidWndWyd2Thn affects the slider, and if not satisfied, user can override by manually typing them ?

Pls expand on WtdHweExpHeSimMix - lost here and/or anyone who see it, cannot make out what may be happening here.

============
1. Is it clear that the first value in each of the first eight lines is the input value (if a guru or an asterisk), or if no asterisk, it’s the expert-input’s Guru-assigned value?

>>> I recall how it was used in MTVxprt, so assuming its the same.

2. Is it clear that the other values on the line after that are:

A. Not answers but rather internal derivatives of the inputs

>> It is not clear, imho.

B. For which of those labeled values can you understand the labels. Which not?

>>> see above

C. Do you think it’s helpful to users for me to provide that info?

>>> Definitely. With only the DLL on hand, its difficult to comprehend as we all are used to going into the OmniLanguage itself of source code to read the "definitions" when things become blurry for recollection.
Top of the page Bottom of the page
JimDean
Posted 1/30/2019 9:47 AM (#9274 - in reply to #9272)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
I've modified Help Panel #1 as shown in the snap to clarify the meaning of WtdHweExpHesSimMix ... please let me know if this clear enough ... ie that it refers to the calc method used for the MA's as being Wma, Ema, Sma, or a mix of those (within the confines of one line of text, that is ;~).
Also ... Help Panel #2 now more clearly describes how to get the single-digit "T" value output. And I've modified the Core version of HP#2 to only show the -1 (thru -10), -11 and -12 options.

I've also implemented a Core vs Xprt distinction re what outputs are provided with each:
Plot: Core does not offer the Plot with the Slope Histo, dual-barV strips, and Stack Histo (now= Xprt option #3)
Return: Core does not offer return-options 11-38, and 39-40 are mapped to 11 & 12 for core. That is, the "net result" return "Wave T.tPpKkB" and its nine breakout-digits-options is provided for Core, but not the detailed "Slope P.QqRrSs" or "Stack K.LlMmNn" or "BarUH B.bUuHh" information.
... I'll likely provide a lower price for the Return subscription for Core, as well as for its DLL

I'm going to provide licenses to BOTH Xprt and Core for people who buy Xprt, so they can use either (both give same answers).

=============

KEN: the SS you sent me last night does not contain the final 140 cases ... just the first 400.

BOTH (esp Ken, since he missed the prior Skype): I'd like to do a Skype-training-demo today for you ... what EST times do you have available? (sooner better than later)

(CVWxprt A1 HelpPanel -1.png)



(CVWxprt A1 HelpPanel -2.png)



Attachments
Attachments CVWxprt A1 HelpPanel -1.png (427KB - 0 downloads)
Attachments CVWxprt A1 HelpPanel -2.png (273KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/30/2019 10:26 AM (#9276 - in reply to #9275)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
Hi, Salim

I posted some answers and a couple of updated snaps about the same time you posted your reply, so you may not have noticed them ... if not, please review my prior post.

I'm discouraged to hear that the bottom half of HP#1 is apparently pretty much a mystery. I'd have thought that the labels which match the listing in the upper half would have clearly ID'd those rows as being related to the various Parameter inputs ... and an equals sign followed by a value would have been fairly obvious as representing the value input by the user. HOWEVER, with the hope of improving it, I am going to REVERSE the meaning of the asterisk ... that is, values will have an asterisk if the Guru set them (ie they were input as zero but "used" as the noted value). I hope that will help. See snapshot below for those revisions.

I'm also surprised that an abbreviation like:
MA pds: Fst= 3, Med= 6, Slo= 12
... is so hard to understand. The user can SEE the MA curves on the output plot ... and the Fast/Med/Slow "idea" is basic to the program ... so why is that abbreviation too obscure? Please keep in mind that the panel has max-height (ie #row) limitations.
... and the other "detail-y" info on following lines seems to be even less cryptic (maybe "bEma" looks odd, but I think they "get it" that it is "sort of" an Ema ... later Help describes bEma)

Regarding your comment about wanting info that you otherwise might have gotten out of viewing OLang:
a. this routine is too complex to report the intermediate values for *everything* that is going on
b. much of the "coding mechanism" is proprietary so I don't want to make it easy to duplicate
c. the other Help Panels (presented previously in thread) describe the calc methods qualitatively ... a LOT MORE than Nirvana typically explains their plugins.


Re WavesBig2Small ... would you prefer "WavesWide2Thin" ? I'm suprised that the idea of a "Big Wave" vs a "Small Wave" is hard to follow ... unless you don't get what a "wave" is. The "wave" is a series of bars that are all the same family of bullish (green) or bearish (red) hues. Again, there is a limit of how many lines I can include on this screen ... what would you suggest?

Re CumSlidWndWyd2Thn ... the final part of the definition line says "or typed" ... which means the user can type in the number of periods they want. I don't really want to encourage people to do that so I don't want to make a big deal out of it ... I considered not even mentioning it except in a video. Is it also unclear to you that the user has an option of manually typing int FFMMSS for the second input? I debated whether to make that obvious, too.

Re WtdHweExpHesSimMix ... I tried to clarify that in my prior post which you probably didn't see before responding ... is the new one-liner adequate?

(CVWxprt A1 HelpPanel -1.png)



Attachments
Attachments CVWxprt A1 HelpPanel -1.png (432KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/30/2019 11:05 AM (#9277 - in reply to #9276)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
OK ... I've modified the HP#1 to define the Guru meanings ... hopefully the explanations of "Fast" for the first one and "Big" for the second will do the trick.
I also added a subtitle line for the headers of the guru sections on the top half (Xprt only), to help clarify how the Guru "drives" the Experts when they are zero.


(CVWxprt A1 HelpPanel -1.png)



(CVWxprt A1 HelpPanel -2.png)



(CVWcore A1 HelpPanel -1.png)



(CVWcore A1 HelpPanel -2.png)



Attachments
Attachments CVWxprt A1 HelpPanel -1.png (461KB - 0 downloads)
Attachments CVWxprt A1 HelpPanel -2.png (273KB - 0 downloads)
Attachments CVWcore A1 HelpPanel -1.png (331KB - 0 downloads)
Attachments CVWcore A1 HelpPanel -2.png (215KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/31/2019 6:51 AM (#9278 - in reply to #9276)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
A few small changes worth noting:

1. I modified the param names for the gurus to use caps: GURU_TymFrmFst2Slo and GURU_WavAgrs2Cnsrv ... required removing "e" from "Wave' and changing "Time" to "Tym" so that it still fit on the param pane and in the message boxes. Reason: later, there *might* be a need for another slider for Master Presets ... if so, I might need to remove the intermediate "Voting Sensitivity" label-line to make room for the new one on top. In that case, the all-caps "GURU" will help visually divide the two sections.

2. I modified the -7 return value to be "B.b" instead of just "B" ... thus including the duration-bars with that output, since the Core version does not offer other alternative outputs to obtain the "b". In line with that, I also modified the 2-digit return values to T.t, P.p, K.k (in both Core and Xprt), and the Q.q, R.r, S.s, L.l, M.m, N.n, U.u & H.h for the other 2-digit return options avail in Xprt. The decimal makes it simpler for the user to parse the two pieces using the functions INT() and FRAC().

3. Minor mods to output format for HelpPanel #2 to reflect the decimal changes above, AND to clarify what the second-to-last option is ... now says "BarVol (neg=cap)" rather than "Capped RawVol".

4. Also ... I fixed the plotting so that when the Sliding Window is off, there are no cyan/pink histo's below the zero line to show what was removed from the left side of a nonexistent window. Along the way, I also tweaked the dot-plotting for that plot so the dots don't get quite so huge (in some y-axis situations), and also are a bit tighter together ... but still readable.


(CVWxprt A1 Parameters.png)



(CVWxprt A1 HelpPanel -1.png)



(CVWxprt A1 HelpPanel -2.png)



(CVW A1 Detail Plot (noSWndo).png)



Attachments
Attachments CVWxprt A1 Parameters.png (173KB - 0 downloads)
Attachments CVWxprt A1 HelpPanel -1.png (460KB - 0 downloads)
Attachments CVWxprt A1 HelpPanel -2.png (275KB - 0 downloads)
Attachments CVW A1 Detail Plot (noSWndo).png (47KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 1/31/2019 11:22 AM (#9279 - in reply to #9278)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
FOUR QUESTIONS for y'all:

1. I'm undecided about my earlier idea regarding leaving out one of the plot options for Core. It's so useful for diagnosing "whys" that I think maybe I should leave it in. I'm talking about the one that is option #2 in your A2 DLL's ... with the Wave band on top, then Slope histo+curves, then dual strips for BarV, then Stack histo+curves. (no matter what, I will limit the Core Return values to just the T.tPpKkB overall set)
Is that plot "high-value and fairly essential" to making the Core version understandable?

... I'm thinking of ways to use the test results to best effect ...

2. I can use the Step6 & Step8 results in combo, to determine what the "most commonly viable" patterns are for each of the two Gurus. That is, the GURU_TymFrm would allow *negative* values, which would assign its three experts based on which have provided the "best" results in the testing (independent of the other Experts). Similarly, I can do the same for the GURU_Wave input, where its negative inputs would assign patterns of its three Experts based on their independent results. The sequence from -1 to whatever (-5 or -10 maybe) would not be Fst2Slo or Agrs2Cons, but rather Best2Good "scoring".
What do you think?

3. I can use the Step8 results to create a new "PreferredPatterns" slider that would be at the very top. It would assign values for *every one* of the 8 experts, based on tier-sorting within the "Fours" results ... (WavAgrs2Cnsrv looping within TymFrmFst2Slo looping) ... all suggestions would (likely) be "four" scores ... using overlap of Ken and Salim results ... probably about 50 options. This would *not* require me to create a Preset Text file ... it would all be internal mapping. If I do this, then the existing Guru's and Experts could override those PP's if they are nonzero (will adjust input ranges for CumMeth and SlidWndo to accomodate). The PP slider would be in Core & Xprt.
What do you think about this?

... of course there's no way to be sure whether the PP's or the neg-Guru options will actually work out to be amongst the "best" for a given user's situation. I'm fairly confident that our methodology, re separate eyeballs, three scoring cat's, and carefully chosen spectrum of symbols will yield generally useful results.

4. This one needs some explanation ...
... the *really handy* think about the PP and the neg-Guru options is when someone wants to pair up CVW with other stuff in a strat, and use StratWiz to find good fits. SW can simply loop thru the ~50 likely-fairly-good PP options, rather than having to do a bunch of not-very-likely-to-be-useful permutations that would be needed if SW was asked to loop thru all the experts in a tiered fashion.
... in fact ... once the 50 are chosen for PP, I might set up a SW run that uses CVW for entries and exits, and loops thru the 50 to determine which are the most profitable for a big list of stocks ... then provide PP a negative range, where the SAME 50 are mapped, but in a different "SW profitability" order (rather than the pos-50 sort described above).
Creating these variants on the Gurus and creating the PP is not terribly difficult ... much easier than (and faster-executiing) than any "preset text file" type of approach.
What do you think about the SW-profitability-sorted neg-PP-input option?
Top of the page Bottom of the page
KenWilsdon
Posted 1/31/2019 3:12 PM (#9280 - in reply to #9178)
Subject: CVW Archived Devel Discussion



Friend

Posts: 45
25
Location:
: ,
Hi Jim,

RE:

Q1.

>>> My thinking is that it is not fairly essential. Core users will probably people who don't care so much about the nuts and bolts, but just want something that works. it will be more likely that Xpert users will want to know the whys. It also provides another plot as a reason to upgrade.

Q2.

>>> That would be a good option for a new user to get started, as well as a good starting place for core users to get experience with CVW.

Q3.

>>> So question 3 is different than Q2. A master slider. Good idea. That allows the Q2 answer to give more granularity than the Q3 option, if a user wants that, without going through all the permutations. If people like to tinker in a broad sense, this would certainly help them.

Q4.

>>> I think that is a good idea. If the markets change over time, then a new SW sort could be done later, and they still would allow their use, even if the order changes somewhat.
Top of the page Bottom of the page
JimDean
Posted 1/31/2019 3:24 PM (#9281 - in reply to #9280)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
Re Q4 (and 3 and 2) - clarification

All of these slider-controlled preset patterns would be hard coded, not driven from a text file, or in any other way modifiable by the user. I’m thinking of killing off the Preset file entirely from MTV and doing a similar hardcoded approach (I’ve sent an email to all MTV owners requesting N/Y as to whether they’ve created any custom presets). Hardcoded means no slowdown to read a file. And a lot less code to implement. I’m thinking ahead about paradigms that I might implement in all the other stradicators.
Top of the page Bottom of the page
JimDean
Posted 2/1/2019 3:48 PM (#9287 - in reply to #9278)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
More changes ... to introduce the "PreferredPatternsMaster" input ... (reposted due to prior errors)

To do this, I had to eliminate one of the gray param-pane "headers", but with GURU capitalized I'm hoping the two separate GURU sections remain clear.

The new PPM input does function, but its predefined set of 50 input patterns is set to gobbledegook for now ... once I get the rest of the Step 8 scores back from the two of you, then I'll be able to analyze them and establish the "50 best" (or some other number).

If the PPM is nonzero, then all the other Guru and Expert inputs can be zero ... and must be, if the entire pattern is to be utilized. I revised the input sliders for the two Gurus and the Cum Method and Wndo Experts to reserve zero for this purpose (shifted the two Expert input-meanings by +1). Help panels have been fixed for this stuff, as well.

The internal logic is really, REALLY versatile. If you set a nonzero PPM, then you can choose to set any of the Gurus &/or Experts to nonzero values ... it will use the PPM pattern for any that remain zero. Of course, if a Guru is nonzero in that situation, then it overrides the defaults for any of its subsidiary Experts (that themselves are zero).

It really does make "obvious" sense once you get it ... but to clarify, I enhanced the bottom half of Help Pane #1 to allow the "default-flag" to be "^" if PPM sets the value ... or be "*" if a Guru sets the value ... or nothing if the Expert itself was nonzero.

Please carefully study the input-combo and let me know if this is clear, from the snaps below.

... Later on, I might also allow PPM to be negative, which would utilize input-patterns that Strategy Wizard determines to be typically profitable ... but I need to wait till I get the NXF-shell routines in place for that (see my RTF file posted here)

... Also, maaaybe (need to see testing), I might allow the Guru's to be negative, to utilize "optimal" combo's of their three related Experts ... this could be pretty handy but it would require one more set of a few hundred Tests (an extension to the results of Step 8). I'll wait to eval Step 8 before I present that to y'all ... if you have the stomach for it. ;~)


(CVWxprt A1 Parameters.png)



(CVWxprt A1 HelpPanel -1.png)



(CVWxprt A1 HelpPanel -2.png)



Attachments
Attachments CVWxprt A1 Parameters.png (177KB - 0 downloads)
Attachments CVWxprt A1 HelpPanel -1.png (476KB - 0 downloads)
Attachments CVWxprt A1 HelpPanel -2.png (275KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 2/1/2019 4:31 PM (#9288 - in reply to #9287)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
Here are the Core snapshots that parallel (as closely as possible) the Xprt examples in the prior post. Note that the Xprt example has two Expert-overrides (one for each Guru) that cannot be duplicated by the Core routine.

Please study the panels, notably HP#1, and let me know if it makes sense.


(CVWcore A1 Parameters.png)



(CVWcore A1 HelpPanel -1.png)



(CVWcore A1 HelpPanel -2.png)



Attachments
Attachments CVWcore A1 Parameters.png (119KB - 0 downloads)
Attachments CVWcore A1 HelpPanel -1.png (369KB - 0 downloads)
Attachments CVWcore A1 HelpPanel -2.png (216KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 2/3/2019 4:33 PM (#9293 - in reply to #9278)
Subject: CVW Archived Devel Discussion



Owner/Admin

Posts: 3806
20001000500100100100
Location:
USA: GA, Lawrenceville
COMBO POST ... I deleted the prior post since this one shows updated snaps. Here's its text:

2/2/19: Decent progress today. Ken says his Step8Fours will be ready tonight ... so it looks like I've got another "free" day or two before he gets me the Step8Threes as well.

I'm using that time to significantly expand and rearrange the Help panels ... there are nine so far, each pretty tall. After HP#2 which you've seen, #3 now gives a general explanation of "how" to use the program (ie Master, Gurus, Experts), followed by a 7-step synopsis of how the program does its calculations (all of this is brand new Help info). Then the next all-new two panels (completed) explain each one of the inputs, up to the final two, with as much detail as I thought might be interesting and useful.

Panel #6 begins with explanation of the versatile new output-window-extent input (2nd to last one), clearly explaining (I hope) its purpose, how it works differently for Plot and Return, and even the special negative-number input option. (prior panels explain optional neg inputs also). I still need to add a clear description of Plot#1 ... the big top band and three breakout strips, with all their colors and meanings and use, etc. Since this plot is likely to be the main one used, I'm going to try to polish the explanation well.

Panel #7 explains Plot#3 and Plot#3 colors and meanings and use, etc. It's not shown since I haven't yet updated it. Panels #8 & 9 will describe all of the 40 return options.

Once it's all done, and y'all have both proofread it carefully, I'll also create a PDF that contains the entire thing, maybe with some snapshots (but I don't want to go hogwild on illustrations).

=================

2/3/19 Progress Update:

I'm still working on the Help Panels. Reviewed my prior-posted work, and cleaned it up as well as clarified a lot of things. Made a bit more progress forward, too, but still need to finish Plot and Return definitions. Also cleaned up RHS of plots, when a negative # used for second to last input.

Main "new" thing I did was to expand the info provided at the bottom of the HP#1 ... it now shows the Volume Cap multiplier, and the Vote-threshold for Warnings (dark red/grn) and Alerts. While doing this, I also provided for the Xprt user to specify the Volume Cap, not just turn it off. Help Panel #4 explains how (special typed values for MAmethod &/or SlopePds inputs). When those overrides (or the FFMMSS MApds override) are used, the HP#1 now is smart enough to show them correctly, and adjust its formatting accordingly so no word-wrap occurs. Of course, those overrides are only available to the Xprt user. This was all pretty messy to accomplish ... but it's done now!


(CVWxprt A1 Parameters.png)



(CVWxprt A1 HelpPanel -1.png)



(CVWxpco A1 HelpPanel -3.png)



(CVWxpco A1 HelpPanel -4.png)



(CVWxpco A1 HelpPanel -5.png)



(CVWxpco A1 HelpPanel -6.png)