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. 

Proverbs 16:9 (ESV) ... The heart of a man plans his way, but the Lord establishes his steps.


Sticky CVW Reference Information
Frozen
Jump to page : 1
Now viewing page 1 [50 msgs/pg]
Jump to room :
JimDean
Posted 4/3/2019 3:00 AM (#9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
This thread was initially built using posts extracted in the very long mid-phase discussion thread (when new Team members were added in June), and will be expanded and kept (reasonably-well) up to date as the project proceeds. Info here will become part of the Help Manual, in one form or another, and as development proceeds, all Team members *must* keep abreast with this, to avoid getting lost.

Note re the “posting” dates ... the earliest ones (prior to the Color Chart) were moved from the archived discussion thread … I've tweaked them bring them fairly well up to date. I’ve noted when I did the updates.

When I update Posts in this thread with new info, I will *revise* the prior post instead of creating a new one, so that any URL-links pointing to it aren't broken. I'll make a brief post in the Discussion forum pointing to the updated post here, so that you get an email notification for Team members. It's *important* to keep up with those posts.

This (active) thread is not open to Team-posts, but questions or comments about it are welcome in the DISCUSSION thread ... as long as a URL-LINK is INCLUDED in the discussion-post, pointing to the specific Post in this Thread that is being mentioned. Click here for steps to create that URL link.
Top of the page Bottom of the page
JimDean
Posted 4/3/2019 1:00 PM (#9646 - in reply to #9645)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: Video Overviews & Training

=== Updated on 6/26/19 ===


Here are links to the three videos produced in May, as an overview of Stradicators, with considerable walk-thru detail of CVW, and some interesting info about the "Robust TP", etc. It's important, before diving in to the posts in the Reference and Discussion threads, to go through all these.

1. Stradicator Paradigm ... 70min ... BY FAR the most important ... many exciting features!
https://vimeo.com/jddean/stradicatorparadigm

2. First-Five Stradicators ... 65 min ... illustrates #1 via CVW, w/quick review of other four
https://vimeo.com/jddean/stradicatorfirstfive

3. Strad Q/A & Special Topics ... 35min ... diverse Q's; Special= Options & Robust TP features
https://vimeo.com/jddean/stradicatordiscussion


I will be adding more training videos, but most likely it will be after the DLL is complete
Top of the page Bottom of the page
JimDean
Posted 5/5/2019 9:56 AM (#9607 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: Color Choices for Plotting

I try to keep "families" of things (trade direction, vote strength, etc) a similar gradient-hue. That is, Bull is greenish with its alert-confirm bluish. Bear is reddish with its alert-confirm purplish. It wouldn't fit to use yellow or white or brown, for instance. For future reference, I'm attaching a color chart to this post so that you can see all the choices that I have to work with.

(A9F50EAE-FEEA-4960-85EE-5E3602F2C33A.png)



Attachments
Attachments A9F50EAE-FEEA-4960-85EE-5E3602F2C33A.png (153KB - 2 downloads)
Top of the page Bottom of the page
JimDean
Posted 5/7/2019 7:24 PM (#9589 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Regarding Optional Filters in a Stradicator

All Stradicators will have at least one "Guru" (the VoteGuru is part of the standard paradigm). Guru's will always by nature be simple inputs, with sliders no wider than -9 to +9 (here, they have limits of 5 and 6). So, I'm using Guru param's as places where the user can optionally type in a 7-digit number, to define how an external filter's analysis should be folded in to the Stradicator logic. Think of this as specifying a filter in the OT Strategy Filter block ... Stradicators provide that same flexibility ... and can therefore include the impacts of the external filter in it's various Equity-stat's and Equity-Vote toggling. Vereee cool, since this means I can sell separate licensed "general-use" filters (ie that work on a standalone basis), which the Stradicator user can "call" (via the typed Guru input). This means I will *not* have to add that filter-code to each Strad that might use it ... and it means in theory that the number of filters are unlimited.

It's tricky to do this and keep the execution time efficient, due to how the OLang Parser handles external function calls. I've documented the "rules" about that elsewhere ... the comments below refer to them ... but suffice to say that this structure absolutely minimizes any "time waste" for unused-but-available filters, to effectively nothing.

I might initially offer one or two filters (ConZone and BigJump) that users can choose to buy. They can be used separately (they plot meaningfully and return helpful values) ... or they can be called by CVW or any future TT Stradicator. If they are called by a Strad, they *cannot plot* any output (that's an OLang limitation) ... but their net effects will be known by the Strad and appropriately utilized (typically, to block or delay an Entry).

If the user spec's one of them in the Strad's Guru-input, but also wants to *see* what that Filter is *doing* with the selected param's, all the user needs to do is to separately plot that Filter as an indicator ... its effects will be seen on the chart (which presumably also has the Strad plotted). If the user has specified more than just one Filter for the Strad, then they can plot whichever one(s) they wish on the chart. The plotted version is disconnected from the operation of the Strad ... and it's up to the user to make sure the same param's are spec'd for the Strad's filter-call (via Guru input), as are used in the separate plotted filter.

The comments in the snapshot below explain things pretty well. It's the first case I've known that uses a seven-level nesting of OLang Indicator calls (the 0-param filters are the seventh, and the initial first-level call is made to TTndGetFltr5P from the Stradicator) I'll distribute these six "GetFltr" routines as OLang source-code Indicators which the user can tailor if they want. Per the instructions, they would manually revise one or more of the case 1-9 "Return = 0" statements with a call to a TradeTight DLL, or to one of their own custom OLang routines (which in turn could be based on a Nirvana indicator, I suppose).

I've attached a Zip file that has all 6 nested routines, ready to go :~) Of course, once I get ConZone and BigJump (or other TT filters) ready, then I'll specify them as 5P case 1 & 2, etc.


(Comments Snapshot from 5P.png)



Attachments
Attachments Comments Snapshot from 5P.png (72KB - 0 downloads)
Attachments GetFlltr Sextet.zip (8KB - 9 downloads)
Top of the page Bottom of the page
JimDean
Posted 5/10/2019 10:50 AM (#9600 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Regarding Hybrid & Classic CumV Methods, and Master-mapping

===== Updated on 6/27/19 =====


I developed two alternate hybrid CumV Method formulae (described in detail in an archive-discussion post), which "bridge the gap" between the inter-day OBV & VPT versus the intra-day AcDst & III.

I'm implementing the "Dean Body Gap" (BG) formula in-between the positive-slider interday VPT and intraday AcDst options. And I'm implementing the other more-complex "body-gap-tail" (BGT) formula between the negative-slider inter and intra-day "Guts|TruR" versions of VPT and AcDst.

After *extensive* review, it appears that the "simple average" of all the formulae variants is also useful ... and seems to be "marginally better" as a *generic default* (+3 and -3 on slider), so I'm offering Method as well - one average for the five pos variants and another for the five neg variants. With this addition, the CumV Method slider now goes from -6 to +6.

The Master Slider now includes 6 CumMeth's in it's "tiered" permutations ... now 270. The default center = 135, which represents CumMeth=3(Avg), TimeGuru=3, VoteGuru=3, CGband=3 ... but the Dean-Meth #4 is very "close-by" ... it's #134: Meth=4(Dean), Time=3, Vote=3, CGband=3.

As far as I can tell from my SBAS of the 15 test symbols that I'm currently using, the Dean formula seems to be more consistently reliable and profitable than any single one of the other four. But it appears that maaaaybe the Avg(5) is a skosh better. It will take more rigorous testing to know. At least the new mapping of the Master has them right next door to one another ;~)

Here is the official and final list for the various CumV Methods ... the formulae are abbreviated so they can fit on the popup Help screen. Abbrev notes: absBody = abs(C-O) & absGap = abs(O-C1) ... these simply are abs() echoes of the terms in the numerator; TR & TruR and TruRng all mean the true range of that bar (not ATR) = max(H,C1) - min(L,C1). If these aren't relatively memorable once explained, please let me know.

... positive slider:
+1. Granville On-Balance-Vol: if C > C1 use +V, else -V
+2. Fosback Volume Price Trend (VPT): (C-C1) / C1 x V
+3. Mixed Avg(5+): BG & Cls|H-L w/ OBV, VPT, AcDst & III
+4. Dean BG: (2*C-O-C1) / (H-L+abs(C-O)+abs(O-C1)) x V
+5. Chaiken Accumulation-Distribution: (C-O) / (H-L) x V
+6. Bostian Intraday Intensity Index: (2*C-L-H)/(H-L) x V

... negative slider:
-1. Granville OBV w/C=>Guts: if G > G1 use +V, else -V
-2. Fosback Volume Price Trend w/Guts: (G-G1) / G1 x V
-3. Mixed Avg(5-): BGT & Gut|TR w/ OBV, VPT, AcDst & III
-4. Dean BGT: (2*C-O-C1 +(L-H-Body)/4) / (TruR+H-L) x V
-5. Chaiken AccumDst w/TruRng: (C-(O+C1)/2)/TruR x V
-6. Bostian III w/TR: (2*C-min(C1,L)-max(C1,H))/TruR x V

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

The attached spreadsheet shows how the Master Slider maps into the two Gurus and two independent Experts, offering 270 patterns (default = 135, mapping to 3,3,3,3). All six of the positive CumV Methods above are included in that set.

The attached ZIP file has a collection of charts that illustrate the distinctions between the "average-of-five" CumV default calcs and the Dean BG and Dean BGT calcs. Each has four instances of CVW ... the first is the Average (+3) of the five positive-slider methods, the second is the Dean BG (+4) method, the third is the Average (-3) of the five negative-slider methods, and the fourth is the Dean BGT (-4) method.

(Comparison of Dean vs Avg methods (15 syms).png)



Attachments
Attachments CVW Master Mapping (270 variants).xlsx (23KB - 1 downloads)
Attachments Charts Comparing Avg to Dean Methods, for 15 Syms.zip (391KB - 6 downloads)
Attachments Comparison of Dean vs Avg methods (15 syms).png (186KB - 4 downloads)
Top of the page Bottom of the page
JimDean
Posted 5/14/2019 6:14 PM (#9625 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Help-Info for XPRT positive +POWR negative +CSTM overrides, w/Param-Panel

================== Updated on 10/11/19 ==================



*** LIST OF XPRT POSITIVE-SLIDER OPTIONS *** (save)

MasterTymVotGapMeth- pos 1-270= 4-tiered-seq-map: TymFrm > VotSpec > CGgap > CumMeth
... def=135= 3,3,3,3; Outer>Tym=1,3,5; Vot=1,2,3,4,5; Gap=1,3,5; Inner>Meth=1,2,3,4,5
CumMethBasic2Refined- GranOBV, FsbkVPT, 3=avg(5), DeanBG, ChaikAD, BostNtraNtnsNdx

=== GURU_TymFrmFst2Slo- positive maps directly to positive of three Experts
MAvgSpeedPdsSm2Lrg- positive 1-5 uses identical FMS pds for Ema=>Sma+Wma & EmH=>WmH
MAvgTypSnug2Gradual- positive 1-5 = WmH, WmH&EmH, WmH&E, EmH, EmH&W calc methods
MASlopePdsLess2More- pos 1-5 = Lrs(2), Wma(Lrs(2),2), Lrs(3), Wma(Lrs(3),2), Lrs(4)

=== GURU_VotSpecTyt2Lax- positive maps to experts: w/NO SmartSize, Adds or PX,PC,PG
BgnVotNetNatcMny2Fw- positive 1-5 NetVotes+Ntry: 6t+c, 6full, 5t+c, 5full, 4t+c
... 1-5 SizePcts: All-In=140,120,120,100,80; Toe-In=60,0,40,0,20; Confirm=80,0,80,0,80
EndVotOpXmwFw2Mny- pos 1-5 OpposVotes+MWcnt: 2V2m2w, 2V3m2w, 3V3m3w, 3V4m3w, 4V4m4w
FuzBandAddCutSm2Lrg- positive 1-5 FuzzBand for neg/~0/pos evals: off,15,30,45,60%
LossCutsGapSzSm2Lrg- positive FullCut: 1-6 => incrs Gap, decrs Tytn-speed (6=off)
... InitGap ATRmult 1-6: FC= 3.2x, 4x, 4.8x, 5.6x, 6.4x, Off (FxdBrokerSM for all cases)
... FastTytn %InitGap/bar: 1-5= 4 > 2%; SlowTytn=half%, when abs(G-G1)<WTR/2
ProfGrabGapSzSm2Lrg- positive FullGrab: 1-6 => incrs Gap, decrs Tytn-speed (6=off)
... InitGap ATRmult 1-5: FG= 4x, 5x, 6x, 7x, 8x, Off (same Tytn rules as for LossCuts)

EquityFilterNone2Tyt- 1-9: min-max by varying Eqty-MaPds, StkSlpVot, Add/Ent Blk, ForcXit

BarWndoTrdDirL3cS6c- Plot=Wdth(>Hre), Rtn=Bb4Hre(9wide); Longs@300+, Shorts@600
OutputReturnHelpPlot- positive 1-5= Plot Band, B+3Strp, Eqty, B+2Hist, Detail (0=Help)


*** LIST OF POWR NEGATIVE-SLIDER OPTIONS *** (save)

MasterTymVotGapMeth- negative used for special "Wizard" patterns (needs license) ... all vers
CumMethBasic2Refined- negative uses Guts & TrueRange; -3=avg(5)

=== GURU_TymFrmFst2Slo- negative maps directly to negative of three Experts
MAvgSpeedPdsSm2Lrg- negative 1-5 "visually similar" FMS pds for S,W,E and WmH,EmH
MAvgTypSnug2Gradual- negative 1-5 = Wma, W&E, Ema, W&E&S, W&S calc methods
MASlopePdsLess2More- neg 1-5 (slower) = Wma(Lrs(4),2), Lrs(5), Wma(Lrs(5),2), Lrs(6)

=== GURU_VotSpecTyt2Lax- negative allows Adds and PX,PC,PG (req's Robust TP)
BgnVotNetNatcMny2Fw- negative 1-5 same ReqV+Ntry, but uses MOC instead of MOO orders
EndVotOpXmwFw2Mny- neg 1-5 allow logic-PExits (not PC,PG) w/Sz: 40%,30%,30%,20%,20%
FuzBandAddCutSm2Lrg- negative 1-5 same Fuzz% plus activates reducing-Adds option
... neg1-5: InitAdd%=80,60,60,40,40 (@nxtAdd-20%); Adds req: Bright w/prev Warn/PX,
... 3 prc rules, possible Pct-Lim by max-TradeSize= TotNtry+InitAdd-PXit (poss +10%)
LossCutsGapSzSm2Lrg- neg 1-5 has FullCut envelopes and also inner PartCut (-6= PC not FC)
... both LC & PG: PartXit Size same as EndVot-PXsize, PX Gap= FXgap/2; PX Tytn same% as FX
ProfGrabGapSzSm2Lrg- neg 1-5 has FullGrab envelopes and also inner PartGrab (-6=PG not FG)

EquityFilterNone2Tyt- neg: Price-vs-ATR SmartSized Entries +/-10% (via TradePlan)

BarWndoTrdDirL3cS6c- ... Plot = #BarsB4Hre, w/128 width; Rtn = width of eval, thru HRE
OutputReturnHelpPlot- neg specifies Return type (pos=Plot,0=Help) ... all versions

*** LIST OF CSTM MANUAL-INPUT OVERRIDES *** (save)

MasterTymVotGapMeth- typed #### reads 10-param tuned sets from (licensed) text file
CumMethBasic2Refined- CCCM: M=CumMethInp, CCC= CC.C CapVmult (100+ => Capping off)

=== GURU_TymFrmFst2Slo- Fltr1=PPPPPFG: G=GuruInp; F=FltrID, PPPPP=parms (Shell Indic)
MAvgSpeedPdsSm2Lrg- FfMmSs: Ff,Mm,Ss = fast,med,slow MApds (neg=same)
MAvgTypSnug2Gradual- GCT: T=MAtypInp; G=#SigFig digits; C<>0 => Hull StkV uses Wma+Ema
MASlopePdsLess2More- WwwL: L=SlopePdsInp; Www= CumWndoPds (neg=slower-slope-set)

=== GURU_VotSpecTyt2Lax- Fltr2=PPPPPFG: G=GuruInp; F=FltrID, PPPPP=parms (Shell Indic)
BgnVotNetNatcMny2Fw- tfvTFV: BgnV=7-4; FullNtr=0-7*20%; ToeNtr=0-3*20%;
... (lower-case tfv defines Short-Trade specs when WndoWdth & BarsB4hre both neg)
EndVotOpXmwFw2Mny- mwvMWVP: EndV=2-5; Med2Wrn#, Wrn2FX#; PXsz=0-5*10% (if neg);
... (lower-case mwv defines Short-Trade specs when WndoWdth & BarsB4hre both neg)
FuzBandAddCutSm2Lrg-abmABMW: W=1-9*20pds; B=0-9*10%; Add=0-4*20%, M=AddLimMeth
... M= 0-4: 0-3= decrem@Add -20%, 1-3= Limit Max#ofAdds; 4= decrem@Add -40%
... (lower-case abm defines Short-Trade specs when WndoWdth & BarsB4hre both neg)
LossCutsGapSzSm2Lrg- spfSPFT: spf for Shorts, SPF for Longs, T for both (SPFT=SLsame)
... T=FstTytn/bar 0-9= 0|1 >5% *InitGap; S|s= PartCut XitSize 0-5= EndVotPX | 10-50%
... Cut-envelopes Gap-distance 0-9 (0=off): Part P|p= .8-4x ATR, Full F|f= 1.6-8x ATR
ProfGrabGapSzSm2Lrg- spfSPFT: T & S|s same meaning as for LossCuts
... T=FstTytn/bar 0-9= 0|1 >5% *InitGap; S|s= PartCut XitSize 0-5= EndVotPX | 10-50%
... Grab-envelopes Gap-distance 0-9 (0=off): Part P|p= 1-4x ATR, Full F|f= 2-8x ATR

EquityFilterNone2Tyt- CVS: S~TimeG, V~VoteG, C=Cntl: 1-4=AddBlk, 2-4=EntBlk, 4=ForcXit
BarWndoTrdDirL3cS6c- MGggWww: Ggg=gap2hre, Www=width, M=mult for G&W #bars
... Pos= BothLS, Neg= JustL; if any sep LvS inputs, then Pos= DiffLS, Neg= JustS

OutputReturnHelpPlot- same as above (no applicable overrides)
Top of the page Bottom of the page
JimDean
Posted 5/18/2019 3:11 PM (#9633 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: Primary Help-popup Panel

==== Updated 9/9/19 ====


I'm rearranging the Help panels so that the most-useful that an experienced user might want is first, then the second-most useful, and so on ...

ACTUAL INPUTS & DERIVATIVES panel is without question the most useful ... it concisely and completely reports all the values that drive the model and are necessary to explain its behavior.

The single most important Help pane is the one that reports what the actual Input values are for each parameter ... AND provides the "derivative" values related to the Master & Guru-mapping, along with the complex Expert-rules. Those derivative values give essential insight as to how the Stradicator logic is working for a given pattern-model.

Note that the Cstm version provides really, really in-depth override-control over virtually ALL the "derivative" values that are reported here. When the overrides are manually typed-in, they can use up to *seven* digits (with a minus sign) ... this eats up space on the Help pane line, and therefore requires smart formatting to accommodate that.

Also, *many* of the "normal" slider inputs create such different derivative inputs, that a lot of fancy re-formatting needs to be done even without overrides. The idea is to provide as much info as possible, within the Help-panel width restriction, and always in the same "logical" location.

I'm attaching two snapshots ... the first is the popup when "pure-default" inputs are used; the second is the popup when virtually every Cstm-version override-able param has manual inputs. There are *many* variants in the formatting, depending on the input values, but the same info is always on the same line.

Note that many of the lines in this panel have alternative wording/formats, based on what inputs are spec'd.


(HelpDetail+Params - Default.png)



(HelpDetail+Params - Overrides.png)



Attachments
Attachments HelpDetail+Params - Default.png (123KB - 0 downloads)
Attachments HelpDetail+Params - Overrides.png (127KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 5/18/2019 6:00 PM (#9635 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: FUNCTION-CALL for OMNISCRIPT help-panel

===== Updated on 9/9/19 =====


This (attached) is likely very useful, being central to many of the Stradicator features such as Dynamic Scans, TradeLogs, and Custom Focus-Lists ... so it's the second Help pane ... and since it requires a list of the various Return-parameter choices, that list is in the panel which immediately follows it. I can't combine that list on this panel, since the "InputBox" format that provides the copy-able function-call string has a "squished" text width.

Return-value options and Plot-explanations may grow to two panes ... even an experienced user might need reminders of the multiple Return possibilities, and what the various digits mean ... and may need reminders of what the Band colors and Histos, etc mean. The Popups won't have really complete explanations ... it would be too long, and they don't permit snapshots to illustrate colors, etc. But hopefully the recap will be sufficient for all but the novice user. There will be a PDF manual for those folks ... &/or Training Videos (probably Videos before PDF).

Following those panes, there are 1-4 more panes (early samples shown in the prior post), depending on the Version that the customer is using. Easy gets one pane, Xprt gets the Easy pane plus another one with detail about the six Expert sliders, for their positive-value ranges. Powr gets yet another pane, explaining the negative-slider options for all but the Master input. Cstm gets another pane, explaining all the manual-override inputs.

A total of 8-9 panes. Of course, by clicking "No" or Esc at any point, ongoing panes are cancelled.

(Help Panel #2 - OScript.png)



Attachments
Attachments Help Panel #2 - OScript.png (199KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 5/18/2019 6:35 PM (#9634 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: Four Input-Param-Definition Panels

===== Updated on 9/9/19 =====


I've used the "crib sheet" info, unchanged but with many line-breaks added with indenting, so that info will fit on the Help pane. Later on, I will reword and reformat, etc ... but at least for now, it's all there. Snapshots show current status (not yet complete).

The arrangement is nifty ...
Help Panel #3 will appear only for Easy version
Panel #4 for all but Easy (dup's #3 info)
Panels #4 & #5 for Powr and Cstm
Panels #4, #5 & #6 for Cstm

"Crib Sheets" for this & Param-pane: http://tradetight.org/forums/thread-view.asp?tid=1445#M9625


(Help Panel #3 - Easy.png)



(Help Panel #4 - Positive.png)



(Help Panel #5 - Negative.png)



(Help Panel #6 - Overrides.png)



Attachments
Attachments Help Panel #3 - Easy.png (248KB - 0 downloads)
Attachments Help Panel #4 - Positive.png (347KB - 0 downloads)
Attachments Help Panel #5 - Negative.png (334KB - 0 downloads)
Attachments Help Panel #6 - Overrides.png (433KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 6/26/2019 2:00 PM (#9556 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Regarding CutGrab "Envelopes"

===== Updated on 6/28/19 =====


PricePane will show *up to* two virtual-exit lines below price, and *up to* two above
... "virtual" = based on Close-price crossover, not H/L
Inner line is for Partial exits, outer line is for Full exits
Exits from lines above called "Grabs" (ie grab your profit)=PG|FG
Exits from lines below called "Cuts" (ie cut your loss)= PC|FC
... Partial|Full exits triggered by Stradicator-specific logic are still called PX|FX
Distances of the lines away from price controlled by CutGrabGapSzSm2Lrg slider:
... an independent expert, NOT tied to either GURU, but part of the Master mapping
... negative slider = turns off Partial Cut (PC) and Partial Grab (PG) inner lines
... FullCut (FC) & FullGrab (FG) "outer virtual" lines show for +1 > +5 slider values
... slider-inputs of 6>10 (pos or neg) turn *off* the Grabs lines, but otherwise same as 1>5
... this slider alters the gaps between price and envelope thresholds, not the size-Percent
Fifth (most-distant) line is the *BROKER* SM order (BC), on the "Cut" side of price
... Broker StopMkt orders assure protection even if OT|feed fails.
... these will trigger based on Long/Short= L/H price crossing the threshold (same bar exit)
... Simple-TP only adjusts when Confirm fires ... otherwise, a flat line
... Robust-TP will ratchet-adjust (never wider) the BC gaps after each new Add (hardcoded)
... the ATR-multipliers for these gaps will be *locked* (no override), since hardcoded into TP
Envelopes can be totally deactivated by manual input of 0.x for this param (not recommended)

PartCuts, PartGrabs, FullCuts and FullGrabs will use "virtual" tradeplan orders:
... "virtual" = internal calc's check "completed" bar and fire the order when the level is crossed
... ... crossover Price will be either Close or Guts (not H|L), to assure complete Fills
... ... Guts crossover used for MoO order models; Close crossover for MoC
... uses "smart time gaps" ... tightening a bit, for each bar that doesn't improve Profit
... these allow Cstm user to *override* their initial-gap ATRmult's, independently

Two TradePlan types:
... Simple TP that supports Full or Toe+Cnfrm entries, and the FullGrab & FullCut line exits, plus standard Stradicator-logic FullExits, but not Adds or PXits of any type (Strad PX, PC, or PG).
... Robust TP that supports all Entry, Add and Exit types (VERY long ... thousands of Conditions)

When a PartialCut fires due to Close crossing a threshold-line, it will "act like" a Warn, in relation to the Med2Wrn count reset rule, and the Wrn2FX count rule ... that is, if a cross of the inner PC line occurs when the Wrn2FX count was one less than its threshold, then that PC cross will trigger a FullExit rather than a PC exit.

When a FullGrab or FullCut fires, due to the Close crossing one of the the outer virtual-profit-lines, it ends the trade. Alternatively, if a BrokerCut fires, due to the L(long) or H(short) crossing the most-distant Cut line, it ends the trade.

When a PartGrab fires, due to Close crossing the inner profit-line fires, it will be treated in the internal logic MW-counts the same as a "bright/strengthening" bar would be. If the bar that crosses the PG line might otherwise have triggered an Add, the Add is cancelled (it would have been a poor pricepoint for an Add, anyways).
Top of the page Bottom of the page
JimDean
Posted 6/26/2019 3:36 PM (#9640 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: CutGrab Envelopes on PricePane

===== Updated on 6/27/19 =====


Here's a snapshot of the various Cut/Grab "envelope" lines. The -3 input CutGrab-param input (shown) has them all active. +3 input removes the innermost red & green partial-exit lines. -8 removes the ProfGrab lines and retain all the LossCut lines. +8 removes both PX and ProfGrab FX line.

To completely deactivate the CutGrab lines (any version), the user must manually type in "0.x" (x <> 0). This is an important feature of the Stradicator paradigm, so I've made it "difficult" to turn completely off.

The farthest pinkish "flat" line is the fixed-loss full-exit *Broker* stop ... which will eventually "stairstep-trail" price when Adds occur (maybe also when PXits occur), if using the Robust TP.

The lines turn yellowish (diff shades) for the final bar of trade, so that if a new trade begins on the very next bar, it's clear where one ends and the next begins.

Note that if the pinkish Broker line is crossed by extreme H/L price, the trade ends. But all the other lines only fire their related PX or FX if the Close price crosses the line (I'm considering maybe making that Guts price ... which is shown by small black dots on the candles).

If a new initial-Entry is just being Signalled at the HRE (not shown in snap), a set of bluish (diff shades) one-bar horizontal dashes appear "ahead of time" (ie before you place the first order), to illustrate the initial gaps, even though the trade may not yet have started (MoO orders). These indicate where the initial Broker stop will be, and any active FX/PX virtual stops.

At this point: the Partial Cuts are not yet tied into the Warning-counts; Full Cuts & Grab crossovers don't yet end the trade; extra colors on Band for envelope-line crossings have not yet been implemented; once I get the PX/FX actions tied in, I will likely revisit the expert-mapping to the Gap sizes &/or tightening. That's coming soon.

(CutGrab Envelopes (not tied to Sig or Band yet).png)



Attachments
Attachments CutGrab Envelopes (not tied to Sig or Band yet).png (26KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 6/27/2019 3:27 PM (#9645 - in reply to #9640)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: Collected Info re Equity Stats Outputs

===== Major Update on 10/10/19 =====


I'm not satisfied, for a variety of reasons, with any of the "standard" single-value measurements that are available in OT (or other packages, afaik). So, I've created a sophisticated new set of metrics, culminating in 2-3 values, any of which that I suspect most users will employ for OmniScan &/or Sorted-Focus-List selection of Symbols that are "best current personalities" for a given Stradicator's input-pattern. None of these "require" the Equity-Filtering feature to be turned on ... but if it is, all of the stat's below will reflect that impact, and likely will be even more useful.

Metrics in the "TLS" TradeLog file, for each Symbol analyzed

BgnYmd (+HM) of first Trade (often after start of test window)
EndYmd (+HM) of final Trade (often before end of test window)

TotBars = test window output Width ... defined in filename, by 2nd-to-last input
PctActv = Total In-Trade Bars / TotBars ... ie "Trade-Engaged Bars"

TotTrds = Total Trades in test window (includes any open trade at HRE)
BndBnT = ( MaxBinT - MinBinT )/2 ... roughly, BarWnd/TotTrds +/- BndBnT is "typ trade length"
HitRate = Number of Trades with positive net Profit

SumProfT = Sum(TradePL,Winners) ... TPL normalized% for ClosePrc & Size% of Basis
SumTrdPL = SumProfT + Sum(TradePL,Losers) ... easy to derive SumLossT & ProfPerTrd from this
ProfFac = HitRate * SumProfT / [ (1-HitRate) * abs(SumLossT) ]

AvgTrdDD = Sum[ Trades' (PeakB4maxDrop - MinB4newHigh) / PeakB4maxDrop ] / TotTrds
BndTDD = ( MaxTrdDD - MinTrdDD )/2 ... roughly, AvgTrdDD +/- BndTDD is "typ trade maxDD"

... next two are "specialized" measures of Reward vs Risk, with different focus ...
SPLvsMDD = SumTrdPL / MxEqtyDD ... focus= worst sustained loss (~Calmar w/o compounding)
HolmarkRR = f(HR, SumWins, SumLoss, NT, SumPL, PPT, AvgLoss) ... Mark Holstius' RTM-focus

BndATQ = ( MaxTrdQ - MinTrdQ )/2 ... roughly, APLvWgTQ +/- BndATQ is "typical Trade Quality"

... next six all are different measures of overall Reward vs Risk, using alternative perspectives ...
(any of these might be good choices for OmniScan searches of "best FocusList to use today")

APLvsWTR = AvgProfitVsDrawdn = SumTrdPL / TotTrds / AvgTrdATR ... ie "Trade Rwd/RskTR"
NgdPLvTR = APLvsATR * PctActv ... ie "Engaged RR" = AvgProf vs AvgATR, factored by %ActvBars

APLvsADD = AvgProfitVsDrawdn = SumTrdPL / TotTrds / AvgTrdMDD ... ie "Trade Rwd/RskDD"
NgdPLvDD= APLvsADD * PctActv ... ie "Engaged RR" = AvgProf vs AvgTDD, factored by %ActvBars

APLvWgTQ = AvgTradeQuality = Avg( { TrdPL*[ 1/(ATR*3) + 1/TDD ] /2 } / BarsInTrd, NT)
NgdPvWTQ = APLvWgTQ * PctActv ... ie AvgTQ factored by "Engaged" test-window %ActvBars

Note: ProfLoss, MaxDD, and TradeQuality *for each Trade* will be output in the TLB file ... where a trade's "TradeQuality" = NetTradeProfitLoss * [ 1/(ATR*3) + 1/TradeMaxDrawDown ] /2

CAGR & Calmar ... metrics which require Account Size &/or including Compounding for their calculation aren't meaningful for single-Symbol analysis, so they are provided via StradFolio calc's.
Top of the page Bottom of the page
JimDean
Posted 6/28/2019 2:14 PM (#9647 - in reply to #9645)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: Color-Key for the Stradicator-Band

===== Updated on 6/29/19 =====


This is an important thing to keep handy. The 27 (!) colors on the wide horizontal Band that appears on most of the plot alternatives "tell a story" about how things are progressing during a "Wave" ... they report what Actions occur (ie orders), and something about why they occurred. Also, the hues hint at what's going on with the NetVotes that drive much of the Stradicator logic.

The sequence of colors is rationally arranged into "families": greenish = Bull, reddish = Bear, black & dark-blue/purple = Between, light-medium blue = Bull Toe-In/Alert, and light-medium violet = Bear Toe-In/Alert. It's taken a LOT of fooling around to pick the best colors ... many of the choices have been driven by what colors might be "next to" one another. Some of the hues are very close ... especially the darker ones ... but usually the distinctions between them aren't very important since they are reporting a similar Action (such as Full Exit).

Note that the sequence goes from the "earliest > latest" Bull events from the Top-Down, and from "earliest > latest" Bear events from the Bottom-Up.

The +/- #'s on the right are for internal-coding use only ... I'll delete them for the help manual. And I might change a few of the descriptions, which are not intended to be fully explanatory. But for the time being, this should help y'all a LOT when viewing the snapshots. Later, I'll provide similar charts (fewer color variants) for the three thinner Strips below the Band, and for the other plot options which have very different formats.

6/29 first Update: The seven asterisks after Bull|End|Bear labels designate Stradicator-logic that is NOT in play for "pure-filtering" mode ... that is, when the Stradicator is used as a Filter rather than a Strategy.

6/29 second Update: Bull and Bear Add-On events had been left out! ... now included as +13/-13, next to the"Norm-Entry" events ... gainsboro and silver colors.

(And btw ... if you're somewhat color blind, I'm sorry but there's just no combo that I have found that would be an alternative ... refer to the chart earlier in this thread for the full set of colors that are available to use).

(Strad Band Colors (CVW).png)



Attachments
Attachments Strad Band Colors (CVW).png (113KB - 0 downloads)
Top of the page Bottom of the page
JimDean
Posted 6/29/2019 8:04 AM (#9648 - in reply to #9647)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: Plot Format Output Samples

===== Updated on 6/29/19 =====


There are currently five (+1 > +5) slider options for plotted output, shown below. It will be expanded as plotting features change.

The FINAL FIVE samples (plot options 1-5) use default positive-slider settings, w/ ToeIn+Confirm MoO orders, with no Adds|PXits. The FIRST sample (plot option 2) shows default settings but with *negative* values toggling on MoC orders, Adds, PXits, and Partial Cut+Grab envelopes.

Powr|Cstm versions offer various negative-sliders that can activate Adds (white dots), PartialExits (orange dots), PartialCuts (dim red inner envelope), and PartialGrabs (dim green inner envelope) ... as well as some additional colors on the WaveBand that echo Adds and envelope-crossovers.

Equity Status #3 is just a "placeholder" for now ... it will have an Equity curve, 2-3 MA's of that curve, and the standard WaveBand across the zero-line ... the Y-axis will be dynamic due to Equity variance ... it will show either the Equity related to the "raw background" Strad trades (if BarWndo input is positive), or the Equity related to the "actual Equity-Filtered" Strad trades (if BarWndo input is negative ... Powr|Cstm versions).

(Plot Option 2(neg) Add, PX, MoC.png)



(Plot Option 1- Just Band.png)



(Plot Option 2- Band & Strips.png)



(Plot Option 3- Equity Status.png)



(Plot Option 4- Full Vote Detail.png)



(Plot Option 5- V, CumV, MAs.png)



Attachments
Attachments Plot Option 2(neg) Add, PX, MoC.png (26KB - 0 downloads)
Attachments Plot Option 1- Just Band.png (20KB - 0 downloads)
Attachments Plot Option 2- Band & Strips.png (23KB - 0 downloads)
Attachments Plot Option 3- Equity Status.png (22KB - 0 downloads)
Attachments Plot Option 4- Full Vote Detail.png (29KB - 0 downloads)
Attachments Plot Option 5- V, CumV, MAs.png (34KB - 2 downloads)
Top of the page Bottom of the page
JimDean
Posted 8/17/2019 6:53 PM (#9669 - in reply to #9648)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Easy, Xprt, Powr, Cstm and optional Equity-Toggle features

===== Updated on 10/10/19 =====


This post defines the distinctions between the CVW versions, without much detail ... as a crib-sheet to help distinguish them, and later, to help explain pricing differences.

All Versions
* Master Input #1 neg values turn *ON* optional Equity-Toggle feature *if* it's been licensed
* Manual-Input optional robust Filter spec via Guru's (plus Guru value) *if* it's been licensed
* Output-control Inputs (final two on param pane) can be either negative or positive

Easy Version
* Seven Inputs (no Experts for the two Gurus)
* Inputs 2,3,4,5 only allow positive values (CumV, GuruTym, GuruVot, CutGrab)
* Plot options +1 > +3 avail; Return options -1 > -3

Xprt Version
* Fifteen Inputs (all Expert sliders exposed)
* Only positive range for inputs/sliders, except Master and two Output-control (like Easy)
* Plot options +1 > +4 avail; Return options -1 > -10

Powr Version
* Fifteen Inputs (same list as Xprt)
* ALL sliders offer negative ranges, providing many additional, more-sophisticated controls
* One Guru & five Expert sliders offer "integer.decimal" Long.Short evaluation-distinctions
* Plot options +1 > +5 avail; Return options -1 > -10

Cstm Version
* Fifteen Inputs (same list as Powr, with positive and negative ranges & L.S decimal options)
* ALL sliders offer highly-detailed manual-input Override inputs to "tailor" the Expert rules
* Plot options +1 > +5 avail; Return options -1 > -17
Top of the page Bottom of the page
JimDean
Posted 8/20/2019 7:59 AM (#9707 - in reply to #9669)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
There are several particularly "cool" features within the Stradicator paradigm, which strictly speaking are not *necessary* for it to do its thing, AND which require a *significant* amount of extra coding, and which add (imho) *significant value* to the tool.

Some users may not feel that they "need" all of these ... or they might be averse to dealing with the related price-premium that the extra development work has warranted. Note that as part of the "paradigm", these features will be available for *all* of the Stradicators.

The TradeTight-Keys license-file has the flexibility of setting "flags" which can tell the routine to allow or to block various features. There are five flag-options available - a client can purchase none, all, or any combo of them ...

===== Updated on 10/4/19 =====

1. Return-Values Subscription ... a feature that all but "pure-discretionary" traders will want, and is paid for by Subscription (required for EqtyStat-Optimized Scans, automated TradePlans, custom Focus Lists, and TradeLogs gen'd from full FocusLists).

The other features are also optional, for an additional *one-time* fee ... in any combination ... none of these "require" any of the other options (but all live features do interact):

2. Equity-Filtering ... if licensed, this activates Equity-Curve voting-logic, which only allows trade-signals to create orders if the recent-past performance of *that* Strad-pattern on *that* Symbol has shown acceptable P/L performance. This is a biggie ... it's a perfect companion to the Equity-Stat Optimized-Scan (though it does not require that to be used), since it can prevent *ongoing* (future) trading of a Symbol which "passed" the historical Scan rules, if that Symbol's personality changes during a trade, or before a new Scan recognizes it.

3. Guru-Filters ... if licensed, the user can input a "filter-spec" in either or both Guru param's, which "calls" an external routine that contains custom OLang rules ... or calls to other Filter routines ... which apply "filter-block" logic *within* the Stradicator. This is extremely powerful, and allows for multiple filtering ... essentially it's a microcosm ATM. If Filters are being used for actual trading, and if Equity Stats are used for Scans or the Equity Filter (#4) is used, then the Filters *must* be modeled using this method, for truly representative application.

4. Trade-Logging ... if licensed, then dedicated input value(s) for the Output param will tell the Strad to create a Log file of all the signals and order details (which the popup Advisor would report) in the calc'd range for the current symbol ... if Return-Values Subscription is also active, then a TradeLog can be created for all the symbols in a Focus List at once, with one click.

5. Master-Wizards ... this licenses ability to utilize "refined" Master-patterns, drawn from a text file. These would likely be derived from StradWiz analysis of a collection of TradeLog runs by the user. I hope to provide TT models as well, derived from some massive analysis studies that I hope to do later, which the user can optionally purchase.
Top of the page Bottom of the page
JimDean
Posted 9/8/2019 10:49 AM (#9703 - in reply to #9642)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
PMAeval Stradicator Features and Logic
(Click here for intro discussion about this in TradeLog thread)


Five Input-Sliders, arranged to roughly echo typical Stradicator parameters:

one @ 1-6, specs EMApds, WTRmults & Fuzz (Fast>Slow EMAs; 1,3,5=Tight & 2,4,6=Normal)
one @ 1-6, where 1,3,5=MoC | 2,4,6=MoO; 1-2=Just Normal, 3-4=All, 5-6=Just Toe+Confirm
one @ 1-4, Signals: All | NoAddOns | NoAddsPXits | NoAddsPXitsBroker (later, neg=EqtFltr)
one @ -256>+256, to spec Width & LastBar (and TrdDir: ShortOnly if >600, LongOnly if >300)
one @ -2>+2 to spec Return|Help|Plot Output (-/+1=Log); later, -/+3=Equity Stat/Curve

The Plot will show four curves: Close and 3 EMAs on the Price chart (each tri-color to denote slopes: green=up, blue=fuzzy-0, red=down). Also shows envelopes related to WTR-gap threshold Cut|Grab rules (colors spec'd below). An Indicator pane will show Signals via a histo (color indicates signal type, pos/neg indicates buy/sell, and height indicates trade %size), plus a line that tracks each Trades's net position-size%. Return value will showing Signals as integer, and %Size for the signaled order as decimal. Later, I'll incorporate Equity eval and Stats, which will provide an extra Plot and extra Return option.

This has much of the broad-brush functionality of the Strad paradigm, as it affects how Trades are managed and is a viable test-engine to generate TradeLogs for the StradFolio sheet. Return values will allow rudimentary Strad-paradigm FL-manual-mechanical trading (later, automatic, via special TradePlans) and also allow Dynamic-Scan optimal-Symbols-search.

The rule-descriptions below are for Long trades ... there will be mirror-rules for Shorts. Believe it or not, these rules are much, MUCH simpler than any of the Stradicators ... but they have roughly similar logical architecture behind them ...

This logic is processed "sequentially" ... if a given rule is met and creates a Signal, then the ones following it are ignored for that bar.


... Initial Entry rules (either Normal single-entry, OR small ToeIn + possible Confirm):


1a. ToeIn entry 40% of Base-size when this full set of rules *initially* becomes true: Med EMA >= Slow EMA (ie Bullish Trend), Fast EMA >= Slow, Fast <= Med, positive LRS(Close,3) slope, and at least one of LRS(Fast,3) or LRS(Med,3) slopes are up.

1b. Confirm entry adds another 80% of Base-size sometime after a ToeIn entry, but only if no Partial-Exit has occurred since the ToeIn, when the "Normal entry" conditions are met. (see #2 below)


2. Normal entry 100% of Base-size if a ToeIn has *not* fired yet, when all these rules initially become true: Close >= Fast, Fast >= Med, Med >= Slow, and LRS(3) slopes of at least 3 of the 4 (Slow, Med, Fast, Close) are up.


... Once trade has started (Toe, Toe+Confirm, or Normal entry) ... check rules in this sequence:


4. Broker (SM-Loss) Full-Exit any time the Low crosses 3xWTR below the most-recent Entry price (Toe, Confirm or Norm) ... pink envelope


5a. Full-Cut Exit any time the Close is below the all the prior-bar EMAs, AND 3 of the 4 current-bar LRS(3) slopes are down ... red envelope

5b. Full-Grab Exit any time the Close is >= 6xWTR above the max of: LLV(C1,3), all the prior-bar EMAs, and the most recent Entry price ... lime envelope


6a. Partial-Cut 40% Exit any time the Close drops >= 3xWTR below the min of: HHV(C1,3) and prior-bar FastEMA ... orange envelope

6b. Partial-Cut 20% Exit any time the Close is below the min of: most recent Entry price and the MedEMA, AND the LRS(3) slope of the FastEMA is down.


7a. Partial-Grab 40% Exit any time the Close is >= 4.5xWTR above the max of: LLV(C1,3), all the prior-bar EMAs, and the most recent Entry price ... green envelope

7b. Partial-Grab 20% Exit any time the Close is above the max of: all prior-bar EMAs and all prior Highs, AND the LRS(3) slopes of 2 of the 3 EMAs are down.


8. Add-Ons Rules and Sizing: (many trades have no AddOns; very rare to have 3 in one trade)

a. Prior Signal was a Partial-Cut (=> pullback), AND Close is above latest Entry & FastEMA, AND Close is 1.5xWTR above LLV(C1,3), AND Close is below HHV(C1,6), AND 2 of 3 EMAs are sloping up. (note: these rules are deliberately tough to satisfy)

b. Sizing: First Add-On is 40% greater than prev Partial-Cut size; Second Add-On is 20% greater than prev Partial-Cut size; Third (last allowed) Add-On equals the prev Partial-Cut size. (note: larger Add-Ons early in trade encourage continuance; limited # of total Add-Ons, and smaller ones later in trade help prevent Adds just before trend-ending reversal)


Notes:
A. EMAs use Warmup of 3*SlowMApds #bars (to assure that early Signals remain stable)
... if first input = 1-3 (tight), Spread between Fast & Med and Med & Slow EMAs is tighter ...
B. "WTR" for this purpose is the WMA(ATR(1),MedMApds) ... keeps TR-basis reasonably current
... if first input = 1-3 (tight), then all WTR-mult rules above are cut by 1/3 (6x>4x, 3x>2x, etc) ...
C. Slope="fuzzy-0" if abs-val is <= 40% of WMA(slope,MedMApds) ... 20% if 1st input=1-3 (tight)
D. Only one Partial Cut or Partial Grab allowed per bar, even if both a & b rules are met
E. A second Partial Cut or Partial Grab on the very next bar will *only* be acted upon if its rule differs from the prior one
F. Partial-Cuts|Grabs based on the SAME rule must be separated by at least two non-Signal bars
G. Internal calcs for MA's & LRS use OL function-calls rather than more-efficient iterative coding
H. Width+LastB input uses (updated) Stradicator-paradigm rules (neg is mirror of pos input):
... Pos BarWndo for plot = width of eval, thru HRE; Neg BW for plot = #BarsB4Hre, w/128 width
... Pos BarWndo for rtn= #BarsB4Hre, w/9-bar width; Neg BW for rtn = width of eval, thru HRE
... Trade-Directions allowed: Input 0-300 = Both; Input 300+ = Long-Only; Input 600+ = ShortOnly
... Override (manual typed input) to model much older &/or different-width output-windows:
... ... +/- MBbbWww: Www=WndoWdth, Bbb=#Bb4Hre, M=mult for W & B (Neg=LongOnly, pos=BothLS)
Top of the page Bottom of the page
JimDean
Posted 10/13/2019 3:25 PM (#9760 - in reply to #9707)
Subject: CVW Reference Information



Owner/Admin

Posts: 3861
200010005001001001002525
Location:
USA: GA, Lawrenceville
Re: TradeLog Use & File-Mgmt Popups (updated 10/15/19)

Every time the TradeLog routine creates, renames, or replaces a file, an appropriate popup-message appears. No message appears most of the time, when TradeLog is just adding a new Symbol to an already-extant log file ... so you won't be "bothered" by repeated unnecessary popup-interruptions.

A pair of TLB & TLS files is created for every unique combination of current bar-length (ie daily, 30min, etc), Stradicator, and Input-Pattern. If you change any of those settings ... for instance, changing just one input of the current Strad with the current timeframe ... then that recalc of the Symbol is Logged into a FRESH pair of differently-named TLB & TLS files.

The seven-digit "PatternID" that's part of the filename is automatically generated from the combo of input-param's you've chosen (excluding the final "output" ReturnHelpPlot param) ... it's used to uniquely identify that Pattern of inputs (for that Strad) in a "short-form" ... if all the actual inputs were concatenated, the resulting PatnID-portion of the filename would be a minimum of 27 characters (instead of 7) ... and the max length (with full manual overrides) of *just* that PatnID could be 140 characters! So ... if you initially balk at the "odd" seven-digit part of the filename, consider how awkward the alternative longer-filenames would be!

That 7-digit autogen'd PatnID is designed to be unique, but it's possible that very rarely, an ID might be gen'd that has been used before on your machine, but with a *different* actual set of param-inputs. The TradeLog routine checks the fully-expanded list of Strad input-params in the second record of the TLS file to determine this. If an existing file which uses that 7-digit ID is found to have different param-inputs on TLS rec#2, the routine will *automatically* find an alternate (unique) ID, and *rename* all related prior files which use that ID to the alternate name ... and the processing for the *current* param-inputs instance will use the "standard" PatternID so that this process doesn't need to be repeated. This all happens *very fast* ... you get a single popup message to report what's been done, and the process drives on.


You're not allowed to add the same Symbol twice to the same file, either via a Strad-instance that's plotted on a chart, or via a Strad-call that's assigned to a custom FocusList column. If the recalc for that Sym on is repeated for the same timeframe and Strad+PatnID, then duplicating its results in the TLS & TLB logs would be a waste of time (and filesize) ... and would confuse/bias the StradFolio and StradWiz spreadsheet analyses which use the TLB & TLS files.

If an attempt is accidentally made to re-log the same Symbol with the same FileID settings, then a popup appears which gives you three options on how to proceed. Usually this popup only appears once (unless you accidentally repeat the re-log attempt). The third popup-option (Cancel) is provided to prevent those repeated popups, but should *not* be your normal response. If you accidentally doubleclick on a charted Symbol that's already been logged, then respond "No" to the Popup, which skips past your "boo-boo", but allows other (new) Symbols to be calc'd and logged.

There is an important purpose for the "Cancel" option, but it "shouldn't" be needed, if you follow the one main rule for TradeLogs: Never turn on TradeLog output when there is a Live Realtime Feed. If you have RT feed active, then a continuous stream of price-prints is occurring at the HRE for all symbols in the FocusList ... willy-nilly ... for some more often than others. *Every* time a new live-feed price-print hits the HRE for a given Symbol, OT recalc's the *entire* history for that Symbol ... and if Logging was turned one, *all* that info would be added to the Log files. Over and over again, for every symbol. Clearly, not what anyone would want to do!

Unfortunately, OT does not provide a way for the TradeLog routine to "know" when an RT feed is active ... so if it is, each time a new print hits for a Symbol that's already in the Log files (whether the Strad is plotted or in a FL column), the dup-Sym check described above occurs, and the Yes/No/Cancel box pops up. If you choose the Yes or No responses, the popup will reappear again and again, very rapidly ... you'd have to use Task Manager to stop the infinite loop by aborting OT ... ARGH!


BUT ... good news ... Task Manager is *never* necessary if you accidentally "break the rule", since you can simply press the popup's "Cancel" button if that happens. That button creates a special "TLA" file which acts as a "flag" for future TradeLog runs ... any time TradeLog tries to store info for any Symbol, triggered by a Stradicator + Pattern-ID using the same BarLen as the TLA file ... it "sees" the TLA file and skips the info-dump request. Forever ... for that BarLen, Strad, & PatnID. If you change your mind and want to add info from new Symbols to the TLB & TLS files that the TLA is "blocking", just go to the folder using Windows Explorer, and manually Delete the that TLA file. That is the *only* manual operation you'll ever need to do ... and that's only if you change your mind.

The TradeLog output will be requested by setting the Plotted or FL-column Stradicator's final Output ReturnHelpPlot parameter to a special value (still deciding what the value should be). As soon as the change is registered (via OK), the Strad-recalc starts Logging, and it stays active until you change the Output param to a non-TL setting. Any time you double-click a Symbol, OT recalculates the chart and/or the FL column entry for that Symbol ... and if the Strad in either of those places has Logging active, it will save data to the TLS & TLB files.

OT "remembers" that setting until you manually change it. If you were to switch Chart Templates, or change Profiles, or shut down OT without first turning off the TradeLog parameter, then the next time you restart OT &/or return to the Profile with that FL column, &/or re-select that Chart Template, the Stradicator will try to "continue" its Logging. If the continuation occurs for a Symbol that has already been logged, then the "duplicate Symbol" popup (see below) will appear, and you'll be able to skip over it (or archive or abort).

However, if you have *shut down* OT for the day (or you just leave it running overnight) ... and you restart it the next day, and you return to the Profile &/or Chart Template that has the Stradicator loaded which *still* has it's Logging flag active ... then the TradeLog routine *notices* that it's been a day since you last saved anything to the file, and it politely pops up a "reminder" to you that Logging is still turned on ... so that you have an opportunity to Abort before adding more Symbols to the Log file without even realizing it (see popup below).

Under any circumstances where you've just turned on Logging ... or where it was already turned on and you just changed another Stradicator-model parameter, or all the param's are the same but you just changed the current timeframe/barlength ... in any of those situations, then the Logging filename spec will be different than before. In the process of doing that, you may have forgotten that Logging is active (it's efficient and quiet in the background) ... so, rather than blithely assuming that the new BarLength &/or Pattern-ID warrants a new Log file to be created, a friendly popup will appear (first one in the snapshot below), asking if you want to do the Logging.

This may all seem like a complex and possibly worrisome explanation ... but it's actually a very natural and sensible "flow" when you're using the routine. TradeLog has been designed to be both "smart" and "courteous", without "annoying" you without good reason. This post has explained the reasons why the popups appear, so that when you see them, you'll be *thankful* for them ;-)

Finally ... regarding the mentions of "manual deletion" of the TLA file to "unblock" logging for a given FileID (mentioned above and in popup snapshots) ... the Stradicator input-paradigm will *most likely* include a means to easily delete the TLA file without having to use Explorer ... probably by adding a decimal digit to the Output param.

(TradeLog Popups.png)



Attachments
Attachments TradeLog Popups.png (1515KB - 0 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.