Spydertrader's Jack Hershey Equities Journal

Discussion in 'Journals' started by Spydertrader, Sep 25, 2004.

Thread Status:
Not open for further replies.
  1. Sevens Ones Zeros
    MTEX TASR DPTR
    AEN NIHD CNCT
    TNP JUPM EAGL
    UTHR HDWR INSP
    TKC URBN CREE
    SIE RIO DRL
    LPX FTI TKC
    PENN DECK CFC
    PMTI NAVR HTCH
    DECK CRDN YHOO

    Hot List:

    TASR
    DPTR
    CRDN
    PMTI
    NAVR
    JUPM

    Watch List:

    TASR - Du Cycle - Du5
    HRT - Du Cycle - Du5 - Du10
    ESMC - Du Cycle - Du5
    NVEC - Du Cycle - Du5
    UBET - Du Cycle
    IDSY - Du Cycle - Du5
    CALM - Du Cycle
    ANIK - Du Cycle
    SWIR - Du Cycle - Du5 - Du10
    TRMM - Du Cycle - Du5
    HANS - Du Cycle - Du5 - Du10 - Du20
    GMAI - Du Cycle
    ENWV - Du Cycle - Du5 - Du10 - Du20
    LIFC - Du Cycle - Du5
    NAVR - Du Cycle
    SMTS - Du Cycle - Du5
    ASKJ - Du Cycle
    HLEX - Du Cycle - Du5 - Du10 - Du20

    (HRT & NVEC fall utside normal float parameters)
     
    #181     Oct 10, 2004
  2. Volume on ESMC came close to signaling a possible breakout, but failed to follow through by our 11:00 AM deadline. In fact, volume and price have declined since ESMC reached its high this morning. The majority of other stocks in our Watch List contained even lower levels of trading volume compared to their respective averages. As a result, the system did not generate any trade signals for today.

    After reviewing the weekend posts, I noticed I failed to report the sale of HANS on Friday, 10-8-2004.

    10/08/04 10:51:15 25.33 25.31 25.36 200

    You may recall we sold 200 shares (half our holdings) when the price of HANS reached a 5% gain. We held the remaining shares (cost basis $25.10) for the total of four days (our time limit for this journal). Since we neither reached our target price, nor our stop loss price, we consider the trade 'breakeven' (or 'wash' under Hershey Terminology) for score keeping - even though the trade itself resulted in a small gain.

    However, failing to post the results of this trade in a timely manner creates an 'appearance' of 'hindsight' trading rather than real-time actual results. In an effort to avoid any perception of impropriety with the results, we will NOT include the results of this trade when tabulating the overall performance of the methods used in the journal. Since we moved our stop loss price higher (when price reached 5% gain) to 'breakeven' plus commissions ($25.15), feel free to use that price point for record keeping to avoid any notion of skewed results.

    Again, I apologize for the delayed post. Sometimes real world priorities require attention to the detriment of 'virtual' responsibilities. I hope you find the above information useful.

    - Spydertrader
     
    #182     Oct 11, 2004
  3. This version includes a specific request to score the final universe. A "Final Universe Sheet" has been added to score each ticker. The only required inputs for this sheet is the ticker. All other information will be automatically filled in. The scoring here is subjective. I'd appreciate if someone could verify the following logic which I have used to score.

    If Today's Price (Close) > 1.01*Yesterdays Close then
    Price Score = 1
    Else
    Price Score = 0


    If Today's Volume (Close) > 1.1*Yesterdays Close then
    Volume Score = 1
    Else
    Volume Score = 0


    If Today was in Accumulation then
    AD Score = 1
    Else
    AD Score = 0

    *I thought I recalled that A/D had to be greater than .3 to be considered accumulation, otherwise 0-.3 indicates a change from accumulation to distribution and vice versa??? Y/N?

    To run the macro, select the ScoreFinalUniverse macro...

    REGARDS & NJOY!
    G33M4K the Newb
     
    #183     Oct 11, 2004
  4. Quick question about PVA/D concept. As far as my NLP for PV, this has become very clear how change in V is a precursor for the action in P. Now I'm looking to complete the picture with A/D. I'm pretty sure that I've done too much reading of other's idea on A/D and hence, I've developed a mental barrier which is denying the fact that distribution occurs during Price increases. ARGGHH! So, I'm taking a more mathematical approach since the semantics of math are unambiguous, at least at my low level it is. A curiousity that I was trying to work thru is whether there is a 1st and 2nd order derivative (dP/dt and ((dP/dt)/dt)) to prices which can be applied here. I could crudely approximate by using (Price Close - Prev Price Close)/ 1 day but I haven't found a single google shred validating this... The PVA/D relationship has the smell of derivatives given the fact that changes in the variables precede the action of the slower variables. My curiosity is whether I can look at P, V, A/D as different fuctions over time along with their respective 1st and 2nd derivatives as I'm sure the derivatives will allow us a means to anticipate the price action. Some will argue this is not possible and perhaps even nonsensical. However Stock Price action is stochastic, hence stochastic calculus can be applied if need be (this should be unnecessary tho since we are only anticipating over an extremely brief window to position ourselves correctly in front of price action). Am I overengineering this (ie. creating too many tools)?

    G33M4K the Newb
     
    #184     Oct 11, 2004

  5. I've coded the math but you won't find it in v0.17 since I haven't figured out how to properly apply these items. Once I've worked thru the proper use of these items, I'll add these variables.

    G33M4K the Newb
     
    #185     Oct 11, 2004
  6. Sevens Ones Zeros
    ECSI RIO UTHR
    ENWV TASR CTSH
    CTSH HDWR ENWV
    UTHR CRDN DECK
    CREE URBN CHTT
    GIVN DPTR CYTC
    LPX CFC CREE
    INSP HRT GIVN
    ALDN PENN LPX
    TRMB FMT

    Hot List:

    TASR
    HRT
    ALDN
    DPTR
    CRDN
    ENWV

    Watch List:

    TASR - DU Cycle - DU5
    HRT - DU Cycle - DU5
    NVEC - DU Cycle - DU5 - DU10 - DU20
    IDSA - DU Cycle
    UBET - DU Cycle
    LWAY - DU Cycle - DU5 - DU10
    ANIK - DU Cycle - DU5 - DU10
    SWIR - DU Cycle - DU5 - DU10
    ALDN - DU Cycle
    DPTR - DU Cycle
    HANS - DU Cycle
    GMAI - DU Cycle - DU5 - DU10 - DU20
    CRDN - DU Cycle - DU5
    PMTI - DU Cycle
    LIFC - DU Cycle - DU5 - DU10
    NAVR - DU Cycle - DU5 - DU10 - DU20
    JUPM - DU Cycle - DU5 - DU10
    SMTS - DU Cycle - DU5 - DU10 - DU20
    ECSI - DU Cycle - DU5 - DU10 - DU20
    HLEX - DU Cycle - DU5 - DU10
    LSCP - DU Cycle - DU5 - DU10 - DU20

    (HRT, NVEC, IDSA & LWAY have float levels outside normal parameters)
     
    #186     Oct 12, 2004
  7. The above calculation mirrors the articles I recall reading from Jack and others regarding A/D.

    I had trouble with this as well during the coding for Wall Street Analyzer. I could not decide whether to use absolute values for A/D or relative values. In other words, if A/D is negative on day t-1, but is LESS negative on day T, does that also equate to a (ever so) slight increase in accumulation? Conversely, if A/D is positive on day t-1, but LESS positive on day T, do we see a slight increase in distribution? Using absolute values (as you appear to have chosen) resulted in a slightly different score values for the stocks on which I tested both formulae.

    I settled on a simple comparison of Balance of Power between Day T and Day T-1. Either method 'works' as a method for scoring these stocks. However, the use of absolute values for A/D may prove to be more accurate. For comparison purposes, I have attached the Compare-Scoring.zip file. Inside, please find a Master List.xls file with our current final universe list scored. Also, I have added an .htm (20041011EODscore.htm) file reporting the scores using Wall Street Analyzer. Notice the slight difference in scores for many of the stocks.

    Thanks again for sharing your efforts, I hope you find the above information useful.

    - Spydertrader
     
    #187     Oct 12, 2004
  8. Many derive benefit from the tools you have shared. Keep up the great work. In regards to using the derivatives of P, V & A/D to anticipate price, creating a separate tool for this purpose would allow multiple individuals to test the theories. Analysis from various viewpoints often results in improved feedback. Such 'iterative refinement' may lead you into a new direction or provide insight into alternative data / result correlation. This in turn may lead to further optimization of your formulae / theories. In short, the only way to know for sure if you 'overengineered' this stuff is to test it.

    Thanks again for sharing your efforts.

    - Spydertrader
     
    #188     Oct 12, 2004
  9. After fussing with WSA, I've managed to get it to work on my comp. Anyhow, I've noticed a couple of things about WSA in that sometimes the indicators are unstable. Using HRT, the BOP would change depeding on the order in which I viewed the graphs (ie. ALDN -> ANIK -> HRT vs BCSI->ASKJ->HRT). Both seequences occusionally yeilded different BOP values. I'm sure this is a development issue for WSA that has something to do with refreshing the display environment. Just as an example, using HRT, the cycle score should in fact be 7 according to WSA since P,V,AD change from the prior to previous day were, P: 23.19 -> 24.5, V: 150100 -> 225600, A/D -.88 -> +.74. As a note on A/D, I'm using the actual values of A/D since I am only interested in the current day's value and not the change between two days. Therefore, I would only be using .74 to cycle score. Additionally, I use actual value since Accumulation and Distribution are complements of each other (ie. A/D -.74<>+.74 since the latter is strong Acc while the former is insignificant Acc, the complement of Strong Distribution). To use a 2 day A/D, we would need to modify the formula to compose the 2 day open, close, hi, and lo from consecutive 1 day bars.


    G33M4K the Newb
     
    #189     Oct 12, 2004
  10. I have added some idiot proofing to this version in order to address the ScoreFinalUniverse ordeals (ie. EOD vs Intraday scoring). Often times, I am the idiot against which I am proofing. This version is strictly EOD, meaning that the cycle score is indicative of the previous closing day and its prior closing day, instead of scoring according to current market levels given the prior days levels. Additionally, some have PM'd me about problems with time outs in the app. This should resolve those issues since the only way that time-outs can be produced is if one puts data that is either a misspelled ticker or a non-ticker. In any event, I absolutely despise making guinea pigs out of anyone except myself so I apologize for any inconvenience the prior version may have caused...

    REGARDS & NJOY
    G33M4K the Newb
     
    #190     Oct 12, 2004
Thread Status:
Not open for further replies.