SP500 fundemental forecast

Discussion in 'Technical Analysis' started by sportsguy, Nov 30, 2002.

  1. This forecast is about 10 years of data, which includes the tech sector unlike the 80's, and estimates the SP500 weekly average of the daily (open+high+low+close) . . . . linear regression tries to drive the line dead through the middle of the data. . . . and i have for 10 years of data

    The dark blue is the actual week avg sp500 through Thanksgiving week, the boundaries aroung the purple forecast are the limits of the mean absolute error of the regression. . from 13 weeks ago.

    The forecasted data is the standard and poors operating earnings from the S&P earnings site. . .
    The operating earnings are diced from quarterly to weekly, and then I calculate the P/E based
    upon the 52 week annual trailing earnings. . . . i forecast the future operating earnings based upon
    purely fundemental data, including:

    LEADING INDICATORS, M1, MFG UTILIZATION, CPI/PPI COMMODITY PRICE INDEXES,
    CHAIN STORE SALES, USDOLLAR

    The Operating P/E is derived from

    ANNUAL EARNINGS GROWTH RATES, CONSUMER LOANS, CPI/PPI COMMODITY PRICE INDEXES, CHAIN STORE SALES, USDOLLAR, LEI, CHAIN STORE SALES, YIELD CURVE, INDUSTRIAL PRODUCTION

    The interpretation is the the market could possibly get to the 975 area (9100 on the DOW) before retreating. . . one variable in the inital forecast from my previous weekly close forecast was dropped, making the forecast less over optimized (better regression statistics) Watch for deterioration in the McClellan Osc for signs of the topping process. . .

    sg
     
  2. downloaded 49 times, and no comments? i expect to be flamed for using fundemental data in a TA department, although the model uses stochastics and momentum. . .

    or is everyone waiting to see if it works, so to speak. . . which is fine, I AM TOO!

    sg
     
  3. dottom

    dottom

    Try posting trading summary based on mechanical method of following the forecast and you might get more interest...
     
  4. if the model uses stochastics and momentum then it IS a technical model, since stochastic and momentum are purely focused on price and time.
     
  5. not on price, but on the fundamental indicators. . .

    price is not an independent variable in the model

    sg
     
  6. Here is the thing with models like yours that I dont get.. are you predecting more upside or the rally is over?? You are saying we can go higher then lower... to me thats double talk I hear on CNBC...


    --MIKE
     
  7. the forecast is basically an estimate of fair value given all the price action from the past 10 years. . . wars, famines, plagues, etc. . . so it is an estimate of where the variable relationships of the past put the price estimate of the conditions today. . .

    doesn't mean that all variables of today have been reflected in the price, but forecasting the stockmarket 13 weeks ahead is hard enough and so it is not perfect. . . but gives me an idea of where the market should be without alot of noise. . . or local worries that are unquantifiable for the short time they are here

    sg
     
  8. dottom

    dottom

    Have you created a mechanical way to trade the values produced by your model?
     
  9. so if we are very far away from fair value, I expect a rebound, or if it is trending in a direction, i will have a preference for playing that trend. . . it is a forecast, not a system. . . . to have an idea where the fair value WILL be in the future

    just a guideline

    sg
     
  10. dottom

    dottom

    My point is that if you are sharing "just a guideline" or an indicator then you are likely to not get any responses around here. People are interested in how well this value or indicator can help in their trading, and the only way to evaluate that objectively is by looking at mechanical methods and trade results using such methods.
     
    #10     Dec 5, 2002