Sunday, September 18th, 2016

Sugar trading model… feedback please……………..?

I Was Asked to come up with a model for soft commodity trading at recently. I’m a noob, and I do not really know much / have any real experience. All comments welcome. 1. Technicals – Finished backtesting and Have The Most Accurate Indicators for weekly, daily, hourly and 10-minute charts for the last 6 months2. Fundamentals – Finished Researching. Relatively Have Detailed breakdowns of the structures of major producers. Not Decided Each individual weight factor to adjuntos There’s no way to accurately weight for me Each one with what i have, And Also Would not sure if it make a difference. The idea of ??the model is to use TA to find trends and projections. The Appropriate options Then strategy is implemented (bullish, bearish, sideways) Upon breaking news, if the news strengthens the current trend (eg bullish news for a bullish trend), Then the bullish position is added to Immediately. If the news weakens the current trend, Immediate exit the market till you decide on a direction. * Relatively inelastic demand and como is at a steady rate (Directly proportional to Population) My questions: 1. Is this far too simplistic to realistically work? 2. What else Should I do to make this more Accurate? 3. Is There a point to weighting fundamentals? I feel like I do not need to refer for Theoretical price, just the POLARITIES of market directions to trade with. tyvm!

Comments

3 Responses to “Sugar trading model… feedback please……………..?”
  1. Connie says:

    Serving as a model is absolutely a respected vocation

  2. BIll Q says:

    “1. Technicals – Finished backtesting and have the most accurate indicators for weekly, daily, hourly and 10-minute charts for the last 6 months”

    When backtesting, you should find what the most accurate indicators were for the period from 12 to 6 months ago. You then test these indicators on the period from 6 to 0 months ago. You will quickly find that the technical indicators that worked during one 6 month period are often not the ones that work in the next 6 month period. (This also tends to hold true for 1 year periods, 3 month periods, etc.)

  3. Barry789 says:

    TA. Step back and look at the graphs of the various time frames. Are there any differences in them, other than scale? My hunch is that they are the same, or self-similar.

    Your fundamental analysis, while simple, is pretty good.

    What you’ve described is a momentum or trend-following trading strategy. In essence, you work to identify the trend, then increase or reduce exposure as new information comes in that might lead you to modify your opinion of the persistence of the trend. Think Bayesian analysis — modification of probabilities in the light of new information.

    You are saying that the trend will persist until it doesn’t. Notice that this statement confirms that there is a serial dependence that violates the random walk idea.

    I’d say you have a good first-pass model.

    I agree that weighting the fundamental factors might not add a lot to the accuracy, but consider that quite a bit of cane is going to fuel in places like Brazil and there is quite a bit of corn being channeled into fuel here in the US. Think about what this might mean for demand for sugar. The mess in the Near and Middle East are raising oil prices, which increases transportation costs — how does this factor in?

    If you build a regression model, you might build in some 0,1 variables, just to see what happens. My guess is that they won’t add a lot day to day, but might be good at turning points, which is the precise time when the momentum model will be reversing itself.

Speak Your Mind

Tell us what you're thinking...
and oh, if you want a pic to show with your comment, go get a gravatar!

*