I had a question the other day asking “Does term “sudden” in your chart relevant to second derivative? In any case I want to create graph of SD but I don’t know where I can get put/call data series on GLD and/or GDX/GDXJ. Is this publicly available information?”
I’ll answer the question above and a few others I’ve had in an attempt to describe the model without giving the whole model away.
Firstly a conceptual recap. When investing in a market which has no cashflows and hence no valuation capability such as gold, I’ve found that you always need to know where you are in the sentiment cycle. You don’t want to be the last one piling in at the top, or the last one giving up at the bottom. Entering and exiting (or simply topping up or reducing exposure) should be done inversely to sentiment turning points. That is, when sentiment is suddenly accelerating positively then more often than not, the last of the investors’ are piling in and you want to be selling into this. Conversely when sentiment is suddenly crashing and all the investment flows are skewed to the downside, then investors are fearful and fully positioned for downside price movement – in those instances you want to be buying in the face of downside fear.
This approach works most, but not all, of the time. Sometimes the downside fear, or upside enthusiasm, is vindicated in price action and prices keeping moving for longer than expected as external factors drive further price movements. For example the Model has no idea that Janet Yellen was about to speak last night, and hint that U.S. rates would stay anchored for longer than the market was pricing in. However, using this model serves to minimise the damage when you are over-exposed in a falling market and maximise the reward when you are fully exposed in a rising market. Over-time you can start to out-perform by ignoring the market noise and almost holding positions inversely to the view of the masses.
Back to the question, the answer is yes the model is all about the second derivative. Or put another way, it’s about measuring and monitoring turning points in sentiment over time, as opposed to basing decisions on absolute sentiment levels. The Model tends to change from “buy” to “sell” only about 6 or 7 times a year, and lasts anywhere from a few weeks to a few months at a time, before sentiment swings around again and the model switches from buy to sell, or sell to buy.
For the model to switch the “sentiment score” (measured as aggregate Put Premium divided by aggregate Call Premium) needs to change by a pre-specified % amount over a pre-specified number of trading days. These pre-specified model inputs are what drives the success of the model and have been hard coded into the model after years of testing.
For input data, I use the entire COMEX option premium array for the most liquid maturity futures contract – eg Gold Options on April 2015 COMEX futures. I wouldn’t use GDX or GDXJ as these are equity indices. GLD is not a bad proxy, but options data is less liquid. The other benefit of using COMEX and other exchanges is that from experience I know that the gold Over-the-Counter (OTC) market uses the exchanges to arbitrage away any OTC action. Therefore you are picking up the sentiment of the gold OTC market which is the biggest market in the world for gold trading.
Finally I get my data from my Bloomberg subscription. However in my early days I remember going onto exchange web-sites and simply typing down each day’s closing prices for each outstanding option contract. I did this everyday for a year or two, before launching my Fund inside a hedge fund group and latching onto the time saving Bloomberg access. However at $50,000 a year it’s not available to everyone, so try the NYSE website for GLD or COMEX website for Gold Futures.