Phil_Stacker
New Member
Before I start with a couple of negative points I want to just say that my views match yours somewhat. I too think the sharemarket is hope (consumer confidence) and not reality. I feel the true is same with real estate.
You state that every crash follows an interest rate hike, but then you state that it comes too late. I think that an interest rate hikes should not occur at all, and the fact they shouldn't have occurred leads to the market crash.
So now on the graphs - a couple of points - there is no such thing as 100% probability or likelihood in statistics. Until a decision is made there is some possibility/probability/likelihood of a decision not being made. This is especially true since the decision is being made on data not yet released, on a meeting that hasn't happened by people who are still monitoring the economic state. Hillary was nearly 99% likely to be president. One other quick point is that there is an expectation on what the rise will be, but what happens if there is a rise but it is minuscule or too high? Nobody is talking about "what change" will happen other than "100% probability of an increase between 75 and 100 basis points". My response is "sure, but there's only a 50/50 chance of that.
My prediction is a rise of 50 basis points, which will semi-shock the market and make gold come back to 1200 until next year's data clarifies things. I base my guess on the same statistics that most proponents use (I made it up just now - shhh don't let anyone know the market prediction secret).
Second point - you have a graph of correlations but every correlation has two corresponding probabilities, the probability that the correlation is JUST RANDOM but you accept it, and the probability that the correlation is NOT RANDOM and you reject it. The easiest way to show this is by displaying the confidence intervals (potentially the box and whisker plot as is commonly used with market data).
In other words, the correlations that go up and down could be just due to random data and have no relation to anything and, in standard statistics should not even be shown.
At very least the probability that the data IS RANDOM should be shown - i.e. (p <0.05). Further, was there any smoothing of the graphs, if so what technique was used?
The point? Currently it looks like gold is negatively correlated to the share market - that when shares go up gold goes down. This may be the case, but you didn't supply any actual figures for me to confirm that conclusion.
You state that every crash follows an interest rate hike, but then you state that it comes too late. I think that an interest rate hikes should not occur at all, and the fact they shouldn't have occurred leads to the market crash.
So now on the graphs - a couple of points - there is no such thing as 100% probability or likelihood in statistics. Until a decision is made there is some possibility/probability/likelihood of a decision not being made. This is especially true since the decision is being made on data not yet released, on a meeting that hasn't happened by people who are still monitoring the economic state. Hillary was nearly 99% likely to be president. One other quick point is that there is an expectation on what the rise will be, but what happens if there is a rise but it is minuscule or too high? Nobody is talking about "what change" will happen other than "100% probability of an increase between 75 and 100 basis points". My response is "sure, but there's only a 50/50 chance of that.
My prediction is a rise of 50 basis points, which will semi-shock the market and make gold come back to 1200 until next year's data clarifies things. I base my guess on the same statistics that most proponents use (I made it up just now - shhh don't let anyone know the market prediction secret).
Second point - you have a graph of correlations but every correlation has two corresponding probabilities, the probability that the correlation is JUST RANDOM but you accept it, and the probability that the correlation is NOT RANDOM and you reject it. The easiest way to show this is by displaying the confidence intervals (potentially the box and whisker plot as is commonly used with market data).
In other words, the correlations that go up and down could be just due to random data and have no relation to anything and, in standard statistics should not even be shown.
At very least the probability that the data IS RANDOM should be shown - i.e. (p <0.05). Further, was there any smoothing of the graphs, if so what technique was used?
The point? Currently it looks like gold is negatively correlated to the share market - that when shares go up gold goes down. This may be the case, but you didn't supply any actual figures for me to confirm that conclusion.