Fun and Profit; Predicting the price of GLD

21

Yes, that’s what his model does. I make no assessment of how successful it might be nor how you might trade based on the predictions.[quote=“TripleB, post:20, topic:23”]
I’m highly skeptical anyone can predict the underlying gold metal price using any statistical models.
[/quote]
It’s good to be skeptical, but all those computer algos trade stocks everyday with statistical models. Pretty soon only the good ones are left, the others having gone broke or been shut down. It’s not easy, but clearly some people can do it.

2 Likes
22

Sounds like you have a good thing going on TripleB. Hope your success continues.

This specific model predicts the weekly high and lows of the ETF named GLD. It could also inform your decisions about the price of the metal itself, but the target of the prediction is GLD.

The correlations posted above from backtesting are true and accurate. You will see over the weeks the accuracy is confirmed.

2-3 times a year it will be inaccurate. This will happen in on both sides of the prediction.

Its up to you to figure out how to use it. I sell options on GLD against it. Occasionally trade shares on IB. Don’t buy options based on this model.

It’s so accurate that even with the low volatility a novice could figure out a steady, gradual profit. YMMV

23

@cogman I don’t understand how that could be possible. Can you elaborate on how the algorithm works? And if it does work, wouldn’t a hedge fund with a billion AUM come in and make the same trades, compressing down the profit to nothing?

I suppose my biggest issue is that you’re looking at historical data to predict what will happen in the future. It can’t possible know if the Fed will raise or lower interest rates, if a new gold mine will be found, if India bans gold ownership spiking demand, etc. And all of those events should have huge swings.

1 Like
24

The statistical methods are derived from social science statistical methodolgies and engineering/process manufacturing methodologies.

You are correct in that unforeseen events can cause erratic changes in the prices causing the predictions to be off a few times each year on both sides.

The final judges are the correlations between the predicted and the actual. This can be easily done with Excel or any stat software. After about 25 weeks here there will be enough data posted for anyone to run the calculation.

1 Like
25

For GLD the Blind Squirrel prediction of hi/lo for this week ending 10/20:

HI: 124.58
LO: 121.52

26

It’s .50 from the predicted low. Keep an eye on this.

27

This looks like a buy area now. Make your own decisions, there’s always risk in losing money in the stock market.

1 Like
28

GLD will bounce way up once BTC crashes. Both attract the same ‘investors’.

1 Like
29

Be nice to know when.

30

For GLD the Blind Squirrel prediction of hi/lo for the week ending 10/20:

HI: 124.58
LO: 121.52

The actual hi/lo for the week:

ACTUAL:

HI: 123.97
LO: 121.39

31

For GLD the Blind Squirrel prediction of hi/lo for the week ending 10/27:

HI: 122.79
LO: 119.97

32

ACTUAL:

HI: 121.90
LO: 120.21

33

For GLD the Blind Squirrel prediction of hi/lo for the week ending 11/3
:

HI: 122.55
LO: 119.92

34

ACTUAL:

HI: 121.96
LO: 120.18

35

For GLD the Blind Squirrel prediction of hi/lo for the week ending 11/10
:

HI: 122.91
LO: 119.97

36

ACTUAL:

HI: 122.62
LO: 120.26

1 Like
37

For GLD the Blind Squirrel prediction of hi/lo for the week ending 11/17
:

HI: 122.62
LO: 120.26

38

ACTUAL:

HI: 123.17
Lo: 120.77

1 Like
39

For GLD the Blind Squirrel prediction of hi/lo for the week ending 11/26
:

HI: 122.93
LO: 119.93

40

ACTUAL:

HI: 122.95
LO: 121.04

1 Like