Estimation and Prediction Trading
- Thread starter aquarian1
- Start date
You all need to read George Taylors Trading Technique... Deals exactly with what this thread is focused on. Great book.
============
First, Mr Rock , congrats on the forcast of !380-1390;
but prediction would be 1380 or 1390., not both.Prediction is sililiar to prophesy; its much more precise than forcasting weather
.Prediction is not useful in markets at all; if it were, limit of position size would be stupid.And with prediction one could wisely own any market shortly.[assuming exchanges would permit that].
Of course you did sort of ''hedge'' your word prediction with probable...Some newer dictionarys, like new traders, muddy/ confuse the word probable with predictions... Not helpful or accurate.
Mr Rock;
Dow Theory would be a good place to start.
. Except like William O Neill, substitutes Nasdaq for DOW Transports; I prefer NasdaQQQ & SPY combo myself.
Forecast, prediction, whatever ... I'm not interested in semantics, I'm interested in what gets the job done. How can you congratulate me now; SPX hasn't reached the 1380-90 zone yet, and could continue upward from there if it reaches it.
Damn ⦠another interesting thread where Iâm too late to the party!
Perhaps you can be tempted you back?
George Angell, in Sniper Trading (2002) wrote about âBuy and Sell Zonesâ. These were price ranges determined EOD for the next day by calculating various metrics (from the Highs, Lows and Closes of the session just closed and each of the two days prior). With these price ranges, the idea was that an estimated range for the day could be deduced, and that this could be traded off of swing highs/lows with a high probability with the overlay of some further entry filters. The âmarket microstructureâ justification for the metrics was that they were all well-known to pit/floor traders (e.g. levels calculated off the daily pivot), and so would often behave as self-fulfilling levels.
So, perhaps some similarity to your Predictive Trading�
Two years ago, I automated the strategy and backtested it over a universe of 100+ US ETFs using bar data for the preceding four years (2007 - 2010), and an ATR based stop. I found only 3 instruments (out of the 100+) that gave passable performance over two years, and none that were good over four. And the ones that were passable over two were almost certainly just curve fits! So I lost interest in pursuing it further â¦
[⦠and then I also learned that George Angell had in 2002 settled an action with CFTC for allegedly defrauding clients (http://www.cftc.gov/files/enf/02orders/enfangell-order.pdf) which further discredited the book IMO! LOL!]
But, there is something tantalizing about the approach, I have to admit; there really is some relation between the zones and price action on an ES chart for example ⦠so I may revisit the analysis at some point.
I designed my test as though the price action was a mean reversion, and that was probably a mistake; itâs actually seems more of a momentum strategy to me now...
Any quantifiable approach that has real merits and is published, will lose its value pretty quickly, that's obvious. Adding to that, the author apparently is a known fraudster. Perfect for party people to base an approach on 