Monday, 21 May 2007
The book reads rather quickly and to the point. Not much notes on the web, except reviews at Amazon, (which provides search inside the book). That being said, the strategies may have a few opportunities, and they are not back-tested. More like some quick patterns studied on trading simulation systems.
Overall, good quick read, good ideas for rare events. First chapter on "gap trading" is basically conditional mean reversion when gaps occur, with 6 techniques being extra conditions to the model.
To demonstrate his strategies, he has no sense of back-testing properly via In-Sample and Out-Sample, but instead simply reports on how well it did on history. Also, many of these are low frequency strategies, having a few occurrence in multi-year period. Besides the Buy/Sell signal, he has no strategy on how much to buy, so instead, he just take some random size, e.g. 30% of his initial capital.
Strategies:
- gap to signal reversion strategies
- pair trading on SP500 vs QQQ markets
- buying bankruptcy (a few cases only)
- TICK as indicator of market sentiment, to trade QQQ
- Bollinger Band (20-day EMA with 2 Std Dev band) to indicate if a stock is overbought or oversold.
- Stocks less than $5 are avoided by investors
- Turtle trend indicator: 55-EMA vs 22-EMA, crossing indicates momentum
- QQQ crash: defined as 1.5 std dev down from 10-day EMA. A good time to buy
- Fed Yields, bonds: competes with the market in asset allocation.
Buy if 10-yr yield is 25 bp down from 1 month ago, and Dow is down 2% in last week.
junk bonds: buy the market when SP500 underperform Merril Lynch High Yield Index by 5% over 3 months. (market undervalued)
- Addition to an index expects 5% increase
- Deletion from an index: going low until effective day, then large snap back due to over-sold
Using the 200-day Moving Average indicator for SP500:
- buy when going above 200-MA, sell when going down: 70% bad
- buy when going above 200-MA, hold for 1 day and sell: no gain
- contrarian: buy when going below 200-MA, hold for a month: 10% gain over a few instances
End of quarter, end of month: fund managers may oversell on the last day to get performance.
Outside month: a high volatility month usually follow by a low vol month, so if last month is higher in vol than the 2 previous month, buy on the first of the month, and sell on the last. 2% gain.
-10% panic: if stock is down 10%, buy and hold it til end of day. Variation: hold for a month. Let the over-sold stock bounce back. Variation 2: buy index, hold for 2 months.
Option expiry date (3rd Friday):
- don't fight the trend. if trending up, buy and hold for a day.
- OED and QQQ (Nasdaq100) anti-correlate.
Convertible arbitrage: pair trade Preferred vs Common stocks when company issues convertible bonds, and the stocks have fallen 10%. Rely on mean reversion.
Intraday Bollinger Bands: 5-min and 10-min reactions, preying on over-reaction of market to bad news: act when large jump, then get out of your position in 5 or 10 mins.
Four straight days: people try to time the market trends, and bet that it reverts after 4 consecutive trendy days. So you bet that the trend continues for 1 more day.
Wednesday as the reverting day of the week:
- buy QQQ when low of Tuesday is 5% less than high of Monday, and less than low of Monday. Buy at open of Wed.
- sell when 2 consecutive up closes occur. sell at 2nd close.
What does not work:
- gut: never trust anyone who listens to their gut
- confirmation: waiting for rebound before buying may cause you to lose a few %
- candlestick patterns: no longer work
- seasonality: no longer work
- low/high P/E: don't say anything
- options: don't leverage the power of compound
Only registered users can write comments.
Please login or register. Powered by AkoComment 1.0 beta 2! |
Home 
