Predicting a crash is a very complicated process, particularly in stock markets. The target of this research is to provide a formal verbal definition and measurable conditions to identify crashes, and its different types, as well as segregates it from other downward
patterns, such as corrections, bearish trends and crises. In addition, to design a simple and efficient model to identify speculative bubbles those precede only stock market crash patterns. The proposed bubble model mechanism is based on three blocks. The concept of Bollinger Bands is the keystone in measuring the first building block which is "Bubble Momentum and Size”. The study applies amendments to it to maximise the benefits of its simplicity and spontaneous application, particularly by public investors and to overcome its inaccuracy of assuming the normality of price returns. The second block is “Bubble Returns” which is based on the concept that a crash pattern is a shorter reversed trend of bubbles. The third block, “Bubble Price Patterns,” is built on the finding that the speculative bubbles at their highest levels, particularly before turning points, are characterized by some common patterns.

The American and Egyptian markets have been chosen as case studies for the research. Both markets, represented by Dow Jones and EGX 30 respectively, are used since their inception until the end of December, 2013. A data collection process comprising five stages was applied in order to increase the quality of the research data and to guarantee its accuracy, completeness and consistency. For optimisation purposes, the model tests different value levels of bubble periods, size and momentum which generate exit signals
before stock market crashes taking into account the different number of working days and the liquidity and volatility levels between both markets. The outcome scenarios are scanned and filtered by means of a chart analysis generated on METASTOCK software in
terms of the number of crashes that were preceded by exit signals at considerably high prices. The selected scenarios then enter the evaluation phase in which they undergo a series of tests in order to back test their predictive power in predicting crashes in terms of
frequency rate, timing quality and pricing efficiency. In both markets, the model was able to predict all the great stock market crashes, all the crashes that were preceded by historical high prices, as well as all the famous and well known crashes to public investors such as the 1929 Great Depression, the 1987 Black Monday, and the 2008 Financial Crisis in the American market, and the 2006 crash, the
2008 Financial Crisis and the 2011 crash after the Egyptian revolution in the Egyptian market.

A comparative analysis between the two selected markets shows that, although the Egyptian market is more prone to crashes, applying models adjusted for liquidity could generate higher profits. On the other hand, the American market is much more
complicated. However, it is stronger and much more digestive to negative events.