In this writing, fair pricing models for stocks are derived based on probability theory. The models are validated by US stock market 5-minute mid-quote to mid-quote reversion, not due to bid and ask bouncing, with theoretical values highly close to market observed ones. The relationship between volatility and fair value time decay is calculated. Minimal looking forward time for full fair value realization is derived for given risk tolerance. The results explain why high frequency trading is necessary to catch short term gains and provide regulators new insights into stock market and how we could build a better market. Many examples are presented to demonstrate how the models may be used to explain things happening in stock market. Copyright © (Meihong Shan) All rights reserved.