Expectancy = (Avg. win x Win%) - (Avg. loss x Loss%). Therefore there are two major factors that will determine expectancy before a trade is put on. The probability of a trade scenario/set up (a set of conditions like a hypothetical set up for the SPY could be to buy the next open if there is a 2% increase on higher volume in the SPY after a 10% decline and the result of 15 years back testing may show that this set up leads to the SPY being at least 5% higher 60% of the time after 5 weeks. So if you are looking for a 5% move in this instance the odds of success will be roughly .60 for avg. 5 week holding time) being a winner based on a certain exit strategy (like an ATR trailing stop or close below a moving average etc.) and the second factor is the extent to which you expect price to move in your favor expressed as a ratio of the amount risked on the trade i.e risk/reward (this could be based on historical precedent like CANSLIM which looks for a 20% - 25% move from the base break out or it could based on Fibonacci extensions or support resistance or projections based on the length of the previous momentum thrust) .
So before considering taking a trade you must ask: how well does this set up meet my criteria?. The stricter you are the higher the win rate will be but the number of opportunities will decrease and vice versa if you are looser. This part of the decision making is very subjective and experience plays a huge role in selecting best ideas from many opportunities. This aspect of the decision making is very important because not only will it tell you if you can take the trade it will also let you determine risk per trade and hence position size. With higher probability trades you can risk more per trade because 1) the odds of a huge gap through your stop is lower (assuming you avoid planned news releases like earnings announcement) and 2) the odds of a huge draw down are lower (at 60% win rate the odds of 5 losing trades in a row is only 1.024% however at a 40% win rate the odds of 5 losing trades in a row jumps to 7.776%! So with set ups that have better odds you can risk more than on set ups with lower odds.
So assuming your set up has passed the first hurdle and can now be safely called a "high probability" set up then you must consider risk/reward. There are essentially two types of systems. One where you know what the average move of the asset is likely to be as in the case of CANSLIM 20% - 25% and ones where you do not know and use a trailing stop. For systems with trailing stops it is useful to know where price will LIKELY get to before the trend reverses. This could be a swing high, all time high, bollinger band etc to establish a rough target and then in combination with you stop loss to determine a risk/reward profile for the trade. For systems that have essentially predefined exits/targets it is just as important to identify likely reversal zones which will tell you in the case of CANSLIM, for instance, a stock might have a lot of resistance before it can get to the 20% target and you know immediately that the reward may likely be lower than 20% and hence that the risk/reward may be worst than what you typically look for.
It is limiting your trades to set ups that have a high probability of success (i.e. those that fit your pattern) in combination with excellent risk/reward and repeating it day in and day out like a robot that makes money. Even further, if you do not have the two or either one for a set up or tip or idea you heard on the news then its OK to pass and do nothing! If you don't believe me go take a look at Mark Minervini's blog. One of the best traders ever and yet he can go for weeks with you trading because nothing fits his profile!
So before you put that next trade on run it through your probability and risk/reward checklist!