The scenario is unrealistic. In reality, we should be buying a stock that has a greater probability of its price rising than falling and we should only hold as long as that probability remains skewed to price rise. Whether price really does rises or fall after you have bought it ironically isn't the key deciding factor. So price could rise from 120 to 120.1, but now your TA says that it is more probably going to fall than rise. Leaving out the question of how far it might or might not fall, if you are certain that price is probably going to fall, why hold the stock a day longer? On the other hand, price could fall to 105, but now your TA makes you even more certain price will more probably rise than fall. Logic dictates you should actually buy more at 105.
oh, yeah, sometime I saw in the figure it happened again and again, but it is the first time i GOT THIS definition, thanks bro
Yes I would agree that it doesn't make any sense to be down $50 to make $20. There are very few situations where it would make sense to hold on but I think the answer falls out of your model. To use sporting analogies each trade is like an individual shot. In golf you're not just going to find out what you're hitting average is, you would break it down into driving statistics, iron statistics, chipping statistics and putting statistics. You would even categorise the environment in which you hit the ball, for example if you're playing a par 5 or a par 3 or if you're playing over water etc. Tennis, baseball, football, basketball, boxing all have the same dynamic. So if you are a discretionary trader and you have no common elements such as similar chart patterns or extreme levels of fundamental indicators, then you need to assess each trade on its risk / reward basis and then measure how good your judgement is by the profitability of your results. If you do have some common elements to your discretionary trading, for example you always use 5 or 6 of the same set ups to get into trades, then for each setup you need to measure whether the risk you're taking is justified by the return, on average, for that setup. If you're mechanical then just follow the rules you backtested. So in answer to your question, the risk reward is contextual to the data set. You should be comparing it to the other entries of a similar type and then you are solving for the break-even plus some margin of error. If you are buying the break out of a 200 day moving average amongst other trading setups then your risk reward for that particular type entry should only be guided by your summary trade statistics for that particular trade entry. If you're also trading a mean reversion set up at the same time then it's important to distinguish each strategy's statistics. If one strategy is working, is profitable and another is not then immediately cease the one that is not until you have identified an Improvement to get the performance past break even. It also means that you need to categorise each of your trades so you know which setups are working and which are not. To use a baseball analogy, measuring your overall batting average is fine but it's going to be much more meaningful if you know that you are pretty poor at hitting curveballs but your hit rate on fast balls is excellent. Therefore you should get better at identifying curve balls so you don't hit them. Over time maybe you can work on curve balls, but the immediate response should be identify specific problems and limit damage. Then plan, test and execute.