This topic is impossible to delve into without discussing personal opinions. So, I will state some personal opinions first and open myself up for criticism.
1) Price action is not ALWAYS random and therefore doesn't follow a normal distribution. That is, movement of the underlying is random much of the time, but there are points at which buyers/sellers group together to cause a non-random event.
2) OTOH, 95% of TA is complete BS.
3) Options are the best trading vehicle because they make it easier to gain an edge and take advantage of non-random patterns.
4) The zero-sum argument is flawed and inaccurate.
Ok, so in answer to your statement, POT is pretty much useless if you are using it to identify a mathematical positive expectancy for a credit spread. If you randomly enter credit spreads and systematically exit the spread if a certain price is hit, you are a living definition of neg. expectancy. The credit you receive is reduced for the likelihood that the spread will go ITM and then return OTM before expiry. The above mentioned strategy will not take advantage of that likelihood.
OTOH, if you enter your spreads under certain conditions and have some artistic license in exiting/adjusting them, you might be able to use POT.
Let's say you've sold a SPX call credit spread and don't plan on holding to expiry, and in fact you don't want to hold the position longer than 3 weeks. You have some sort of adjustment planned for a market rally.
You're looking at the charts and notice strong support 30 points below the current market price (your profit exit), but no other s/r is noticeable. Barring any economic events, non-random movement is unpredictable in this case. Using the 3 week date you can now figure the likelihood of hitting your profit exit point. IOW, the POT is the best estimate of your probability of profit.
POT also gives you a gauge in forming an opinion on the "true" POE. For me it isn't enough to simply say that I don't think SPX will reach ### by a given date. I want to know what my chances are. I use POT in determining whether or not a point will be touched due to random fluctuation. Let's say it's 25% by a certain date. Then, I use my forecasts to determine the current p/l if that were to happen, as well as the likelihood of breaking through that s/r.
In the end, if my "weighted" probability of profit doesn't present a positive expectancy, then I look for a different trade.
