I think more in terms of probability, anticipation or better yet a projection (a prediction made by extrapolating from past observations) than what I'd normally think of as a an outright "prediction". But then this has been discussed and argued Ad nauseam and always disintegrates into a debate of semantics. I never did figure out why the prediction crown need to project their view of it onto everyone else.
Like others who have commented before me, I don't try to predict anything. I look at probabilities. I take trades based upon market context and statistical analysis of key areas of support/resistance. I look at risk (how much I have to lose to cross the statistical threshold) and potential return (what's the probability of price moving from point A to point B). That's it. No indicators. Just context and statistical S/R yielding a net positive expectancy per trade.
Fundamentals + probabilities. Assume over the longer term the fundies win out, in the shorter term it's all random. Via options you choose two out of the three directions: up, sideways, down. Up or down is determined by the fundies, of course. Get vega on your side (this is a little complicated), make theta line up with vega, figure out the timeframe you want to take your bet. Sacrifice a rubber chicken to The Mighty Favog. If you have it all right, over time you'll win.
Yes, I've had a few journal threads over the years, all of which have been profitable. Here's the most recent one but its progress has been slow because QLD hasn't been moving much in the last few months. I just closed that position out today. I cannot predict price despite trying everything out there, and when I finally accepted that price is random, I changed the way I thought about everything and became profitable. Note that price may not actually be random, but since I cannot predict it, it is random to me. If you can predict price and are profitable that way, then definitely keep doing what you are doing.
With a computer... and large amounts of price and volume data. The analysis consists of identifying support and resistance areas based upon volume (buying and selling activity); and then extracting statistical reliability from each data segment that makes up the support/resistance area. I then use that probabilistic model in combination with overall market context (market bias) to determine trades. Another way to look at it is that I'm only willing to take trades at specific prices under a predefined set of circumstances.