Yeah. Agree completely. However many people are doing this without realizing by applying a moving average to compute the trend. Whether you apply filtering in the time domain with a moving average, or in the frequency domain, it's equivalent. The same assumptions of stationarity are being made even though people may not realize it, which is dangerous. A trailing moving average also introduces a delay which I'll bet many people do not take into account.
Trend following is certainly predicting. If you identify trend properly you will be able to predict continuation of trend.
In other words, once you have identified the trend, the next step is to identify the price movement within that trend which indicates a high probability of the trend continuing or reversing. However, unfortunately, most people do try to use TA to predict where prices will be at year's end, on their dog's birthday, by when their stock will double or triple etc. This type of prediction is one of the biggest pitfalls for those who use TA as it was never meant to be used.
Without quantification, the discussion becomes almost "philosophical". Experience of such conversations in forums, over the years, has taught me that the word "predicting" inevitably causes a little consternation and some confusion: people use it with subtly (but significantly) different meanings and understandings, and this sometimes produces some talking at cross purposes. It seems to me that the key distinction is the one between individual and collective/statistical "predicting" (or between specific and probabilistic "predicting", to put it another way). With every individual trade I open, I can't "predict" its outcome (as anything more than a research-based, empirical probability-function), but I can tell the overall, collective outcome of my next 300 trades, within about 3% accuracy, with more than 95% confidence. We all decide for ourselves whether or not we think that's "predicting".
I have no idea why these 'semantics' debates continually show up in discussion forums. You can predict a coin toss. Your accuracy rate may be 50% +/-. You can forecast a coin toss too. You can predict/forecast the weather too. Accurately or not. You can predict/forecast anything. It's up to you. What really matters. What should, imo, really be debated is how accurate are the predictions/forecasts. Not whether we should call them this name or that name. ___________ That said, I can enter a trade and believe/say to the effect of, if not in these exact words: "I'll set my stops, both loss and target; and see how it goes." Or, "The trend has been up, let's see if it continues." The above are not predictions. The trader either doesn't have a guess/prediction/forecast...or doesn't care (a straddle, for example). Or I can guess/predict/forecast: "I expect/forecast/predict/guess that the trend will continue/reverse." Or, "I expect/forecast/predict/guess that the price will rise/fall." Imo, very few enter a trade without some sort of expectation that it'll eventually go in the direction that brings them profit. Even with a straddle, the expectation/forecast/prediction/guess is that the instrument will move, rather than stay flat.