Bollinger Bands and MACD are fantastic tools and I put them very near the head of the class. The correction combination of tools is Prudent Risk Management, MACD, RSI, Bollinger Bands, multiple time frames and price action. The latter 5 help you see the picture and the first one is your edge.
CHALLENGE: The MACD is not my favorite indicator, but it does work. So, I offer you a challenge -- you post a graph of any underlying you wish -- any market, any calendar, any calendar length. Two restrictions: it has to be available on TradingView.com (thus, we both have access to the same platform, and the same data), and second, that the periods (i.e., "candles") be between 1-day and 1-month duration, inclusive. You post what you think demonstrates MACD failure. Your post should exhibit your explicit Entry signals, Exit signals, and claimed total points gained/lost based on the trade available with the start of the next candle. (Thus, no trading until the "signal" candle has closed.) For a year's worth of day-candles, this takes about 30 minutes to mark entry(s)/exits on the indicator, and another 30 minutes to read the resulting points from the chart. I will follow your post with my own, "tweaked". And regardless what your totals might be, my totals will be positive, likely beating the market return, AND they will be of less draw-down than the market experienced. (Translation: LESS RISK.) But here's the kicker: you'll be able to take my MACD setup and apply it to pretty much any other underlying, and it will have a similarly high probability of making-money-with-reduced-risk. Sweet, eh? To paraphrase, I appreciate the 40-year-old MACD, but there is better -- and from the very same vintage.
I have one indicator based signal, multiple divergences across time frames. The great majority of my signals are PA based with a momentum oscillator which serves to filter and qualify the signal. It works for me but it probably won't work for so don't try
I told already in previous TA threads: this discussion does not make sense. The essential question is: What does it mean that" it works"? The answer can be: I make 10% a year I make 10% per month I make 10% a day I make 10% an hour I double my money every day etc... How good it is working should be linked to published performance. Till now "it works" means that you don't lose money. All the rest is a guess. There is no clear and accurate definition of "it works".
Somebody forgot their brannnnnn flakessssss. "It works..." is quite simple: it makes US$1.00 "It works well..." might go a step further -- it begets 80% of market buy&hold, with reduced risk. "It kicks ass..." might be that it *beats* the market, *while* reducing risk over buy&hold. C'mon, man. You know this.
I believe the premise of the question is can an edge (big or small) be gained through technical analysis and technically derived trade plans. To those of us who trade technically for a living, it's a silly question but to a new comer or to one who hasn't figured it out yet, it's an important one.
And importantly, it's not so much about *predicting* "where the market will go next" as it is in *accepting* what the market has done so far, and then placing a righteous STOP where the market has demonstrated that a true turn/trend would hit. That's it! No magic! Measurement + numbers = {Entry/Exit/Stop} And where did the TP (Profit-taker) part of the bracket go? With nods to any differences of opinion thus far, let us pay tribute to the mantra of trend-tappers: "Let your winners run; cut your losers short." I always thought that was a *hilariously* dumb, ignorant, ill-informed aphorism. Either you're *IN-vested* (I thought -- "Buy&Hold"ism), or you're a short-timer -- a trader[!!] -- just looking to ride a quick up-or-down, and get out while the gettin' is good. I've had to have my eyes opened to a *marvelous*, de-stressing, rules-based world of actually holding positions for an indeterminate length of time, with an exposure-reducing exit already pre-planned and ready for execution. It's simple: with real trend following, you first declare a trend (your MACD,etc Entry signal), and when that price movement violates your declared settings, you exit (following the received Exit signal). Long or short -- you don't care. Quickly or extended? You don't care. You are mechanistic, formulaic, codable, detached, nearly-bored. And if that market goes for 7 days or 17 or 27? You're good. A turn will take you out, not a TP. You are letting any winners ride ride ride, while cutting those losers short. ("Like the ol' man sez!") Are profits guaranteed? Nope. ("'Course not." ) But expectancy? Positive expectancy? Yep. From 4-5 recent examples, you might expect 25-35 trades in a year, from a MACD system based on a 1-day candle. In a rising market, most will be long positions. 65% will be winners, gaining 25%-35%. (Individual) Losers will be 1/3 of market drops. Winning trades average 7 days, losers average 3 days. So! Three steps: 1) Tune the indicator (MACD under current view) to where its actions seem to match best with the asset/market you wish to trade. 2) Mark off entries and exits on the MACD. 3) Transpose those entries/exits to the asset/market chart (on the open, following the signal's closed candle), and start adding up results. For the empirically-minded, there are 10,000 *obvious* ways to tighten this up, including ParabolicSARs, trailing stops, Bollinger-derived stops/reversals, dynamically-derived indicator parameters, yadda-yadda-yadda. This is just the least little nibble at a sizeable Fun With Numbers apple. Tiny tiny tiny little nibble.
I agree, I use targets based on average MFE in my tf's. It keeps things simple and it works and that's a good combination. More often than not an extended run will give multiple entries. If one doesn't then it doesn't...that's trading.
I use RSI as a trend indicator with settings other than the defaults but here is it with usual length of 14 and OB/OS of 70/30 showing divergence at the Dow top of the crash of 1929: Everything works and everything fails to a certain degree. Does it really need to be said use what works best for you?