Hey everyone, I am an undergraduate student and one of my course this semester requires me to create my own trading strategy with backtesting results. I thought of using ADX and OBV to generate possible buy signals. The buy signal is generated only if ADX is bullish (ADX rises from below to above the 20 level) and OBVMA crosses above OBV. I tried searching for "ADX OBV" online but i'm unable to find similar trading strategies using these oscillators. So i'm wondering if this trading strategy is logical and worth backtesting as no one seemed to have suggested this combination before.
As much as I love the intuition behind the OBV, I have never seen it work. I have checked it out short term, long term, swings, intra-day to 1-minute candles. I have tweaked it's look-back further and shorter, and never seen it survive a contest with a fair coin. If you are *forced* to utilize OBV, I would make its conditionals on your system *very* generous -- like, "If OBV>0, then ADX" and just get it the hell out of the way. The ADX, on the other hand, is mui fun. Play with its look-back; and test it over various time frames. And do not be trapped into thinking that the condition that justifies an *entry* needs to be the type of condition that gets you an *exit*. HAVE FUN. (But moreover, do not be online looking for (us) to pimp your homework for you: for the time you've spent looking for an answer, you could've completed the assignment, and tucked a lot of learning under your belt. )
Probably not. OBV is worthless. You have a difficult task. Your professor probably has never traded and likely doesn't know "come here" from "sic 'em" in the real world, yet he wants you to "develop and backtest" a mechanical strategy. Unfortunately, likely impossible for you. Some of us on ET understand how to trade successfully, but highly unlikely in such a way to mechanically backtest concepts to be able to draw a conclusion to satisfy your task requirements. (Personally... what I do I call KISS...."keep it simple, stupid"... it's somewhat intuitively subjective and therefore very difficult to define in such a way as to mechanically backtest in any way to "prove/support" a conclusion.... which I'm sure is what your task requires and your prof wants.) The entire trading/financial world.... recruits the best brains, commits $Millions, uses the most powerful computers.... to try come up with what your prof wants from you. Unless your prof expects you to come up with, "I tried everything I could think of, and nothing worked"... then he shouldn't have given you this assignment. The most successful strategy is to "front-run" orders like the HFT guys do, but I don't know how you could backtest that.
OBV doesn't have a lookback by default. I presume you rolled your own? Building a successful strategy (not that I would know...) is less about the specific indicators (except it is) and more about how you approach it. Figure out what it is you want to find out from the price/volume/fundamental data and then find or create indicators that match. First you have to decide: can I predict the market? If the answer is no, then your job is to figure out the direction and strength of the price movement and see whether there is enough fuel to give you some profit. If the answer is yes, go fuck yourself.
I was introduced to this by HobbyTrading and I was surprised once I looked it up as I had previously created my own version! This is legit advice.
It would take more effort to program, but stick to charting, trendlines, triangles, wedges, moving averages. You don't need to re-invent the wheel for longer term approach, less trading means less fees, taking advantages of gaps and dividends.
I thought I saw an indicator something like this many, many years ago, but I couldn’t find it anymore, so I too created my own version a few years back, which I started calling “dynamic support/resistance envelopes.” (Last week was the first time I'd seen the term again in years.) But I had a problem with my "makeshift Donchian channel" (above on the left) though. For at the time, I was strictly scalping and I wanted an indicator that would only make contact with the candlesticks at their very “tippy tops” or when they were scraping the very bottom of the barrel. If I had been using real Donchian channels, I would have been out of luck, but because I was forced to come up with my own version, I was able to make modifications (the version on the right) which does a much better job of giving me a definitive idea as to where I might want to take profit or start looking for bounces off of what I call statistical support/resistance. Then today I saw something called an Adaptive Price Zone Indicator, which sort of looks like what I created on the right. Had I come across it years ago, it might have saved me some trouble (except I have not yet seen anywhere online where one can download a version for MT4).