see not really in the same class "these are real indicators" advancing issues over 1300 might wanna be long tick flips between 800 and -800 choppy tick over 1000 might trend
Thank you for all your suggestions. I should have be more specific on the strategy and time frame of my trading. What I am working on is market-neutral or long-short pair trading, in which I hope the long-side stock will beat the short side stock. Due to slippage of some stocks, I can not enter and exit the positions so frequently. I enter the positions during the market hour and exit at market close, which means the holding time is from a few hours to tens of minutes. Is it possible to find some statistical edge for pair trading with these conventional indicators? Or any "edge" I find based on historical data are likely due to overfitting?
Hi Aisone, What do you consider to be the best way to use ROC... only take buy signals when it's above zero and only sells when it's below... or is there a divergence aspect to it whereby price might be going up but ROC is going down?
Generally I'm a seller at high positive roc levels, and buyer at low negative levels because I am a contrarian trader. But I use many different smoothed roc's of different time frames, and compare them to raw price roc to determine trades. At any given time, some smoothed ones are positive and some are negative, so there is no hardfast rule for everything. But I will always take a contrarian position on the raw roc. It's the relationships between everything that counts though, so for example, some high roc levels can be a great sell, and sometimes they would be a very bad idea to short. But one thing I've learned about using roc's is that the price prediction changes as time passes, so even if the actual price isn't moving, roc values are, so sometimes it just takes time before an opportunity for a particular trade can present itself (and alternatively, time can dissipate potential opportunities as well.)
The most popular indicator is the invisible one. The market moves in various pattern / manner / speed / behaviour / rate of change.... ie indicators probably work for a day or two out of 1 month / year. HAPPY INDICATORS HUNTING !!!
Using ROCs on different time frames is an interesting idea. Thanks, I'm a TA nerd and will enjoy looking into this.
TD can chart stock differentials and combine the volumes. It even has VWAP for baskets, spreads, pairs, etc. I like a linear regression moving average against VWAP. Example using the XLK/XLF tech/banks sector spread (synthetic index). You can also use the Dow tick internal -- $TIKI The white data is basically $TIKI - LRMA($TIKI). You can see the raw data in black if you look closely. There's a whole ecosystem of bots using VWAP. My favorite is regression error analysis. Implied vol is important. I like indexed IV with custom weights.