MGJ, I think it's the latter. I simplified the topic to motivate the discussion. At any rate, it worked since I've received couple of helpful points on where to look for more info. Thx for the heads up on the NP response. I didn't have an e-mail alert set there. http://www.economics.ox.ac.uk/members/andrew.patton/Patton_IER_2006.pdf -bluelou
Tick Bar Charts are better than Minute Charts but not as good as CVSB (Constant Volume/Share Bar) Charts. GLOBEX started combining transactions about 18 months ago so any advantage that of constant Tick Charts evaporated at that time. I personally Intraday trade the FX markets, eMini indices, eMini Crude, eGrains and Bonds. I have no need to trade markets I can't read because of insufficient and inconsistent data is all that is available for them. I Swing trade about 300 other markets and about 1200 contract months in those markets. My philosophy about trading is exactly like watching a movie, why would I ever want to watch a movie where I had to read the subtitles when I could watch and listen to it and eliminate the distraction? Answer . . . I wouldn't. There are so many instruments out there to make profit from that there is no need to trade those that aren't as consistent easy to ready as others. I've literally spent years assessing hundreds of markets to determine those that are easy to read and the increments that are the easiest to read of those markets. those are the markets and increments I trade. No sense EVER making things harder than they need to be.
That's ProfLogic's deal. I've never looked into CVB. I don't know how or if constant volume bars are any different than regular volume bars. I can plot volume bars in NinjaTrader no problem but I don't know if that's the same thing. Let me know if you figure it out. Sorry, I just work with tick bars. -bluelou
Apologies Lou, the question was directed to ProfLogic - I should have been more specific. But since we're far off topic anyway I hope you don't mind if I ask YOU this question regarding stationarity: System A is in drawdown; in order to determine if the underlying market conditions have changed or this is 'normal' behavior can one create a cummulative probability density function over a look-back period or take a moving sample of the CPDF and create a surface. Then compare the most recent data to the surface to see if it is statistically different from anything seen in the past? I created another thread but did not find the answers sufficient. I was curious in your opinion if you think this makes sense or if you had applied anything similar?
I currenty use MultiCharts and don't believe in backtesting, only forward testing. I believe that for any method to be validated it must be tested in real time situations. Backtesting isn't valid because historic bars change from real time bars. For most traders this isn't practical because they don't have the patience to test anything over physical years of time. They are scared the edge they might discover will evaporate by the time they validate their process.
knocks420, I've never heard of the method you've suggested so I don't have an opinion on it one way or another. Are you outside of your risk of ruin calcs? Have you tried any of the statistical tests for structural change? Just my opinion here - but given what the mkts are like these days I wouldn't be surprised if your results do indicate a structural chg. I guess another way of thinking of the question is: is the change just in the 2nd moment (variance) or has the mean chg'd as well? Moreover, I would think that your risk of ruin calcs would override any of the descriptive statistics. FWIW, I had a drawdown in Oct - wasn't life threatening but the number of losing trades was beyond my risk estimates. I had to shut down for several weeks and recalc my exit parameters among other things. -bluelou
The difference between Constant (or capped) Volume Bars and regular Volume bars are quite simple. Let's say you create a 500 Volume Bar Chart that isn't capped to make it constant. Now say that the current bar you are on contains 450 contracts and only needs 50 more contracts to complete the bar but the next order comes in at 150 contracts. If the bars aren't capped or Constant then the software will simply add all 150 contracts to the bar making it a 600 contract bar instead of the 500 contract bar you had designated the bar to be. By doing this your software arbitrarily decided for you that what you wanted is meaningless. The purpose of keeping all bars identical in volume is to make the bars stable and the environment quiet of the distractions of normal chart idyosyncrasies. Again, you have to see this for yourself to understand it.
So if your historic bars matched your real time bars then backtesting would be valid for you? That is a pretty easy problem to solve.
That is part of the problem and yes, that part could be solved but the bigger part of the problem is the psychological part. Backtesting only goes so far but it takes real screen time to trust pulling the trigger on trade set ups. Backtesting doesn't build personal confidence and THAT along with an edge is what makes a trader profits. Ever wonder why 95% of traders fail? It starts with trying to take shortcuts and backtesting is a shortcut.