Cheers
- #22
- Apr 12, 2010 8:59am Apr 12, 2010 8:59am
- | Joined Feb 2009 | Status: Price Stalker | 1,197 Posts
Economists have forecast 9 out of the last 5 recessions
[quote=spekkiefx]
They only use them during black outs, to light up the room (-:
[/QUOTE/
I like this one
Now ask yourself if it would have made any sense to use candlesticks in those graphs. Would it have helped to describe the problem or present the study results? I'm not surprised not to see any candlesticks in http://dl.dropbox.com/u/190212/ff/Qu...ent%20News.pdf. There's not a single graph that represents price in it.
The point I was trying to make is that it's unwise to judge what looks 'quanty' and what doesn't especially if you don't take into account what a particular graph is meant to represent.
One of my best performing systems is based on volatility adjusted currency indices that I like to plot as candlesticks for some sort of analysis. Is it inappropriate to call it 'a quantitative system' then ?
Cheers
They only use them during black outs, to light up the room (-:
[/QUOTE/
I like this one
I have browsed through tens of quant papers. Not even one graph in them was done using candlesticks. All were continuous lines.
The point I was trying to make is that it's unwise to judge what looks 'quanty' and what doesn't especially if you don't take into account what a particular graph is meant to represent.
One of my best performing systems is based on volatility adjusted currency indices that I like to plot as candlesticks for some sort of analysis. Is it inappropriate to call it 'a quantitative system' then ?
Cheers
[quote=Ipso;3631121]
You see, I knew that somebody will say that the information in that particular article is not appropiate for candlestick display.
But like I've said, in tens on articles, NOT EVEN ONE used candlesticks. And some were analyzing stuff like EUR/USD or SPX.
But these articles were not trading systems, they were technical methods which improve your edge (like non-linear time frames).
EDIT: Here, one with prices: http://dl.dropbox.com/u/190212/ff/0811.1561v2.pdf
They only use them during black outs, to light up the room (-:
[/QUOTE/
I like this one
Now ask yourself if it would have made any sense to use candlesticks in those graphs. Would it have helped to describe the problem or present the study results?...
But like I've said, in tens on articles, NOT EVEN ONE used candlesticks. And some were analyzing stuff like EUR/USD or SPX.
But these articles were not trading systems, they were technical methods which improve your edge (like non-linear time frames).
EDIT: Here, one with prices: http://dl.dropbox.com/u/190212/ff/0811.1561v2.pdf
You still don't get it. These guys are not predicting O, H, L, C so there's no place for candlesticks here either! It simply doesn't make sense! Also, would you really use candlesticks to plot 2500 data points in a 15 cm wide box or would you use something a bit more readable ? (thanks for this paper though, it's quite interesting)
Anyway, keep an open mind and refrain from thinking you've already seen it all. End of thread for me at this point. Cheers.
97% of retail traders lose. 99.9% of them use candles.
100% of the quant papers I've read don't use candles (maybe I'm interested only in the sub-fields which don't use candles, I don't know). 100% of the quants are paid for their work.
So we have this:
97% losers - 99.9% candles
100% winners (not in the trading sense) - 100% non-candles
I know with what group I wanna be
WARNING: I'm not saying that there is some sort of correlation between using candles and profitability. So don't flame me on that. Just an observation.
BTW, I'm also using candles for MY visual display. But my analytics are done with continuous time series.
Anyway, I don't want to fight with you, you're a good HFT guy
Interesting you bring it up. It is a very efficient transform algorithm, but I find it's got issues with spline interpolation. Wavelets yield best results in my research.
I'd be really interested in any opinion you might have on non-linear time frames. For others reading this, think of an M5 candle chart, but the candles between the NY/Asia session are longer (let's say 15 minutes), while the ones during the NY/London overlap are shorter (let's say 3 minutes).
So on one chart you have different candle durations, based on the typical market activity at that particular candle time.
- #37
- Apr 12, 2010 12:26pm Apr 12, 2010 12:26pm
Like I said, I've browsed quite a lot of quant papers. I found some interesting nuggets along the way
I'd be really interested in any opinion you might have on non-linear time frames. For others reading this, think of an M5 candle chart, but the candles between the NY/Asia session are longer (let's say 15 minutes), while the ones during the NY/London overlap are shorter (let's say 3 minutes).
So on one chart you have different candle durations, based on the typical market activity at that particular candle time.
However, when it comes to quantitative analysis, you actually want to look for the "hive mind". When 3-5% of traders sell or buy a security on established time frames, and techniques like RSI, chart formations (and your other usual stock), they leave a definite imprint, which, after thousands of times can become relatively clear. So instead of watching the indicators, you watch market transformations.
-Marko
- #38
- Apr 12, 2010 1:28pm Apr 12, 2010 1:28pm
- Joined Apr 2009 | Status: Live and learn. | 25,165 Posts
"If The Fool persists in his Folly he will become wise." - William Blake
- #40
- Apr 12, 2010 3:19pm Apr 12, 2010 3:19pm
- Joined Apr 2009 | Status: Live and learn. | 25,165 Posts
"If The Fool persists in his Folly he will become wise." - William Blake