Systematic trading | Page 13 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Trade another day through Systematic Trading 203 replies

Nonparametric Association Measures in Systematic Trading 57 replies

Anyone use Systematic from Alpari? 1 reply

A Systematic Approach to Markets 68 replies

Systematic Trading vs. Discretionary Trading 14 replies

Attachments: Systematic trading
Exit Attachments

Systematic trading

  • #241
  • Aug 12, 2009 2:06am Aug 12, 2009 2:06am
I brought the high frequency by Olsen (of Oanda fame), to be honest it didn't have much useful in it (for me), the contents of the book you are looking at seem very similar.
The breaking of a wave cannot explain the whole sea.
  • #242
  • Aug 12, 2009 2:35am Aug 12, 2009 2:35am
Except from stocks, I would be interested to combine metals (or other commodities) with FX pairs and see what comes out. Does anyone have any experience about this? I suppose we would need to trade a future on the metal, right? Is there other way to "bet" on metals? spread betting perhaps? (but I realise spread betting is probably available only for the UK)

Quoting Craig
Without wanting to flog a dead horse, as they say in the article, this is also one of the advantages of stocks, you have more markets to choose from, one can construct synthetic pairs etc, where as in forex it's all a bit circular. With crosses, you really need ECN access to get the lower spreads.
  • #243
  • Aug 12, 2009 2:48am Aug 12, 2009 2:48am
Quoting challenger78
Except from stocks, I would be interested to combine metals (or other commodities) with FX pairs and see what comes out. Does anyone have any experience about this? I suppose we would need to trade a future on the metal, right? Is there other way to "bet" on metals? spread betting perhaps? (but I realise spread betting is probably available only for the UK)
ETFs perhaps?

Here is some random link from google: http://etftradedfunds.com/metals-etf
  • #244
  • Aug 12, 2009 4:44am Aug 12, 2009 4:44am
Quoting mikkom
By the way has anyone read this book? I'm thinking of ordering it, especially the parts about market microstructure sound very interesting.

http://eu.wiley.com/WileyCDA/WileyTi...fContents.html

edit: uh apaprently no because it's not released yet
She has a web site on HFT where you have to pay for access. I was far from impressed when I tried the service, but since it's pretty much the only book on HFT I will definitely give it a shot when it is released.

Olsens book contains some interesting information on micro structure and how to process high frequency data, but it is a bit outdated.
  • #245
  • Aug 12, 2009 12:54pm Aug 12, 2009 12:54pm
I'm looking for more sites to read - I'm mostly looking for sites about systematic/quant aspect of trading so please dump your favourite sites.

Here are some sites that I read, some frequently, some [very] infrequently - some are much more dead than others.

http://www.nuclearphynance.com/

http://www.wilmott.com/index.cfm?NoC...=Yes&forumid=1

http://www.quantnet.org/forum/

http://www.tradingblox.com/forum/
  • #246
  • Aug 12, 2009 2:17pm Aug 12, 2009 2:17pm
Good blog: http://marketsci.wordpress.com/
The breaking of a wave cannot explain the whole sea.
  • #247
  • Aug 12, 2009 5:36pm Aug 12, 2009 5:36pm
If your into automated trading there is no way around

http://www.elitetrader.com/vb/forumd...?s=&forumid=32

Also this one is good

http://www.maxdama.com/
  • #248
  • Aug 13, 2009 7:33am Aug 13, 2009 7:33am
I'm pondering on using amount/significance/numbers on news releases as predictor of volatility. There are some obvious pre-volatility effects like usually week before NFP the volatility is lower, on NFP day, volatility is higher and so on.

I would like to compare volatility to number of news (probably weighted by importance), variability from predicted value and so on.

So my question is, is there any longer term data available on news releases, something like the front page of FF but longer term.

I don't mind paying for it if it's reasonably priced (I don't need articles or anything like that, just dates/times and predicted values+actual values)
  • #249
  • Edited 10:07am Aug 13, 2009 9:27am | Edited 10:07am
10 years of minute data (2,8 million candles), simple donchian 4/1 week system with fractional position sizing = ~1.1 seconds runtime. After all the optimization looks like this is best I can do. Quite an improvement to Metatraders horrible testing efficiency anyway.
  • #250
  • Aug 13, 2009 2:30pm Aug 13, 2009 2:30pm
Quoting mikkom
So my question is, is there any longer term data available on news releases, something like the front page of FF but longer term.

If you only need US news then you can crawl these sites:

http://biz.yahoo.com/c/e.html

http://www.briefing.com/GeneralConte...ate=08-13-2008

We build a database this way in connection with some research we did a while back. As I remember the data goes back to 1998.

If you need EU or other news, then you could go to e.g

http://www.economy.com/dismal/

- but it is a pay service. It seems that non-US news are difficult to get to if you want +10 years of data. You could of course try Bloomberg or Thomson-Reuters but they will probably rip you.
  • #251
  • Aug 13, 2009 2:54pm Aug 13, 2009 2:54pm
Here is a good source for euro data (if someone is interested)

http://epp.eurostat.ec.europa.eu/por.../bulk_download

The thinng is that I'm spoiled and trying to find data similar to ff front page, csv or xml. Does anyone know where ff gets the data?
  • #252
  • Aug 13, 2009 3:11pm Aug 13, 2009 3:11pm
Quoting mikkom
Here is a good source for euro data (if someone is interested)

http://epp.eurostat.ec.europa.eu/por.../bulk_download

The thinng is that I'm spoiled and trying to find data similar to ff front page, csv or xml. Does anyone know where ff gets the data?

I don't know if can be useful, but if u open the folder of every news on the ff front page you find the source with the link.

Lubo.
If I know how, I tell You!! Few but Good!!
  • #253
  • Aug 13, 2009 3:34pm Aug 13, 2009 3:34pm
Quoting LuboLabo
I don't know if can be useful, but if u open the folder of every news on the ff front page you find the source with the link.

Lubo.
Maybe I should hire few indians [or any other low wage country people who are advertising their work online] to write a csv of news events in the past.. Not really an impossible option.
  • #254
  • Aug 13, 2009 4:07pm Aug 13, 2009 4:07pm
Quoting mikkom
Maybe I should hire few indians [or any other low wage country people who are advertising their work online] to write a csv of news events in the past.. Not really an impossible option.
I don't know.... , maybe,...... but u have to find people able to do it....

BTW ...sorry for the useless info.

Lubo.
If I know how, I tell You!! Few but Good!!
  • #255
  • Aug 13, 2009 4:13pm Aug 13, 2009 4:13pm
Quoting LuboLabo
I don't know.... , maybe,...... but u have to find people able to do it....
It's not really an impossible option, I have been in IT business for a while (10 years or something) and I know who to ask/where to look.. Really thanks for the comment, I really didn't think of that option before (also after that the copyrights for the stuff will be mine which is a plus because I'm probably not the onyl person who needs this stuff - yes I have been in business too long).,
  • #256
  • Aug 14, 2009 1:44am Aug 14, 2009 1:44am
Quoting mikkom
10 years of minute data (2,8 million candles), simple donchian 4/1 week system with fractional position sizing = ~1.1 seconds runtime.
2.8 million candles in 1.1 second...??? Oh my... long way to go (for me) to achieve something even near that.
Very impressive!

Thanks for inspiration.

Bluefish
  • #257
  • Edited 3:05am Aug 14, 2009 2:28am | Edited 3:05am
Quoting Craig
Good blog: http://marketsci.wordpress.com/
Thanks a lot for that! It will be on my list from now on.

Keep the good links coming if you have them!
  • #258
  • Aug 14, 2009 8:37am Aug 14, 2009 8:37am
Quoting bluefish
2.8 million candles in 1.1 second...??? Oh my... long way to go (for me) to achieve something even near that.
Very impressive!

Thanks for inspiration.

Bluefish
Mikkom, has been quite successful doing forex and has one of these in his attic:


http://www.cray.com/Home.aspx

(-:

Mikkom thanks for pointing out the proportional sizing of positions! Results and risk control are getting more impressive with this in place!
  • #259
  • Aug 15, 2009 5:22am Aug 15, 2009 5:22am
Mikkom, carrying over from another thread, you've said you found fixed %R the most profitable in your strategies. It makes market sense to me because with a fixed amount per pip you inadvertently tilt your strategy to favour days of higher volatility, which isn't necessarily a good thing.
But wouldn't weighted position sizing surely improve a strategy if you could determine which conditions had a higher probability of success than another? This is obviously very strategy dependent - personally, I've seen some weak correlations between certain conditions and profitability on the strat I trade but they're not strong enough to make me move away from fixed %R - but isn't this something to be aiming for?
  • #260
  • Edited 6:00am Aug 15, 2009 5:35am | Edited 6:00am
Here are the results for 4/1 week donchian breakout since 2001 if anyone is interested (based on minute data) 2% risk per trade.

The framework is still quite rough as you can see but it works.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot-7.png Size: 184 KB
0 traders viewing now