Australian (ASX) Stock Market Forum

Pull ya DAX up!

Jump to new
Gringotts Bank said:
Yeh everything is fitted, but not everything works on unseen data.

I had to look up sigma. Six sigma came up a bit on google - basically a big move (6 std devs)and what happens next. http://www.marketcalls.in/trading-l...-about-six-sigma-events-in-stock-markets.html
Click to expand...

Well I guess that's the risk we're willing to take by trading, can only minimise the risk, make the best call we can based on recent history/price action. Every trade we take is going to play out on unseen data, otherwise you have a serious gift and can see into the future :D

Yeah there are the outliers which would also provide opportunities, on the day and after it I would imagine.
 
ThingyMajiggy said:
Well I guess that's the risk we're willing to take by trading, can only minimise the risk, make the best call we can based on recent history/price action. Every trade we take is going to play out on unseen data, otherwise you have a serious gift and can see into the future :D

Yeah there are the outliers which would also provide opportunities, on the day and after it I would imagine.
Click to expand...

You missed the point. Testing on unseen data is possible with the right software. If the optimized parameter fails on unseen data, the edge isn't really an edge. If it passes, then it's a tradable edge.
 
Joules MM1 said:
Fari Hamzei (constant top 5 US market timers according to timerdigest.com) is now making calls on $DAX
he's a big user of sigmas ...take some time to check out his work....at least watch a cupla vids of his if you want a diff perspective

https://twitter.com/HamzeiAnalytics
http://www.hamzeianalytics.com/Educational_Webinars.asp
Click to expand...
Thank you once again, Joules!

Is timerdigest worth subscribing?

Mathematical analysis has been an eye-opener and I see a vast potential in this area. I have only used it on equity trades. Been mostly trading based on understanding of fundamentals, news, calendar events and my half-baked comprehension of chart patterns.... and mostly using excel and manually importing data from yahoo.

I have enrolled in couple of financial markets courses and programming courses via coursera - am currently considering quandl for my data source. I have yet to figure out if ninjatrader can do what I want. Lots to explore over the next few weeks.

rvm
 
rvm said:
Thank you once again, Joules!

Is timerdigest worth subscribing?

Mathematical analysis has been an eye-opener and I see a vast potential in this area. I have only used it on equity trades. Been mostly trading based on understanding of fundamentals, news, calendar events and my half-baked comprehension of chart patterns.... and mostly using excel and manually importing data from yahoo.

I have enrolled in couple of financial markets courses and programming courses via coursera - am currently considering quandl for my data source. I have yet to figure out if ninjatrader can do what I want. Lots to explore over the next few weeks.

rvm
Click to expand...

timerdigest, no
definitely worth the time to watch cupla of Fari's edu videos, he uses self-written proprietory indicia + sigmas
and has developed his own routines based on those indicia...if nothing else it'll open a few ideas
 
Gringotts Bank said:
You missed the point. Testing on unseen data is possible with the right software. If the optimized parameter fails on unseen data, the edge isn't really an edge. If it passes, then it's a tradable edge.
Click to expand...

Yep I don't get the point. So its just random data it tests against or something? I don't get how it can forward test when I'm using data/info from one instrument. I'm calculating std dev from recent months, and updating every week. I don't understand how or the point of forward testing that. Its not a set of rules or mechanical.
 
ThingyMajiggy said:
Yep I don't get the point. So its just random data it tests against or something? I don't get how it can forward test when I'm using data/info from one instrument. I'm calculating std dev from recent months, and updating every week. I don't understand how or the point of forward testing that. Its not a set of rules or mechanical.
Click to expand...

All you need is software which does walk forward testing (eg. Amibroker). Say you're trading ES. You can choose how AB will break up the historical data of this single instrument, so that some is in-sample (training or 'best fit' data) and some is left as out-of-sample (unseen).

eg. Use 2000-2001 to optimize the best SD. AB finds best SD = 1.6. Now it will take that number (1.6) and run it on data from 2001-2002 to see whether it's still profitable. 2001-2 is data which has had nothing to do with the optimization/fitting process, so it can be regarded as unseen. This way we get to know whether an optimized number from last week/month/year will hold up and be profitable today and going forwards.

Howard Bandy has written a lot on this.
 
Gringotts Bank said:
All you need is software which does walk forward testing (eg. Amibroker). Say you're trading ES. You can choose how AB will break up the historical data of this single instrument, so that some is in-sample (training or 'best fit' data) and some is left as out-of-sample (unseen).

eg. Use 2000-2001 to optimize the best SD. AB finds best SD = 1.6. Now it will take that number (1.6) and run it on data from 2001-2002 to see whether it's still profitable. 2001-2 is data which has had nothing to do with the optimization/fitting process, so it can be regarded as unseen. This way we get to know whether an optimized number from last week/month/year will hold up and be profitable today and going forwards.

Howard Bandy has written a lot on this.
Click to expand...

Yeah I understand the process, but it doesn't make sense with what I'm doing and what I've made here. There is no 1.6 like number that can be used to optimize or tweak what I've made, what I'm calculating IS the std dev of the range of X period and it's just drawing that for me on the open of each day, it's not like one number is going to be better than another number, just different time period std devs, daily, hourly etc. there's just one number, not like indicator settings where you have multiple options to change to see how it changes performance etc.

I am using the most recent market data available to us so it's as recent as I'm going to get therefore the most valid data to use, there's no point me calculating the std dev of the range from a period of time back in 2000 or back in the early days when the Dax was created as the daily ranges were stuff all and projecting that on today's trading session as it's fairly irrelevant, hence why I'm doing just 2016, or the last 12 months and then updating it weekly from then on.

The better calculations to be doing would be to do what Can and myself mentioned earlier and figure out how often in % terms we hit an upper/lower band or two levels up/down and stuff like if we hit two levels up/down what are the chances that we don't, or that we hit the opposite levels tomorrow etc.

I am just curious to see how these(if at all) help in my trading as it seems like an idea that makes sense to me, they seem to be helping me with expectations of how the day will play out and some decent entry/exit points, in conjunction with everything else obviously, just another aid/tool in the process.

I have been away(been answering on my phone over the weekend) but am back now to see how Friday and Monday looked as I wasn't here to watch. Friday didn't hit either first levels up or down, just bounced around between the hourly levels, quite nicely, definitely some nice places to enter where it bounced. Last night we hit one std dev lower and hovered around it basically bouncing around between the hourlies again with a great entry short where the gap closed, it all lines up.

FDAX 06-16 (3 Min) 26_04_2016.jpg
 
6 trades, 5 winners, +107 points, +$3900 AUD.
Largest winner 38 points, largest loser half a point. Did a bit of 2 lot trading tonight so I could hit a target, then let the other go.

Bands giving a good outline again, another nice play around the high + gap close. I have also added weekly and monthly std dev bands(not visible on this chart). We hit level one to the down side yesterday, which also lines up with level one on the hourly levels today, which may or may not have helped. Let's see how the day plays out :)

FDAX 06-16 (1 Min) 27_04_2016.jpg
 
kid hustlr said:
Cool mate - same here.

Just got excited when you wrote 4k day thats all :)
Click to expand...

Yeah I probably wouldn't say anything if it was live, mainly use this thread for sim/testing things, doubt anyone would care or read this to justify posting all my details, so I'm surprised you're even reading here :D Cheers! All the best with your trading :xyxthumbs
 
ThingyMajiggy said:
Yeah I probably wouldn't say anything if it was live, mainly use this thread for sim/testing things, doubt anyone would care or read this to justify posting all my details, so I'm surprised you're even reading here :D Cheers! All the best with your trading :xyxthumbs
Click to expand...


Hey Sam ..... Just bumping this thread after reading through it again in its entirety last night.

No pressure intended so whether you carry on with anything etc is totally up to you but given some of the results you had achieved over a short period i reckon you were pretty close to something potentially life changing in your trading so this is just a post to encourage you to keep cultivating your ideas:xyxthumbs Cheers.
 
barney said:
Hey Sam ..... Just bumping this thread after reading through it again in its entirety last night.

No pressure intended so whether you carry on with anything etc is totally up to you but given some of the results you had achieved over a short period i reckon you were pretty close to something potentially life changing in your trading so this is just a post to encourage you to keep cultivating your ideas:xyxthumbs Cheers.
Click to expand...

Wow thanks barney! Surprised you maintained enough interest to read the entire thread :D Sorry I totally missed your post with all the other threads being more active.
 
ThingyMajiggy said:
Wow thanks barney! Surprised you maintained enough interest to read the entire thread :D Sorry I totally missed your post with all the other threads being more active.
Click to expand...

Lol .... Its a good thread:cool:. Some of the trades you started generating were pretty impressive so whatever ideas you were working on would definitely be worth following up I reckon.
 
After being encouraged from barney and a new year with hopefully more time, I think it's time I was more rigid and did this properly. So might as well line it up with the start of the year and trade sim for a few months, or however long it takes before my results look decent again, feeling very rusty as you'll see from my attempt at trading the aftermath of non-farms last night :D

I have at times funded a CFD account trading the indices and occasionally some commodities with real money, did well off Brexit and Trump's election in 16, and also turned 1.5k into about 8k in around 4 weeks or so. Then had a string of a couple days of stupid decision and had some losses that really made me angry at myself, I didn't blow up the account or anything, just was down about it and mad at myself and also had to withdraw a big chunk of it for things outside of trading, so that was the end of that for a while.

But I had multiple little cracks at the market like that with just fiddling around with 1k or so, I did a motorbike tour through Vietnam in October and paid for the trip when I was over there with a nice trade, so that was awesome.

I'm also off to South America in Febuary this year. But will probably take my laptop and try to do some trading over there as I'll be there for 2 months so I really wanna try and keep it up over there as well.

Will post my horrendous trading for Thursday and Friday a bit later, am about to head out and melt in this 41 degree heat.
 
ThingyMajiggy said:
did well off Brexit and Trump's election in 16, and also turned 1.5k into about 8k in around 4 weeks or so.
Then had a string of a couple days of stupid decision and had some losses that really made me angry at myself, I didn't blow up the account or anything, just was down about it and mad at myself
paid for the trip when I was over there with a nice trade, so that was awesome.
Click to expand...

Good stuff Sam! .... I see a 2:1win ratio above already:D

1.5K to 8K = over 500% return in 4 weeks is in the gun trading category:cool: .... There will be plenty of fellow punters interested so good luck with it.
 
So this is the hilarity that was last night, shameful to look at after it's all happened, thought I'd stay up and watch the NFP release, can sometimes get a nice trade out of it, the session was terribly quiet in the lead up to the release, only a 40 point range or so on the dax, which is pretty quiet for the dax, everyone still on holidays I guess. Anyway as you can see from my sea of entry and exit arrows, fade, exit, fade, exit, fade, exit, finally got the drop back to pre-release area which made up for the nice hole I'd dug on the way up, then went to bed. Woke up this morning to see the rather nice trend after that but oh well, bed was calling! :D

I am still fiddling around with my StdDev lines, seeing if I can improve them at all, they seem to be a reasonable guide for areas to at least pay a little more attention. But you do get those outlier days where it just goes beserk in one direction all day and blows any multiples of standard deviation out of the water.

Was/am also investigating some machine learning stuff, but there's a lot to learn with that, Math, Python etc. So that can be like a side goal maybe, I'd rather work on myself and my own trading first. I have reasonable confidence, I know I've done it before and can do it again. Just needs fine-tuning, I tend to want to chase the market so I get filled, when I should probably leave my original order in and let it get hit, but it's hard to tell sometimes with the Dax, it can take off in a hurry so you have to be ready and anticipate! Also need to improve trade management, the two nights I traded this week prove how important that is, with 3 winners and 7 losers yet still profitable.

Also traded Thursday night, so only the 2 nights but for the week:
Total ticks/profit: 53/$662.50
Total trades: 10
Winners: 3
Losers: 7
Avg winner: 22
Avg loser: 6
 

Attachments

  • 6-1-17.PNG
    6-1-17.PNG
    70.8 KB · Views: 145
Market hasn't been open long but feeling tired and have to go to bed soon, damn early starts at the moment :(

Total ticks/profit: +101/1,262 EUR
Total trades: 6
Winners: 4
Losers: 2
Avg winner: 28 ticks
Avg loser: 6.5 ticks
 
Top