Deadline June

NinjaTrader

Just discovered NinjaTrader. This does practically 95% of what I need. Version 7 beta has portfolio backtesting with combined results.

Time to use the Fonz emoticon: (y)
 
Importing 8 gigs of tick data for each symbol, and NinjaTrader 7 has fallen at the first fence. I got Blue Screen of Death and a core dump.

This eats up my time. I've been suckered into beta testing NinjaTrader 7. If I used NT6.5, I would have to split the tick data files into small chunks under a million lines each and then import each one one by one.

This is all diverting me from my key objective at the moment - to get NinjaTrader up and running, programming my system in NinjaScript and then backtesting it, walking it forward it, simulated trading it and then hopefully going live with it.

Somewhere in that project plan I need to find about 8 hours per symbol to import the data for 9 more forex pairs - I've got EUR.USD in there already.

6 weeks and 3 days to go.

Time to apply the "Baseball Bat of Management" to my plans.
 
The timestamp on my NinjaTrader6.5 directory is 7 days ago.

The Baseball Bat of Management is asking what the **** I achieved over these last 7 days, if I'm not trading yet and if I want to achieve something in the next 7 days.

The Baseball Bat of Management is not very understanding, it's motto is "Just ****ing do it", and it reminds me of an old sales manager that I used to work for, whose big weakness was to sell a year's worth of our team's programming effort for £10 grand after 18 holes of bad golf. To be delivered in 2 months. But the Baseball Bat of Management actually comes from a department head in Bankers Trust when I worked in the Compliance Dept. He didn't let you go off sick. You'd have to be in hospital to be able to have a day off sick, or he'd phone you up and hassle you until you went in. I was never sick but I saw him do it to others. He'd patrol the open plan office with a baseball bat. I wouldn't strictly say I was intimidated by him, but I did pick up a sense of urgency about his plans. He looked like Morgan Freeman and sounded like Denzel Washington, but the content of his conversation was pure Ray Winstone.

Anyway, how did I manage to drag out what little I achieved this week to make it take up so much time?

There was the ridiculous amount of data from Disktrading that wouldn't import into TradeStation4, which lead to downloading NinjaTrader6. I then spent hours trying to work out how to operate the thing, and how to import data into it. That took up a disproportionate number of CPU cycles.

Sadly, it took up virtually a whole day to import just one symbol's tickdata for the 12 years from 1998, and I did it wrong. I had no clue that NinjaTrader would assume that the timestamps on the tickdata was in the time zone of my computer. I should have realised, but I didn't. So my timezone is London, the tickdata is NY, so I ended up with tickdata offset by 5 hours.

I just deleted it and I'm re-importing it. In NinjaTrader 7, it politely asks you what timezone the data is timestamped with. That's how I found out - who knows if I would ever have realised if I hadn't upgraded to NinjaTrader7.

So that accounts for what, 3 days? The other 4. Ermm.... OK I spent a while messing around getting NinjaTrader's "Strategy Analyzer" going, and then translating my TradeStation EL script into NinjaScript. I had to go onto the NinjaTrader Support forums to find out a few things.

Anything else? Wasted time as I posted earlier trying to get NinjaTrader7 to import a huge file. Essentially I guess I don't have enough RAM. First time, I got a blue screen of death, second time it just ground to a halt and refused to do anything else. Wasn't even using the CPU, just sat there. That's the kind of stuff you get with Microsoft - I can't figure out why they tied their colours to the Microsoft mast. Just do it in Java, then you've got platform independence. As it is, I'll have to get a new PC if I want to go live with NinjaTrader.

Oh yes right, that's a couple of hours that I spent - checking out potential new computers.

I think i spent a while trying out strategies simultaneously on multiple symbols and checking out the combined results.

I also spent a morning trying to trade the Euro.USD just using the chart. I would have lost money on that of course, but it's an outrageous challenge that I know I'll have to take up one day. All I learnt though was that I didn't have a clue. I'd like to trade just the price and the fundamentals, but I have less than no clue how.

SO: the Baseball Bat of Management says, finish your system in NinjaScript and get optimising it at least on the EURO.USD to see what the results look like over 12 years of hourly bars.

Here I go. Another lost weekend, it might seem, as the sun shines outside and Leonard Cohen drones on the hifi, but I feel optimistic about getting this show on the road perhaps even by the end of this month, 1 month ahead of schedule. I just mean trading the Euro.USD - not the whole 10 forex pairs.
 
In my own personal experience, but I am slow learner, it took me from 2002 to 2008 to develop anything worth trading in terms of automated systems, but I must add that, since starting in 1997, I've wasted 50k and most of my time trying to be profitable at discretionary trading. So, if you entirely skip that, you might reach your goal faster than me. I'd advise everyone to skip discretionary trading altogether. It's totally useless, it gives you no good ideas, it just wastes your money and time.
 
It generally takes me about 3 months to get into the groove of trading strategy development. I guess I'd say I'm in one now. I had been in this position perhaps 4 or 5 times before. Maybe the first time took a lot longer, I don't remember. The problem is moving to the next stage of implementing the plan.

If I can't sort out something I'm happy to trade using tick data, with $50K and looking for 5% per month, by June, then I'll move into day trading the Euro futures.

On a different mailing list, this one discretionary trader basically issued me a challenge:

If you want to survive at something, you first and foremost fight for yourself. You don't put all your time and energies into teaching an idiot computer how to take your place. That comes much later when you have the time and luxury to train it how to behave more like you because you're ALREADY successful at trading live markets to begin with.

You're not desperate because of the technical difficulties you're facing trying to put together a fully automated trading system. That's covering up the greater truth: you need to face the market head-on yourself and succeed there first. Then, you can do anything in the markets you set your mind to.

I don't consider myself a great trader, but I am 100% sure that you are doing the wrong things in trading to ensure your success.
I don't really agree with him about what I'm covering up, but I agree I should face the challenge.
 
Yeah, he's wrong. About almost everything he says. What makes him even more wrong is that he is "100% sure". I often add at the end of my posts something like "i could be wrong". But this guy talks as if he were infallible, which is upsetting.

You don't have to first become profitable as a discretionary trader, in order to then become a profitable automated trader. I actually just met an automated trader, who is a professor of statistics and he even loans his systems (to millionaires) for a percentage of returns. He told me he didn't place one discretionary trade in his life: just one for a technical test (so it was not a discretionary trade, but just a manual random trade) and he immediately closed it. He is profitable, his systems are so good that millionaires want them, and yet he never placed a discretionary trade.

You don't put all your time and energies into teaching an idiot computer how to take your place.
The "idiot computer" this guy talks about, is much better than a human at not doing the things that cause humans to lose: impatience, emotions...

"Idiot computer" is always objective, doesn't do any revenge trading, nor doubling up on losing positions. He doesn't get impatient nor trades out of boredom. "Idiot computer" does each time and without efforts what most humans cannot do with their best efforts.

Then, you can do anything in the markets you set your mind to.
Wow, by now this guy owns the world, because with his will power, he can do anything he sets his mind to. He's richer than Bill Gates by now, but somehow managed to stay unknown. Or maybe he is Warren Buffett. We should be honoured to have heard his advice... not! Nothing could be further from the truth. If you have stand in front of a truck, armed with all your will power, you won't stop the truck, but you will just die. You cannot do all you want thanks to your will power.

you need to face the market head-on yourself and succeed there first
This above and your concept of "facing the challenge" is exactly the attitude that caused me to lose money for 12 years of discretionary trading. Wanting to do what I could not do. Sure, a few people can control their emotions, their impatience, and all that you need to control while trading. But I couldn't. Why did I have to be so stubborn and go the hardest way about it? Probably because I used to be like that guy: I felt I could succeed at anything I set my mind to. And I didn't want to "quit" ("quit" meant in a negative way). But instead with some things it's better to "quit" immediately.

As I said in another post, "quit" can be meant in two ways:
1) "quitting" a race, which is considered bad
2) "quitting" drugs, which is considered good

With trading you don't really know if you're dealing with a drug or if you're in a race, so you don't know if you should "quit" because you're trading out of addiction (compulsive gambling), or if you should not "quit" because you shouldn't give up until the finish line.

I think you should quit in both cases, because even if had no elements of addiction and if it were just a race, you should not be in a competition just for the sake of it. Competing is not good in itself. "No pain no gain" is bull**** in many cases. You can achieve more by using your brain than by using your pain.

Just thinking out loud. I am not 100% sure about anything I said.
 
Last edited:
Importing 8 gigs of tick data for each symbol, and NinjaTrader 7 has fallen at the first fence. I got Blue Screen of Death and a core dump.

Carefull - this is NOT a ninja problem at all. Ninja trader is a use level application, not installing anything in kernel level. At most it should crash, taking itself down.

A BSOD starts with a kernel level error - hardware, graphics driver, some other driver are the culprits of that. Could be the operation first time stressed your memory and it "broker". But whatever it is - a normal application CAN NOT trigger a BSOD. Simply not possible (taking out driver errors, but then this clearly is a driver problem).
 
Just thinking out loud. I am not 100% sure about anything I said.

:jester:

As you said. I'm sometimes 100% sure about something but I don't say it like that guy said it and yes he was wrong on a lot of fronts and I guess he only ever made enough money to pay the rent, but at the end of the day I'm sure I could learn and succeed at discretionary trading and the difference between me and you then I guess is that I want to.

But I admit his message did need a good wipe to get rid of the excess testosterone on it :LOL:
 
Carefull - this is NOT a ninja problem at all. Ninja trader is a use level application, not installing anything in kernel level. At most it should crash, taking itself down.

You just reminded me how much I despise Microsoft and all its creations and how much I love Linux, but the problem is, NinjaTrader does everything I want and it's there ready and waiting if I can only bring myself to use Windows.

I wonder if Windows 7 is more stable. I'm going to upgrade my hardware before I go live and the new system's probably going to have Windows 7 bundled on it. :sick:

Can you recommend a good RAM tester, or better, something more comprehensive that sorts out defragging, registry, etc like Symantec Utilities?
 
Well, this is not a NInja problem still. A BSOD is the equivalent of a kernel panic under linux - normal software can not cause it.

Win 7 is a LOT better and has actually a good ram tester included ;) Started with Windows 7 - you can from the control panel boot into a ram tester. RAM is often the culprit - sucks, but there is a lot of cheap material out in the wild, and RAM is a typical culprit. If you need more, there is always http://www.memtest.org/ where you can get an ISO file to burn ;)

Symantec utilities - I dont like crap, and all from Symantec sadly is crap. I dont use any defragging except the one on board, and a registry clenaer is not wanted ;) My computes only have on it waht I need - for playing games I boot into secodary installs ;) I use a lot of virtual machines to isolate programs. Seriously, you dont need that stuff.

Half of the crashes fom forom ridiculous software people install, and sadly Symantec accounts for a lot of it.
 
Well, this is not a NInja problem still. A BSOD is the equivalent of a kernel panic under linux - normal software can not cause it.

Win 7 is a LOT better and has actually a good ram tester included ;) Started with Windows 7 - you can from the control panel boot into a ram tester. RAM is often the culprit - sucks, but there is a lot of cheap material out in the wild, and RAM is a typical culprit. If you need more, there is always http://www.memtest.org/ where you can get an ISO file to burn ;)

Symantec utilities - I dont like crap, and all from Symantec sadly is crap. I dont use any defragging except the one on board, and a registry clenaer is not wanted ;) My computes only have on it waht I need - for playing games I boot into secodary installs ;) I use a lot of virtual machines to isolate programs. Seriously, you dont need that stuff.

Half of the crashes fom forom ridiculous software people install, and sadly Symantec accounts for a lot of it.

Thanks for the warning on Symantec. I always thought they were meant to be super brilliant. Guess I listened to their marketing hype too much.
 
almost a month left

Playing with EUR.USD now.

Using the same system that I had for daily bars, but this time with tick data.

Still importing the tick data into NinjaTrader. Really hope there's no deal-breaker in NinjaTrader's bug portfolio, I don't want this to be wasted time.

I'm looking at the average true range across different time frames.


daily 0.0130
60min 0.0030
30min 0.0022
15min 0.0018


So I gave up a big chunk of volatility going from daily to 60min bars, but the next step down the decrease in volatility is a lot less, and from 30 to 15 min bars is even less.

Obviously another parameter to test in my optimisation routine.

Really running the clock down here. I want to find something by Thursday because I'm away for another long weekend, in Zurich. It will be good to have got it under my belt, if I can.
 
more nose to the grindstone

Actually importing 9 gigs of tickdata into NinjaTrader is not that difficult, it just takes 3 hours. What is bad though is when it goes wrong, wastes 3 hours and more, and requires finding bad ticks in the 9 gigs of data. Almost like a needle in a haystack, except really slow because the editor doesn't want to load up 9 gigs into memory.

Discovered that NinjaTrader optimizes a basket of instruments nicely, but won't combine all the trades from all instruments, so won't give me some of the combined stats that would be nice, like number of open trades at one time, maximum daily drawdown, etc.
 
Sharpe Ratio

Now I've got at least half of my tick data into NinjaTrader, I'm trying out some strategies (what NinjaTrader calls the scripts for the trading rules).

I look at the results across a basket of 5 forex pairs and see that my Sharpe Ratio is 0.08.

That mean nothing to me, until I read in the Help that the author consider 1.0 to be good and 3 to be great.

Let's see what other people think. I've got a new thread for this:

Sharpe Ratio
 
Going to Zurich now for a long weekend, so I can knock 5 days off my remaining time straight away now if I liked - but fortunately the whole thing looks good. I'm starting to get results from backtesting and optimising on hourly bars in the forex markets and the returns look promising for a beginning.

So less than a month to go practically - times are getting interesting.
 
inputting CADUSD

This'll take 4 hours - time to water the toms.
 
Man from UNCLE

After Travis's great post My drawdown should not exceed my "uncle point" I thought about it and came to the conclusion that you should be able to train yourself to have a deeper Uncle point.

The maximum or deepest Uncle point you can have is when you are some kind of non-emotional type or autism suffer, and you can emotionlessly watch your account go to zero. Not so good.

On the other hand, logically, if you are Spock or some other Vulcan, you should be able to work out a range of performance statistics that describe your back-tested or simulated trading returns, which gives you the minimum worst loss figures that you should expect. Then Mr Spock decides logically how much further beyond that simulated loss scenario he is prepared to go. How would you do that logically? At what point? How much further beyond the worst you have seen in testing or simulation? 20%? 50%? Is there any logic you can use to guide yourself? I am not sure.

And did Mr Spock ever panic? No. So I guess he would have made a plan beforehand at what point he should place his system stop loss. And had it ever been reached, he would have calmly stopped trading the system. That is of course if Captain Kirk hadn't at the last minute discovered what the problem was ;)

So Mr Spock's Uncle Point would be exactly where you would logically place your system stop loss. The question is, how do I train myself to be Mr Spock?

That is illogical, Captain.
 
Importing more bad data

Really trying to keep everything on track here with my trading system development, optimisation and testing and keep getting knocked off the straight and narrow onto some nice but dim tangent.

Got another problem with Disktrading data, in the EUR.CHF tick data. Has problems importing into NinjaTrader, "timestamp on line 133456898 less than previous timestamp" and after lots of CPU usage I finally get the line up on screen to find out that it's a bug in NinjaTrader so I have to split the file at that point and load the two files seperately boohoo sob sob I should just say **** the EUR.CHF.

There's the Pomodoro Technique: you put a kitchen timer on 25 mins and stick religiously to your task until the pinger goes. I might need this.
 
Well congratulations on entering the real world.
From what you have been writing i think everything has been thoroughly planned and the risk factor is not so high. So are you nervous about being on your own or what?
This way at least you know you only have yourself to look after if things don't work out, but forex traders tend be a stubborn breed and they manage to to strive in all sorts of conditions. Keep posting and then we will all know how you are doing. Until you start in June it is going to be hard to know what is going to happen but I wish you all the best.
forex.lady
 
Top