Deadline June

[Adamus]

3 weeks 6 days to go

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.

That's the truth. Thanks for your wishes.

I hope I don't need 3 weeks and 6 days. I think I can get this system on the road quite soon.

First up, I have to polish off the trading system. At the moment, I'm optimizing a trading system which buys or sells on pull-backs and exits on a trailing stop, using a moving average based filter to trade in the direction of the trend.

I'm not entirely happy with the results, although they are already trade-able, but I haven't finished entering all the tick data for backtesting yet and I'm assuming my optimisations will pass the walk-forward tests - the system did on EOD data earlier back in March.

I want to try it with parabolic stops rather than the simple trailing stops, and then I'll put it on the simulator with the IB paper trading account. I have to purchase a NinjaTrader execution license for that.

And while I'm in simulated trading, I want to get the trades for the whole basket of currencies out of NinjaTrader and into my own analysis software, to see how many trades it puts on at any one time, get some correlation stats for the returns from each currency, and see what the maximum time in drawdown was over the ten year backtest period.

Then I'll turn it on live, although I'm expecting a whole raft of problems with implementation in simulated trading, possibly also finding out I need a new PC.

 

[Yamato]

If you're someone who can stick to his plan, that's a very good start and that's saying a lot about your potential profitability. I was never able to stick to any plans for longer than a a few weeks (even when they worked). Also, someone who can watch the markets for a few months, like you did, and not place a single trade seems very promising to me. You have the right attitude and psychology for the markets: patience and self-control. As long as you don't catch the gambling addiction once you start trading, but that is unlikely.

On the other hand, I cannot look at the markets and do nothing or just paper trade. If I have capital and I look at the markets, pretty soon I feel compelled to place a trade (I see opportunities that I just cannot miss). Then, once I'm trading, I don't stick to any plans, and I'm out of control: especially as far as exits. If the market moves against me, I double up and all that stuff I explained on my journal. So, I couldn't tell someone unprofitable (like me) how to become profitable, but I can tell when someone is different from me, and you seem to be quite different in all the things I mentioned.

I think there's a small minority of people who have what it takes for profitable discretionary trading, and they're profitable pretty much from the start. The others have ego, self-control and patience problems, and it's close to impossible for them to become profitable discretionary traders. After almost 13 years of losses, I can safely assume that. Yes, I am probably the most untalented person as far as discretionary trading, but I am quite confident that I share many personality traits with the large majority of unprofitable traders. These are my conclusions after trading unprofitably for 13 years and writing many posts about this topic (e.g.: here): you need a certain personality to be a profitable discretionary trader, more than intelligence and knowledge. If you don't have it, it's almost impossible to become profitable just as it is almost impossible to change your personality.

I have acquired a lot of knowledge about the markets and certainly i am not stupid, but I am not profitable, and never have been. Speed doesn't do it. The right broker with cheap commissions doesn't do it. Having an edge doesn't do it, because even just 10% of losing trades, if handled wrongly, can wipe you out. The right instruments (futures) won't make you profitable either. I am just as unprofitable today as I was when i started. In a way my personality makes me stupid. I know by now that I should not double up on a losing trade, but close it. I shouldn't put myself in a kamikaze position where I either am right on a trade or I blow out my account. And yet I do it every time, why? I am not stupid, not by the dictionary. When I am trading, my pride and impatience blinds me and takes over. There's no reasoning any more: in fact I do become stupid.

What maybe could be done is to be profitable via automated trading, but even then you need to be able to not place any trades and that is really hard when you're looking at the markets more than once a day, to make sure your systems are behaving correctly.

 

[Adamus]

If you're someone who can stick to his plan, that's a very good start and that's saying a lot about your potential profitability. I was never able to stick to any plans for longer than a a few weeks (even when they worked). Also, someone who can watch the markets for a few months, like you did, and not place a single trade seems very promising to me. You have the right attitude and psychology for the markets: patience and self-control. As long as you don't catch the gambling addiction once you start trading, but that is unlikely.

Nobody's perfect and I'm certainly not nobody. I should have been trading a year ago. But I got burned twice before and the fear from that makes me go off on the nearest decent looking tangent that allows me to kill some time and not get round to trading.

Now I've got a handle on that, I can generally recognise when I'm veering off course and I can haul myself back onto the straight and narrow without wasting more than a few hours or a day.

Everyone's got their cross to bear.

I find it fascinating that you have no fear and you dive in again and again. That's a major plus point in my books, although of course it doesn't help if you only dive in to blow yourself up.

 

[Adamus]

I have acquired a lot of knowledge about the markets and certainly i am not stupid, but I am not profitable, and never have been. Speed doesn't do it. The right broker with cheap commissions doesn't do it. Having an edge doesn't do it, because even just 10% of losing trades, if handled wrongly, can wipe you out. The right instruments (futures) won't make you profitable either. I am just as unprofitable today as I was when i started. In a way my personality makes me stupid. I know by now that I should not double up on a losing trade, but close it. I shouldn't put myself in a kamikaze position where I either am right on a trade or I blow out my account. And yet I do it every time, why? I am not stupid, not by the dictionary. When I am trading, my pride and impatience blinds me and takes over. There's no reasoning any more: in fact I do become stupid.

Two things: the first is humorous - there are gangsters out there who sometimes need to lose money, I don't know why, something to do with money laundering, but they need to lose it for real and have the account statements which show they lost it. They probably re-use the statements multiple times or something. Anyway, you could work for them. Tell them you'll lose on the tenth trade guaranteed. The whole lot gone in one orgy of loss.

Secondly, I think you have to fix the trading issue you have with being able to take a loss. Otherwise no trading is ever going to work for you. I don't know how, but there must be a hundred different techniques out there to train yourself. Your brain after all is just an organ, you can change it. Neuro-linguistic programming or zazen meditation or something. Maybe kickboxing. Take up full contract kick boxing. The pain of losing a round would be comparable to taking a loss in the markets. Whatever. Good luck.

 

[Yamato]

Eh eh, kick boxing... why don't i shoot myself in the balls each time I double up on a losing position or just don't close it? Seriously, I don't think it will work, ever. I don't believe we can change our personality as much as you think. If you can do it, it's because you're ready for a change or not that far from it. Maybe in 20 years, I'll be able to trade discretionary. Right now I take it too personally. I cannot be objective. It will be like this for a long time. It's not up to me to be ready for that change. If I'll be a different person, maybe 20 years from now, then it will work - but not because of my efforts.

Or say I was in a different situation, like I either make money or I die, then it might work and I might change. But this is not the case right now. I am not that desperate, and I am glad that I am not. Also, I can't pretend I am that desperate in order to be profitable, because I can't bluff myself.

It's as if you told me: keep the pastries and the cakes at home, and keep yourself from eating them. It's not easy and maybe not even possible. The easier thing, which should be done, is not keep cakes at home. I can't control myself at trading, so I should not trade. But even that is quite hard.

It will be a good achievement if I'll manage to at least keep away from discretionary trading in the future.

 

[Adamus]

but you've proved you can't keep away from discretionary trading. it seems you've proved you can't find a solution that allows you to trade and leave your psychoses unchanged.

 

[Adamus]

How about this: I'll put out a contract on you this time next year unless you make a ton of money trading. Won't cost you anything - if you succeed, I won't need to. If you don't succeed, you'll be dead. That way it also puts pressure on me to succeed to pay for the hit. Oh and don't take too long to read this post - I'm deleting it in 24 hours :jester:

 

[Adamus]

another glitch in the data

The NinjaTrader Historical Data Manager is quite good at catching and rejecting errors in the tick data that I'm trying to import. I should be grateful but every error takes half an hour of faffing around to sort out. Just got another one in the EURCHF file.

edit: found out it was my own fault. the zipped file from the data provider was fine. it must have corrupted during the unzip or the split processes. an excellent opportunity to blame windows for not having true multi-tasking. i shouldn't have tried doing anything else on the PC while I was unzipping and splitting them.

 

[Yamato]

How about this: I'll put out a contract...

I see these options in front of me, with different outcomes:

1) keep on doing discretionary trading and probably lose forever and even become homeless (right now I am in debt with my bank because of it).

2) try to quit discretionary trading but keep on doing automated trading, and if I succeed, I might make a lot of money, but it will take a long time - unless someone donates 50k to me (with the risk I'll gamble it).

3) close all trading accounts and quit all trading (both discretionary and automated) and this certainly would stop my compulsive trading, but, after all this work, I would never want to do it.

So it's either, take a risk of becoming homeless but also have the hope of making some money, or take no risks, quit everything, and have no hopes.

If I were to compare my compulsive trading as an automated trader with the compulsive pastry eating of a baker, it would go like this:

1) keep on working at the bakery and keep on eating sweets (and I'll become fat)

2) keep on working at the bakery, but try to not eat any sweets (automated trading forces me to look at the markets)

3) quit my job at the bakery, and this would most likely stop my pastry eating, but I am a very good baker after all these years, and I would never want to stop.

So basically, I want to keep on working at the bakery, and here's what I'll see every day:

 

[Adamus]

Eh eh, kick boxing... why don't i shoot myself in the balls each time I double up on a losing position or just don't close it? Seriously, I don't think it will work, ever. I don't believe we can change our personality as much as you think. If you can do it, it's because you're ready for a change or not that far from it.

I have to disagree. Your 'personality' is not some black box inside your head. It's a dynamic set of characteristics, each of which is a function of your genes, your learning, your environment etc. Essentially your personality is a collection of habits, many of which are so deeply embedded you don't recognise them as habits.

However humans - including you presumably although I've no evidence - possess the ability to change our habits. We're not Pavlov's dogs or Lorenz's ducks. In the time between an event and our reaction to it, comes an instant of decision, a moment of choice that we can use to control ourselves and our next actions - difficult as it might seem.

The brain is remarkably malleable.

 

[Yamato]

Well, I can't do it.

 

[Adamus]

NinjaTrader 7 beta has died on me

Not good. I can't import the tick data for EURGBP. NinjaTrader shuts down as soon as it's processed half a gig or so. Tried everything. Hmm, haven't downloaded the zip again, perhaps that's the issue. Although I figured it wouldn't unzip if it had been corrupted during the download.

 

[Adamus]

NinjaTrader can't cope with the EURGBP data and their support can't duplicate the exception. Totally error-free data and it makes NinjaTrader shut down. Another wasted couple of hours chasing someone's software bug in a dumb beta version. I'm like Travis, offer me a beta version and I can't say no. Sorry to use you as a bad example, Travis.

I knew it when I upgraded to the beta version that I shouldn't be doing it but I did it anyway.

Guess I can live without the EUR.GBP pair.

 

[Adamus]

Finally got them to reproduce the error at their end - they hadn't been applying all the settings I had used. I wonder when they're going to send me my free beta-tester beanie with the naff NinjaTrader logo or perhaps a T-Shirt with "I wasted 3 days and counting on NT7beta15 and counting! AHAHAHAHAHAHA but it was worth it because I'm a geek!"

 

[Adamus]

How to lose my mind........

How did I manage this? I overoptimised a sample with 50,000 bars and 3000 trades per instrument.

Attachment one: equity curve for optimisation window, 7 currency pairs from 1998 - 2009.

Attachment two: same strategy results, including walking forward 2009-2010.

I guess I need to go back to basics to figure this one out. OK, how many ****ing parameters did I optimise on? 4

One was the MA length of the filter which decides the direction to trade.

Two was the choice of algorithm to describe volatility (35 different types) (e.g. ATR, close-to-close, etc)

Three was the factor to multiply the volatility measure by to obtain the range from the close to place the limit entry orders

Four was the factor to multiply the volatility measure by to obtain the distance to place the first stop

Hey, maybe I can sell this :jester:

 

[Adamus]

Re: my very own Holy Grail

Here's one I can sell

I was minding my own business, laughing at the 'Holy Grail' threads on T2W, doing some more system/strategy development and then out of the blue, along comes my very own Holy Grail.

There is obviously something wrong with (a) my strategy, (b) my data, (c) NinjaTrader, or (d) my sanity, because this strategy has produced the holiest Holy Grail of trading systems that I have ever seen, apart from when I hacked TradeStation to buy every high and sell every low.

A fairly intensive inspection doesn't reveal anything unusual, so I figure I'll leave this until I've bought the NinjaTrader order execution license and then I can see what it does on the IB simulation account.

It's interesting to note though how feeble minded I am - I actually went so far as to walk-forward the strategy and am unable to convince myself that this is impossible.

Imagine these stats in real life:

[FONT="Courier New"][SIZE="2"]Total Net Profit	$273855.00
Gross Profit		$312595.00
Gross Loss			$-38740.00
Commission			$18360.00
Profit Factor		8.07
Cumulative Profit	$273855.00
Max. Drawdown		$-1990.00
Sharpe Ratio		1.34
	
Start Date		01/07/1998
End Date		30/06/2008
	
Total # of Trades	3672
Percent Profitable	81.84%
# of Winning Trades	3005
# of Losing Trades	667
	
Average Trade		$74.58
Average Winning Trade	$104.02
Average Losing Trade	$-58.08
Ratio avg. Win / avg. Loss	1.79
	
Max. conseq. Winners	60
Max. conseq. Losers		12
Largest Winning Trade	$1020.00
Largest Losing Trade	$-845.00
	
# of Trades per Day		1.01
Avg. Time in Market		82.1 min
Avg. Bars in Trade		1.0
Profit per Month		$2411.25
Max. Time to Recover	167.38 days
	
Average MAE			$63.11
Average MFE			$256.04
Average ETD			$181.46
[/SIZE][/FONT]

It's like finding a little green man at the bottom of the garden who just gave me three wishes. Must have been those mushrooms I had for lunch.

 

[pssonice]

poisoned mushrooms u had for lunch!

rofl

 

[Adamus]

They never mentioned this scenario in The Mathematics of Money Management

CRITICAL

Based on the above information, if you are backtesting in the Strategy Analyzer using order quantities of 100,000 and you now wish to trade live in your brokerage account where 1 lot is equal to 100,000 MAKE SURE that you adjust your strategy's order quantity from 100,000 to 1 to ensure you are trading the correct quantity.

Why is IB asking me if I want to buy 10,000,000,000 Euro? :jester:

IB probably put this warning up after receiving phone calls from people who had actually managed to place mammoth orders without realising it.

 

[Adamus]

3 weeks to go, still more nightmares with disktrading and ninjatrader

Managed to get Ninjatrader support to reproduce a bug in their historical data manager, so now I've got an issue tracking number and I'm waiting for the fix. in the meantime I can't import any more data, ninjatrader falls over at every attempt. Still I got 6 out of the 10 pairs that I want, all imported with 10 years of tickdata.

Still trying to find that elusive strategy that will optimise well and produce a decent walk-forward.

I've had to ditch 2 parameters from my strategy, so now i'm down to 2 parameters - a moving average length parameter, and a stop distance from entry parameter.

But I'm forward testing on 2009-2010 and it's like a brick wall and the little devil on my shoulder is telling me to go back to the Pardo style 2 year optimisation window, 6 month walk-forward, roll windows forward 6 months process.

Most the time when I listen to the little devil on my shoulder, something gets messed up and i end up going back to square 1. I guess I'll stick to the straight and narrow this morning and see how it goes by lunch time.

The 'straight and narrow' is optimising over 1998 - 2008 and walking forward 2009 - 2010. I figure any strategy that makes it through that walk-forward window has got to be worth trading. Like my Holy Grail system no no no no no i didn't mean that :-0

 

[Adamus]

Hmmm, totally inconclusive strategy.

Had great results in the optimisation window with a strategy that I spent most of the day tweaking.I actually ended up with not optimising it at all, and I just ran the final strategy against the walk-forward window to see if the results continued in the same way, and they were narrowly profitable.

So I'm left with the question still, is 2009 - 2010 just a really bad period to do walk-forward / out-of-sample testing on?

Is this strategy I have now worth optimising? Will optimising it improve or worsen the walk-forward? And if it makes it worse, shall I ditch the strategy or try improving the non-optimised version?

I think I need to be wary of putting in more rules, which are probably as bad as optimisable parameters when it comes to curve-fitting.

I changed the type of moving average filter, testing various different version.

I tried out several different entry mechanisms, eventually going back to the plain old enter-on-pullback with the limit order going 1 ATR from the Close.

I tinkered with the stop too, sticking with the parabolic stop which I really like, but adding another rule to get out when things look ugly rather than waiting for the parabolic stop to catch up.

Lots to do. I figure I should have a look at the monthly periods, which NinjaTrader shows nicely in a report, and compare the monthly period stats during the walk-forward period to see how out of line they are with the optimisation windows.

Basically the optimisation window captured $70 grand p.a., and the walk-forward captures $2500 only p.a. - breaking that down by month will be more revealing hopefully.