Tradestation 8.1 (build 3066)

[JTrader]

Hi

I gained access to TS8.1 on Friday, after opening a brokerage account. I've been busy familiarising myself with the platform and all its features, by reading the help files and playing with it. I like what I see, and it all operates in a logical way.
My only criticism so far is that it's not possible to view strategy performance reports in terms of points and pips profit, instead they can only be viewed in $ profit and loss terms.

Now that I'm familiar with all the features, the next step is to develop some robust (and profitable) mechanical intraday strategies.

Cheers

jtrader.

 

[fxmarkets]

you could always adjust the dollar value so that it would simulate points, used to do the same myself. saves working out $1254.00 / $37.50 etc. whatever,l like you say niggle, but can be viewed that way, by adjusting.

lazy buggers

 

[JTrader]

Thanks fxmarkets

how exactly is this done? ..........By selecting a "Dollars per trade" size of $1 per trade within "format strategies" properties, general, "Trade size (if not specified by strategy)"?

Thanks

jtrader.

 

[fxmarkets]

i used to alter the tic details in the pricing universe setup.

 

[TWI]

This should be the least of your worries with Tradestation believe me.
When you start down the automation route you will be making some irrate phone calls transatlantic to morons who are less helpful than a cowpat and more moronic than tube station barrier attendants.
Good luck. I am already on the journey and still getting there.

 

[JTrader]

Thanks for the warning twalker , but I do not plan on fully automating anything, I am happy to manually execute the trade signals generated. At the very least, this would mean that I have something to do!

The tradestation 8 futures account application form makes you acknowledge that mistakes can and will occur when automating strategies on the platform, and that automating a trading strategy does not mean that you should leave the automated strategy running while away from your workstation. So I'll just have to forget about playing golf while my strategies do all the work for me!

Setting the strategy properties to return performance report results in single $ values that are equivalent to single points or pips profit, was as I suggested in my previous query. Here is a screenshot (please ignore my commission and slippage setting! -

 

[turtle trader]

Hi James

I've never tried to switch to number of points/pips instead of $ values, but suspect the way to do it is to go:
Format Symbol, then hit the 'Properties' tab, then adjust the 'Price Scale' setting.

I remember thinking that I'd like to see point/pip totals when I first switched to TS 8, but you soon get used to the $ amounts, after all, that's what you're trying to make! I suggest you leave it as it is for a while & see if you don't soon prefer it.

Simon

 

[JTrader]

Hi Simon

I know what you mean....I just like to keep things simple, in the sense that I can see how many points or pips profit or loss a strategy makes. This way, i do not need to worry about the other setting. Maybe I'll adapt in time as I learn more!

One drawback of my approach of wanting to $ to points and pips seems to be that because all the strategies seem to work around $ values for things like stop loss' and profit targets, as opposed to points and pips, I'll also need to convert these values. Perhaps it will be as easy to learn how to do it the TS $ way!

it doesn't seem possible to edit the price scale box contents, as the 1/10000, 1/100 etc. inputs, seem fixed, in order to reflect the particular contract in question (?) as with the "Min. move" and "big point value".

 

[JTrader]

I'm not too keen on the "we are experiencing high call volumes with hold times exceeding 30 minutes" when I try calling the platform support desk (albeit at peak times), but then again if I am to be placed on hold, its good that i know for how long!
Lets hope the trading desks (which I will not be trading through) enable quicker access!

 

[turtle trader]

One drawback of my approach of wanting to $ to points and pips seems to be that because all the strategies seem to work around $ values for things like stop loss' and profit targets, as opposed to points and pips, I'll also need to convert these values. Perhaps it will be as easy to learn how to do it the TS $ way!

You can set exits as a number of points by using the reserved word 'points'! Look at the TS User Guide - the example it gives is:


An exit statement can be written to prevent large losses by:

Sell This Bar at EntryPrice - 1 Point Stop;

where 'point' represents the smallest possible price movement of the security.

 

[JTrader]

Thanks Simon.

As yet I've only played about with combinations of the built in strategies. I am going to have to modify some at the very least because most of the existing exit strategies do not enable intrabar order generation. Therefore I'm going to have to learn some easy language.

I was wondering why my backtest results were returning strange outputs! I've been testing EURUSD for example with the stop loss LX strategy in place. EURUSD minimum movement is $10, this equals 1 pip. I'd got the stop loss set with an input value of 1 (=$1 loss = 0.1 pip stop loss) and profit target LX set with an input value of 5 (=$5 target = 0.5 pip profit target).

I can now produce better performance reports, and make more sense of the generated results!

 

[JTrader]

Am I going down the road of over optimisation and curve fitting?

Hi

the strategy that I've been playing about with consists for 4 seperate built in TS8.1 strategies.
It has an RSI entry point, stop loss & % trailing stop-loss.

I've optimised the four components in turn, and each one improved the total net profit.

Therefore I repeated this cycle, and with each individual optimisation the net profit increased.

Here is a summary, based on $10,000 initial capital -

Total net profit ($) Total number of trades
9744 951
11890 950
12044 972
12044 972
12407 1271
12702 1266
13505 1135
13999 1007

It seems that I could go on forever repeating this cycle. But this sounds like curve fitting or over-optimisation.

But is it?

The total number of trades has grown. I would have guessed that the total number of trades would fall on a curve-fitted system, as entry and exit points become more specific?

If this is over optimisation, where should I stop? I've already completed 2 cycles of 4 optimisations. Is 2 cycles too many, or would 10 cycles = too many?

Theres no definate rules I'm guessing...

My results are based on EURUSD, with 1 years worth of 10-minute data.

All input welcome

Thanks a lot

jtrader.

 

[ZDO]

jtrader,

Short answer about your optimization runs:
Find all the peaks.
Hopefully you will find two clear ones.
Then select a setting between the two and just below one of the peaks.
and don't even dare ask me why

btw: use the Forum for questions. Compose them carefully. Don't waste your time on the phones. You'll stay on hold - only to get poor support and a brush off ...

Welcome to TS hell !!!

zdo

 

[turtle trader]

To check for over optimisation try testing some different time periods - if you've tested the last 12 months then try testing the previous 12 months & the 12 before that. I suspect you'll find the results disappointing, i'm afraid, but you never know?

You also need to look at your average profit/trade. Your final optimisation gives an average of just under $14. Even if you've already allowed for slippage you're never going to be able to achieve that in real trading.

Good luck
Simon

Hi

the strategy that I've been playing about with consists for 4 seperate built in TS8.1 strategies.
It has an RSI entry point, stop loss & % trailing stop-loss.

I've optimised the four components in turn, and each one improved the total net profit.

Therefore I repeated this cycle, and with each individual optimisation the net profit increased.

Here is a summary, based on $10,000 initial capital -

Total net profit ($) Total number of trades
9744 951
11890 950
12044 972
12044 972
12407 1271
12702 1266
13505 1135
13999 1007

It seems that I could go on forever repeating this cycle. But this sounds like curve fitting or over-optimisation.

But is it?

The total number of trades has grown. I would have guessed that the total number of trades would fall on a curve-fitted system, as entry and exit points become more specific?

If this is over optimisation, where should I stop? I've already completed 2 cycles of 4 optimisations. Is 2 cycles too many, or would 10 cycles = too many?

Theres no definate rules I'm guessing...

My results are based on EURUSD, with 1 years worth of 10-minute data.

All input welcome

Thanks a lot

jtrader.

 

[JTrader]

turtletrader - You also need to look at your average profit/trade. Your final optimisation gives an average of just under $14. Even if you've already allowed for slippage you're never going to be able to achieve that in real trading.

I did account for slippage, or the spread at least in these results on the basis that EURUSD has a 3 pip spread. So i set slippage at 1.5 on the basis that TS8.1 will subtract 1.5 on the trade entry, and then subtract a further 1.5 on trade exit (?).

Without downloading data beyond the last 1 year, the profit split between the first and last 6 months of this year is more or less even at least.

The bit that confuses me is about the slippage setting is - based on one EURUSD cash contract being worth 100,000 ($10 per pip) with a typical spread of 3 pips, and with the chart plotting the bid prices. In order to cover just the basic spread only, should I set the slippage setting at $1.5 per trade or $15 per trade?...............I'm now thinking that slippage should be set at $15 per trade, which makes a big difference to the performance report!

Does the $13999 profit = 1399 pips * the $10 value per pip within 100,000 (1 lot) EURUSD cash contract ? (This would mean that the average net profit per trade was $13.74 or 1.374 pips.)

Thanks again

jtrader.

ps. I suppose I should be getting used to posting these questions on TSW. But so far, my browsers haven't been very TSW friendly!

 

[ZDO]

jtrader and other TS users,

This is way off topic but
Go download the Motherloads before it's too late.

zdo

 

[JTrader]

jtrader - did account for slippage, or the spread at least in these results on the basis that EURUSD has a 3 pip spread. So i set slippage at 1.5 on the basis that TS8.1 will subtract 1.5 on the trade entry, and then subtract a further 1.5 on trade exit (?).

Without downloading data beyond the last 1 year, the profit split between the first and last 6 months of this year is more or less even at least.

The bit that confuses me is about the slippage setting is - based on one EURUSD cash contract being worth 100,000 ($10 per pip) with a typical spread of 3 pips, and with the chart plotting the bid prices. In order to cover just the basic spread only, should I set the slippage setting at $1.5 per trade or $15 per trade?...............I'm now thinking that slippage should be set at $15 per trade, which makes a big difference to the performance report!

Does the $13999 profit = 1399 pips * the $10 value per pip within 100,000 (1 lot) EURUSD cash contract ? (This would mean that the average net profit per trade was $13.74 or 1.374 pips.)

Thanks for the tip ZDO!

I fully understand what the above settings should be now. A 10,000 forex position (mini-lot) would equate to needing to take account of $3 of slippage (MINIMUM) for the 3 pip spread ($1.5 per sde/trade) Whereas a 100,000 forex position (a lot) equates to needing to take account of $30 of slippage ($15 per side/trade) (MINIMUM) for the 3 pip spread.

It would appear, that optimisations do have a habit of levelling out and improving no further, with repetition. Applied to my EURUSD chart for example, is 4 individual TS8.1 inbuilt strategies, I optimise each one in turn. During the first cycle, the optimisation of each strategy improvres the net trading profits significantly. The second cycle of optimisation might improve the results a bit more, until you reach a point when each seperate optimisation isn't improving the results from the last optimisations any further.

It's amazing how a stategy (or combination of strategies) can differ from chart to chart. i.e. optimise parameters on 30-min chart, place on 10-minute chart, optimise parameters again, and both the total net profit and the total number of trades can be vastly different.....

(You'll have to excuse me for thinking out loud on this thread :!: )

Cheers.

 

[fxmarkets]

Its quite creative therapy to think and consider a trading strat, I use to enjoy it, however huge chunks of your life seem to go un accounted for, when someone says what you been upto this last year or 2 . errr, umm. crikey ,sunlight....

enjoy, watch out for blurry eyes and dont forget you need to eat now and then...

 

[JTrader]

Thanks fxmarkets

I notice you speak about your TS strategy development in the past tense .
Where did all that strategy developing hard work get you? has it been a happy outcome?
do you still use TS? If so, what do you use TS for - trading mechanical strategies with trade entry and exit signals, semi mechanical trading with a mechanical exit and discretionary exit (or vise versa) or do you/did you just use TS as an advanced charting application?

I would be grateful if you could please sum up your experience of using TS! Has the outcome made it worthwhile, or, was it a waste of time for you?

Cheers

jtrader.

 

[fxmarkets]

hmm this was going back quite a few years, and at that time I was busy being "normal" juggling, then ended up getting married, kids, moving, renovations, time was needed or directed else were... distraction from trading is a good way to put it, experience with systems development, not a waste of time, but how I look at trading now, and how I trade its simply not needed,for me, that said I know I will at some stage would like to see if I could programme some elements that I know now into the beast.... but you need to dedicate a fair run of time months, getting into these idea's.

Not sure if it can be done though..... heavy discretion reading of the market, thats what I mean by just not needed. can TS cope with false breaks, eyeing targets, spotting trends that have turned but the price doesn't look nothing like the such....manipulation, on and on...

the thing is price leads most indicators, so they are late, i'm well in before they would show up. I know you code on non indicators too, but if the markets not going down any more in whatever cycle its in, i'm looking to get long, simple as that really. ditto shorting it too.

I can see me using it to maybe produce or flag certain conditions, but would I loose contact and not feel tic by tic , become lazy waiting for an alert....i'd just be thinking, come on then you should alert me now.... etc...


so its just, 1 market, simple charts, total focus.