Triple Moving Average Experience

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Triple Moving Average Experience

Post by Ghost11365 » Sun Aug 18, 2013 11:09 am

Hi,

Does anyone have any successful experience using TMA that would be willing to chat?

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Mon Aug 19, 2013 8:43 am

yes i do
although i disclaimer that with the fact that although having made a decent amount fo money in the real World using the TMA that - like many medium term systems- that particular TMA is in a hugely deep draw down - yes thats a huge concern I know
C

Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Post by Ghost11365 » Mon Aug 19, 2013 2:55 pm

Hi C,

thanks for responding. is your strategy geared towards a more longer or shorter term trend? do you implement your TMA strategy over multiple markets (ex. equites & forex)? do you use TB to generate your orders for next day trades? curious to hear from you as I don't know others using this strategy and agree with you that it works knowing full well there are going to be long/steep drawdowns. I think we all know this is expected of the TMA strategy but I think the idea of cutting your loses right away and letting your winners run is a logical process and that sticking with your strategy though the drawdowns is vital. thanks.

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Wed Aug 21, 2013 8:01 am

i run it on 100 markets
next day moo orders

problems are that what pisses me off about a lot of systems is that if use xyz as teh 3 moving averages you may blow up
if you use x, y and z+10 you maybe teh best fund manager in history

Its a fine line
Maybe not that robust ?

C

sluggo
Roundtable Knight
Roundtable Knight
Posts: 2986
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo » Wed Aug 21, 2013 10:52 am

Possibly helpful: viewtopic.php?t=8606

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Wed Aug 21, 2013 11:54 am

Thank you Sluggo

I feel a re-ignition of the old days coming on - good posts on trading blox forum and lots of good input

Here's a question for you then [ all]

You test a system [ TMA] over 20 years of sample data = excellent
you test in on 3 in years out of sample data = excellent
You live trade it for 5 years and compund 20% in it per anum

year 6/ 7.5 draw down = 85 % [ max in backtested and real history 30%]

Good system ? Bad system ? What the hell happened ?

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Wed Aug 21, 2013 11:59 am

my answer incidentally = QE

Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Post by Ghost11365 » Wed Aug 21, 2013 12:49 pm

Thanks Sluggo and Chris for your replies. I'll checkout the link you provided and get back with feedback.

Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Help reviewing trades

Post by Ghost11365 » Mon Sep 02, 2013 9:03 pm

I used the TMA strategy, had equity of $250,000, risking 1% per trade. I ran a simulation from 2006-08-01 to current date (actually whenever the sample equities data ends). Nevertheless, when I review the trades for the period there is a transaction for a quantity of 1996 with an order fill of 21.27 = ~$42.5K; this is not 1% of my portfolio that should be at risk - it's closer to 17% - why the disparity?

Also i'm using the Turtle Edition - why are the cut and copy icons on my toolbar grayed out?

fab1usa1
Roundtable Knight
Roundtable Knight
Posts: 383
Joined: Sat May 21, 2011 8:30 pm
Location: New York

Post by fab1usa1 » Mon Sep 02, 2013 9:29 pm

Dollar Risk Per Trade for a long position is (Entry Fill - Stop Price) * Quantity.

Rearrange terms to get Stop Price = Entry Fill - Dollar Risk Per Trade / Quantity.

In your case: Entry Fill = $21.27, Dollar Risk Per Trade = 1% of $250K = $2500, and Quantity = 1996

Therefore: Stop Price = $20.02

The system assumes that you will close your position if the price drops below $20.02. In this case your dollar risk would be limited to $2500.

The Triple Moving Average system provides two parameters to calculate the Stop Price: ATR Average (days) and Stop (ATR). You can use whatever you like but I use ATR Average (days) = 20 and Stop (ATR) = 3. After the system calculates the Stop Price it then runs the formula to calculate the Quantity.

In this particular example that you gave I can almost guarantee you that the stock is exhibiting very little price movement hence the large quantity.

Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Post by Ghost11365 » Tue Sep 03, 2013 11:00 pm

Fab,

Thank you very much for your analysis - it makes perfect sense and I agree with your thoughts. I was under the impression that the 1% is the amount you're willing to lose on a false identification of a trend (for lack of a better phase). Now I need to figure out how automate a defined amount of my portfolio I want to risk per trade - i'd like to risk only $2500/trade (1% of my cash in account).

Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Post by Ghost11365 » Sun Oct 13, 2013 9:05 pm

What is the formula to calculate quantity? I'm using TMA, 1% risk per trade, ATR Stop, ATR Average = 20, Stop ATR = 1; Equty = $250,000 THANK YOU.

rhc
Roundtable Knight
Roundtable Knight
Posts: 464
Joined: Tue Nov 18, 2008 8:46 pm
Location: Oz

Post by rhc » Mon Oct 14, 2013 12:01 am

If using ATR stops for position sizing;

Position Risk = (Todays close) minus (Todays stop price) . . . expressed in Dollars, Euro’s, Yen, Rai etc

Stop price = (Todays close) minus (ATRmultiplier)*ATR

Therefore, Position risk = (ATRmultiplier)*ATR

Position size = Max risk($)/Position Risk($)

Example;
Todays close in Cotton = 55.55
Todays 20 period ATR = 1.55
ATRMultiplier = 1

Position Risk = 1*1.55 = 1.55 points (BPV for Cotton = $500)
Position Risk($) = $775.00

Max risk($) = 1% of $250000 = $2500

So, position size = $2500/$775 = 3.23 lots . . . rounding down = 3 lots

Does this help?

See here for more information on volatility position sizing
http://www.investopedia.com/articles/op ... -trade.asp

Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Post by Ghost11365 » Mon Oct 14, 2013 10:32 pm

Thank you for the reply. Your example is for futures - I see you multiplied 1.55 by 500; what would you do when using stocks versus futures?

rhc
Roundtable Knight
Roundtable Knight
Posts: 464
Joined: Tue Nov 18, 2008 8:46 pm
Location: Oz

Post by rhc » Tue Oct 15, 2013 7:37 am

The '500' figure simply converts the point value of Cotton to a dollar value since the value of 1 point in Cotton is $500

For stocks there are no conversions necessary since stocks are already priced in dollars.

Example;
Todays close in ABC corp. = $55.55
Todays 20 period ATR = $1.55
ATRMultiplier = 1
Position Risk($) = 1*$1.55 = $1.55
Max risk($) = 1% of $250000 = $2500

So, position size = $2500/$1.55 = 1612.9 shares . . . say 1600 shares

See example No.2 in link below for more information on volatility position sizing for stocks
http://www.trading-plan.com/money_position_sizing.html

Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Post by Ghost11365 » Tue Oct 15, 2013 9:54 am

Brilliant. Thank you very much.

doubleR
Full Member
Full Member
Posts: 15
Joined: Thu Jun 12, 2008 3:25 pm
Location: London

Post by doubleR » Tue Oct 22, 2013 3:16 pm

because markets have different speed to the upside vs downside when considering three moving averages

A = sma 10 days, B = sma 20 days, C = sma 100 days (short, medium and long term for example)

buy when A>B always (> means cross over)

sell when A < B IF B below C (< means cross below) but...

if B has moved above C then ignore A/B signals and sell only when B < C

for shorting reverse conditions.

it is a different way to use them. IF the trend is good enough we can shift "gear" to a longer trend following model. If we make a mistake, it should be limited at the early stage of a failure of a trend. Otherwise, if successful, we have been able to catch a trend much earlier than merely using averages B and C.

Thoughts?

Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Post by Ghost11365 » Tue Oct 22, 2013 6:30 pm

Hey thanks for replying. I see where you are going with your thought but I was thinking something a little different. I'm already using a long term trend following strategy. Assuming today is day one of my brokerage account, if I click generate orders I'm presented with potential positions to enter. However, typically all these trends have been in existence for quite some time. What I would prefer is a trend that has just recently begun. For instance if SMA crossed MMA and MMA > LMA then enter long IF SMA crossed MMA today (for example or maybe like within past week). I'm not sure if this can be done with my TB version - I have the turtle edition. Not sure if this would require writing code or even if using upgraded version (i.e. TB Builder). Please let me know any thoughts or suggestions. Thank you.

Ghost11365
Roundtable Knight
Roundtable Knight
Posts: 112
Joined: Wed Jul 31, 2013 3:33 pm

Post by Ghost11365 » Sun Nov 10, 2013 11:14 pm

I'm using starting equity of $250,000, Order Risk Per Trade (%) = 1%. How does the $2,500 really become my "risk" per trade? The amount i'd lose based on the stop loss order generation for my first trade (below) is $3233.01 ([Entry Price - Stop Loss] * quantity). That $2,500 isn't really the risk to me - it's used to determine position sizing which doesn't have any barring on how much could be lost on each trade (risk in dollars). Any clarification would be greatly appreciated.

Additionally, the sum of the 2 trades = $226,609.90 ([3333 * 46.40] + [626 * 114.95]); why isn't the remaining $23,390.10 invested in another position?
Attachments
Trading Blox Orders for 2013-11-11.htm
(5.88 KiB) Downloaded 266 times

Jake Carriker
Site Admin
Site Admin
Posts: 1493
Joined: Fri Sep 12, 2003 10:32 am
Location: Austin, Texas

Post by Jake Carriker » Mon Nov 11, 2013 10:01 am

The sizing logic uses the difference in the Order Price and the exit stop to size the trade. The order report you refer to shows the Fill Price, which was 22 cents worse than the order price in the case of the Agilent trade, thus the extra $700+ of risk. You can't know the fill before you size the trade, and the slippage incurred on entry (including opening gaps) adds to trade risk.
Additionally, the sum of the 2 trades = $226,609.90 ([3333 * 46.40] + [626 * 114.95]); why isn't the remaining $23,390.10 invested in another position?
Since one trade wants over $150K in capital, and the other wants over $70K in capital, it might be the case that $23K is not enough cash to take another position. Check the Filtered Trades log to see the reason any particular trade is rejected.

Post Reply