• Australian (ASX) Stock Market Forum

Hello and welcome to Aussie Stock Forums!

To gain full access you must register. Registration is free and takes only a few seconds to complete.

Already a member? Log in here.

XAO backtest benchmarks for 2018-2020

Discussion in 'Trading Diaries and Journals' started by qldfrog, Feb 17, 2020.

  1. qldfrog

    qldfrog

    Posts:
    4,209
    Likes Received:
    2,685
    Joined:
    Jun 8, 2008
    Dear all system builders, I toyed with the idea of having a repository of backtest or even paper/real trading results for various recent periods, to use as a rating/grading tool when developing systems of your own; what to beat, what to aim for or reaching for the sky;
    Ideally if backtesting, please ensure you are do not have future leaks..however great the results are :)
    And we all know that backtesting will nearly always be different from real trading action so no point discussing this here
    The rules:
    Portfolio size $100k
    Positions 20
    Realm: All Ordinaries (XAO)
    While my systems are weekly, dayly or monthly system results are welcome

    The periods we will test are recent enough to mitigate the survivor bias and chosen to represent interesting market trends
    01/07/18 to 01/01/19 : a 6 month down period,
    01/07/19 to 1/1/20: a 6 month unstable period,
    1/1/19 to 1/1/20 a full calendar year with great growth momentum overall
     
    Roller_1 and Skate like this.
  2. qldfrog

    qldfrog

    Posts:
    4,209
    Likes Received:
    2,685
    Joined:
    Jun 8, 2008
    upload_2020-2-17_18-32-32.png
    excel details attached
    Hope it helps

    systems 1 has been traded for a year next week with real world trading matching these back-tests but for a few mistakes from the operator,
    system 2 started in August 2019
    thanks to @Skate for these 2 systems initial ideas

    123 system is under development and based on the 123 or ABC(D) indicator as discussed by @peter2 Many thanks

    Still trying to see where my zigzag is wrong..too good to be true so paper trading to detect future leaks if any;
    Thanks to @rnr !!!
    Looking forward to work on the super-trend next (@Skate @peter2)

    I welcome all similar entries :) @Warr87?
     

    Attached Files:

    Warr87 and Skate like this.
  3. Newt

    Newt

    Posts:
    453
    Likes Received:
    603
    Joined:
    Oct 17, 2012
    Might be worth specifying standard commissions per trade and slippage for people to apply if sharing qldfrog. For example, I'm trading with CMC with max $11 commissions, and apply 1.5% slippage for buy and sell. Its surprising how even these mild but hopefully realistic handicaps can change a smooth equity curve to more realistic obstacle course :)

    I guess while on this line of thought, delays should be specified. Will vary depending on strategy and time constraints, but I use 1 day delay on buy and sell. If compounding returns (matters quite a bit after a few years), how many positions are targetted.

    In Amibroker AFL:

    //Slippage modelling:
    BuyPrice = Open*1.015;
    SellPrice = Close*0.985;

    //Commisions for brokerage:
    SetOption("CommissionMode",2);
    SetOption("CommissionAmount",11);

    //Trade delays:
    SetTradeDelays( 1, 1, 1, 1 );

    //Position Sizing and compounding returns:
    Positions = 15; //(or whatever)
    SetOption( "InitialEquity", 100000 );
    SetOption( "MaxOpenPositions", Positions );
    SetPositionSize( 100/Positions, spsPercentOfEquity );
     
    qldfrog, Skate and Nina4 like this.
  4. qldfrog

    qldfrog

    Posts:
    4,209
    Likes Received:
    2,685
    Joined:
    Jun 8, 2008
    Indeed, my results were performed using a weekly system, 1 day delay, no slippage and $10 commission per buy/sell
     
    Newt likes this.
  5. Skate

    Skate

    Posts:
    1,638
    Likes Received:
    7,227
    Joined:
    Dec 28, 2013
    3 Actual trading strategies

    1. The CAM Strategy with StaleStop
    2. The HYBRID Strategy
    3. The PANDA Strategy

    Backtest Settings - 6 month down period
    Start Date: 1st July 2018
    End Date: 1st January 2019
    Period: 6 months
    Portfolio Capital: $100,000
    Positions in the Portfolio: 20
    Fixed Position Sizing: $5,000 (No re-balancing)

    1. The CAM Strategy with StaleStop

    2. CAM StaleStop Capture.PNG




    2. The HYBRID Strategy

    3. HYBRID Capture.PNG




    3. The PANDA Strategy

    4. PANDA Capture.PNG

    Skate.
     
  6. Skate

    Skate

    Posts:
    1,638
    Likes Received:
    7,227
    Joined:
    Dec 28, 2013
    3 Actual trading strategies

    1. The CAM Strategy with StaleStop
    2. The HYBRID Strategy
    3. The PANDA Strategy

    Backtest Settings - 6 month unstable period
    Start Date: 1st July 2019
    End Date: 1st January 2020
    Period: 6 months
    Portfolio Capital: $100,000
    Positions in the Portfolio: 20
    Fixed Position Sizing: $5,000 (No re-balancing)

    1. The CAM Strategy with StaleStop

    2. CAM StaleStop Capture.PNG




    2. The HYBRID Strategy

    3. HYBRID Capture.PNG




    3. The PANDA Strategy

    4. PANDA Capture.PNG

    Skate.
     
    qldfrog likes this.
  7. Skate

    Skate

    Posts:
    1,638
    Likes Received:
    7,227
    Joined:
    Dec 28, 2013
    3 Actual trading strategies

    1. The CAM Strategy with StaleStop
    2. The HYBRID Strategy
    3. The PANDA Strategy

    Backtest Settings - a full calendar year with great growth momentum overall
    Start Date: 1st January 2019
    End Date: 1st January 2020
    Period: 12 months
    Portfolio Capital: $100,000
    Positions in the Portfolio: 20
    Fixed Position Sizing: $5,000 (No re-balancing)

    1. The CAM Strategy with StaleStop

    2. CAM StaleStop Capture.PNG




    2. The HYBRID Strategy

    3. HYBRID Capture.PNG




    3. The PANDA Strategy

    4. PANDA Capture.PNG

    A full calendar year with great growth momentum overall

    The results above explains why I made the post below - (a small Portfolio size of $100k & $5,000 bets) multiply that 3 & 10 fold with re-balancing profits gives an indication what can be achieved. Bear in mind, when I win, I win big & when I lose I lose big, it's all relative.
    Update - of my Hybrid Strategy
    The updated Equity Curve of my Hybrid Strategy is for the last 12 months. The chart confirms trading with the trend still works.

    Equity Capture.PNG

    Skate.
     
    Newt, debtfree and Roller_1 like this.
  8. Roller_1

    Roller_1

    Posts:
    278
    Likes Received:
    54
    Joined:
    Sep 13, 2013
    Sorry to deviate a bit but why don't you participate in the Opening auction to mitigate your slippage Newt? I just set a LMT order the night before (to sell) 5-10% below the closing price, then you get the opening price, i don't have any slippage issues.
     
    MovingAverage and qldfrog like this.
  9. qldfrog

    qldfrog

    Posts:
    4,209
    Likes Received:
    2,685
    Joined:
    Jun 8, 2008
    @Roller_1 and if it helps @Newt
    I do the same: ask abuy higher than previous close and a sell lower than previous close.
    In most cases, it get acted on at the Monday open and I do not have any slippage
    BUT
    Bell Direct sometimes refuses my buy order as they are too high above the estimated open price and so might "tweak" the market.I asked the reason and it is a manually operator controlled action,
    If I am deamed too high my order is refused and cancelled: I have had time when I wanted to by at X$ a share (around 2/3% above previous close) my order is refused at 9:30 Sydney time as too high and it actually opens around my X$ value 30 min later and I chase the price manually at X$+ to pass my order

    so not an universal recipe,
    I also have the case where I need to sell before buying as BD requires the money on the account before allowing the purchase ;
    so once fully invested if I want to sell/buy lots I need to first sell then buy and so only the sells are slippage free, I have slippage on the buys

    In conclusion, in the real world i have some slippage..sometimes so the slight divergence between backtests and real results
     
    Last edited: Feb 18, 2020
    Newt and Skate like this.
  10. Warr87

    Warr87

    Posts:
    204
    Likes Received:
    383
    Joined:
    Jun 28, 2019
    You've called me out so I guess I have too now ;), haha.

    I will try and post something on the weekend.
     
    qldfrog likes this.
  11. qldfrog

    qldfrog

    Posts:
    4,209
    Likes Received:
    2,685
    Joined:
    Jun 8, 2008
    @Skate much thanks, your data is much appreciated
    giving targets which are achievable and emulating to work further and refine the systems further
     
    Skate likes this.
  12. qldfrog

    qldfrog

    Posts:
    4,209
    Likes Received:
    2,685
    Joined:
    Jun 8, 2008
    That was the idea but only if you want to,
    :) one of my issue is the isolation: is that 20% gain on this period a great result or pathetic? how can I know if I have no one to compare to? I can not expect stellar results but where should I put my targets
    Hopefully these "benchmarks" covering various relatively recent periods will help not only me but others in the same situation
     
    Skate likes this.
  13. Warr87

    Warr87

    Posts:
    204
    Likes Received:
    383
    Joined:
    Jun 28, 2019
    You are right. It is hard to tell what would be considered good or pathetic, particularly when you start out designing a system.
     
    qldfrog likes this.
  14. MovingAverage

    MovingAverage Smoke me a kipper

    Posts:
    294
    Likes Received:
    403
    Joined:
    Jan 23, 2010
    That's a shame, I'm living trading a daily swing system but it uses a 20% position size so only max of 5 positions. Would like to post the results but doesn't meet your rules so I guess I'm out :(
     
    Skate and qldfrog like this.
  15. DaveDaGr8

    DaveDaGr8

    Posts:
    110
    Likes Received:
    34
    Joined:
    Jun 29, 2007
    Post it anyway .. Be a rebel !!!! .... I'm feeling like this forum needs a shakeup in MAE/MFE
     
    Skate, qldfrog and MovingAverage like this.
  16. MovingAverage

    MovingAverage Smoke me a kipper

    Posts:
    294
    Likes Received:
    403
    Joined:
    Jan 23, 2010
    that’s all the encouragement I need...I’m here to shake things up baby!!! Not sure why but I had George Constanza’s voice going on in my head when I was typing that out...hahaha. I’ll post my results tomorrow morning.
     
    Skate and qldfrog like this.
  17. Newt

    Newt

    Posts:
    453
    Likes Received:
    603
    Joined:
    Oct 17, 2012
    No probs roller. I put in Limit orders night before trading and generally get close to the Open too, but for aggressive stocks moving quickly you don't always. I guess in answer to your question, 2 reasons:

    1. Model the small number of trades where you don't get the Open - not a huge issue for weekly trend trading but your slippage is always >$0 if you average it over the year, so better to have something in your model than be too optimistic

    2. Assist with an idea of what price to offer for next day's trading - so frustrating when you miss out, but frustrates me even more if I offer 5% or greater premium and price later recovers


    Slippage modelling is admittedly much more important for short term trading (e.g. mean reversion).
    I would challenge you to including a bit of slippage in your backtests - even 1% can make a surprising change to your final equity curve (not just lower - generally much more realistic "lumpiness").
     
    Roller_1 and qldfrog like this.
  18. qldfrog

    qldfrog

    Posts:
    4,209
    Likes Received:
    2,685
    Joined:
    Jun 8, 2008
    Please post but highlight the daily factor, and whatever other specifics
    So brokerage, fixed position size,etc
    I am sure it will be of use.
    looking forward to seeing this
     
  19. Newt

    Newt

    Posts:
    453
    Likes Received:
    603
    Joined:
    Oct 17, 2012
    Here are 3 backtests on the weekly trend system I'm currently trading:
    All Ords universe, no slippage, $11 commissions, 20 positions, $100k starting, no compounding:

    1/7/18 - 1/1/19
    2020-02-18 22_02_51-Symbols.jpg
    1/7/19 - 1/1/20

    2020-02-18 22_03_21-Films & TV.jpg

    1/1/19 - 1/1/20
    2020-02-18 22_04_32-Films & TV.jpg
     
    qldfrog likes this.
  20. qldfrog

    qldfrog

    Posts:
    4,209
    Likes Received:
    2,685
    Joined:
    Jun 8, 2008
    Much appreciated @Newt
     
    Newt likes this.
Loading...

Share This Page