[BenOsborn]

I also hadn’t planned on trading a short system initially but implemented one last August.  It has returned 7.78% over the past 12 months with a drawdown of 3.39%.  Nothing to write home about in % terms overall but when combined with the other MOC systems backtesting shows it has a decent impact over time.  Plus I would be 7.78% further in the hole without it! haha

[BenOsborn]

August 2023 Performance

US
Combined MOC 3.26% (-61.57% DD, -58.09% since inception (late March 2022))

Breakdown:
MOC1 RUI      -0.27%
MOC2 RUI      -0.88%
MOC Deviate   0.09%
MOCCOD        4.73%
MOC Short     -0.36%

NDX Monthly Rotational -2.28% (-2.28% DD, 4.48% since inception Dec 2022)

[BenOsborn]

July 2023 Performance

US
Combined MOC -1.72% (-62.78% DD, -59.41% since inception (late March 2022))

Breakdown:
MOC1 RUI      -0.12%
MOC2 RUI      -1.09%
MOC Deviate   -0.63%
MOCCOD        -0.53%
MOC Short     0.73%

NDX Monthly Rotational 0.27% (6.92% since inception Dec 2022)

Started counting my MOC chickens before they hatched this month.  Was up nicely until 20 July when everything turned around!  You win some, you lose some. Bring on August.

[BenOsborn]

Hi Kate,

This is something that I have been monitoring but don’t have an answer to myself.

Over the last three months I have the following that fall into this type of category:

– Spike down at open (would never realistically get a fill as maybe 100 shares traded at some level below my limit order) – 3 missed trades, -1.47%

– No fill due to low volume – 4 missed trades, -1.16%

– Slippage – 32 trades, -0.42%.

[BenOsborn]

June 2023 Performance

US
Combined MOC 2.49% (-62.13% DD, -58.7% since inception (late March 2022))

Breakdown:
MOC1 RUI      0.60%
MOC2 RUI      0.08%
MOC Deviate   0.75%
MOCCOD        0.87%
MOC Short     0.27%

NDX Monthly Rotational 3.90% (6.63% since inception Dec 2022)

[BenOsborn]

May 2023 Performance

US
Combined MOC -0.76% (-63.06% DD, -59.7% since inception (late March 2022))

Breakdown:
MOC1 RUI      -0.63%
MOC2 RUI     -1.21%
MOC Deviate   0.73%
MOCCOD        0.97%
MOCMR         -0.11%
MOC Short     0.62%

NDX Monthly Rotational 11.53% (2.61% since inception)

[BenOsborn]

April 2023 Performance

US
Combined MOC -4.92% (-62.77% DD, -59.4% since inception (late March 2022))

Funnily enough I was actually very slightly positive for the month until this last week.

The work to reduce selection bias continued into the first week or two of the month.  Some systems have been swapped out for others to get the best mix of systems working together with the least selection bias.  Nothing wrong with the systems alone, the final mix of systems just perform better when combined (in theory anyway!).

Breakdown:
MOC1 RUI      -1.67%
MOC2 RUI     -0.71%
MOC Deviate   0.12%
MOCCOD       -0.38%
MOCMR         -0.89%
MOC Short     -0.56%

NDX Monthly Rotational -3.88% (-8% since inception)

I have noticed that I had made an error in my monthly calculations which overstated the drawdown.

Previous results should be:
Dec 2022                 -6%         
Jan 2022                  +3.22%
Feb 2022                 -1.12%
March 2022             -0.2%

[BenOsborn]

Sorry, every time I try and delete it just reposts it.

[BenOsborn]

Probably stating the obvious, but I always find it interesting looking at the impact different ranking formulas have.  The 10 below (ran one twice) show the vast differences they can have on a system:
  

[BenOsborn]

 

[BenOsborn]

Thanks Glen, that’s pretty handy. I hadn’t played with Scan before.

[BenOsborn]

Noting that my results don’t match Nick’s (except for the benchmark ), this is how I coded it:

Data:
    Liquid:    C > 20 and Avg(C*V, 50) >= 300000
    ATR5:    ATR(5)
    Sdev50:    StdDev(ATR5, 50)
    Sdev200:    StdDev(ATR5, 200)
    SdevCalc:    Sdev50/Sdev200
    Uptrend:         C > MA(C,200)
    Longset:    Uptrend and Liquid and SdevCalc < 5    
    Ranking:    100 – ROC(c,3)
LimPrice:    L – 0.5 * ATR5

CAGR 56% DD 52%.  Max DD was only in mid 20s until 2022 when it got creamed.

I don’t use scan but changed the rules to SdevCalc > 5 and got no trades, so this rule isn’t having an effect for me either.

[BenOsborn]

March 2023 Performance

US
Combined MOC -12.5% (-61% DD, -57% since inception (late March 2022))

Breakdown:
MOC1 RUI      -2.95%
MOC1 RUT     -2.63%
MOC2 RUI     -2.23%
MOC2 RUT     -2.31%
MOCV             -4.54%
MOC Deviate   -0.02%
MOC Short     1.28%

NDX Monthly Rotational -7.57% (-20.04% since inception)

Thanks to Nick who took a look over my MOC systems.  Although I had reduced selection bias compared to how my system was running earlier last year, it was still likely having more of an impact than I estimated.  I have since undertaken a significant amount of further testing, swapped some systems out with others and reduced orders, resulting in a system with selection bias on 0.42% of days.  RealTest makes a huge difference running so many tests easily, would have taken me forever in Amibroker with all the changes.

[BenOsborn]

Wow Terry, that is impressive!

[BenOsborn]

I agree with Terry, it does not take much time to run the day trade systems when you have everything set up. Takes me about 10 minutes now if I do it quick to load trades for multiple systems. Those using Interactive Brokers can probably do it quicker as I think I have a few more hoops to jump through with TradeStation.

It also takes a few minutes each morning to quickly check that live trading matches the backtest and investigate any differences. I used to keep a more detailed spreadsheet re any differences, and this would take longer, but once I was comfortable with everything I stopped doing that.

Not advice, but you there is a good chance that you may be considered a trader with your rotational and swing systems anyway so I wouldn’t let that influence any decision. I’m sure your accountant will let you know anyway.

[Author]

Posts