A Mechanical Strategy Journal

[nickm001 10]

correct... while the absolute values of the continuous contract may not match old contracts ( due to shifting up of all old contract values at rollover, when spliced to new contract), the difference between entry and exit ( P/L) should be the same.

[Dogpile 10]

nickm, could you comment on this:

 

I show the Russell contract in 2002 trading under 1400 contracts per day. (ie er2h06)... and under 10,000 contracts per day in 2003 (ie er2h03).... 2002 had good 'range' as the market was in freefall but 2003 & 2004 show about 1/2 the range (ATR) of that achieved by the completed Sep 2007 contract (15 pts vs 6-8 pts)... tick values are the same so todays contract is quite juicy compared to just 5 years ago. pretty amazing. in you opinion, is it valid to use strategy results from 2002 when the market dynamics (virtually non-existent volume and 'dollar value range') were so different?

 

seems to me like growth of futures contracts traded changes the futures market quite fundamentally. you simply couldn't have many traders operating with only 1400 contracts traded per day to work with.

[TinGull 10]

Ahhhh, I was using a 24 hour chart for my opening prices.

[nickm001 10]

Low volume will impact RT trading ( slippage), but as far as strategy testing, there is no impact.

Range should be considered as relative range. I usually "normalize" the range results by dividing range with value of the contract. For daytrading, I would use range as metric, not ATR. ATR is important only if you are holding overnight. In 2002 the peak range was in July ( ~16) and contract traded at ~400. This year peak range was 24, but contract is trading at 800 level, so the range actually has contracted.

 

For testing strategy, you need as many different market condition as possible.. Proper testing , optimization and walk forward testing ( the only real testing) is pretty complex process. If you have access to TS forum, there is tone of stuff on that subject.

[Dogpile 10]

New Order Triggered

 

Long 1 RUS 788.30

 

-----

Existing Positions:

Short 1 ES

Short 1 NQ

Long 1 ER2

 

(will try to follow this for all 3 contracts)

 

in looking at historicals --- the time this strategy gets nailed is when it is short into a relentless multi-day climb like 6-8 straight up days. even including those, the performance has been quite good -- but you have to be able to stomach some occassionally serious drawdowns.

[Dogpile 10]

New NQ Order Triggered

 

Long NQ 2027.00

 

-----------

Existing Positions:

 

Short ES

Long NQ

Long ER2

[Dogpile 10]

End of a wild day update.

 

We are buried in ES but this is offset by NQ net gain, and RUS is absolutely nuts.

 

How our account stands on paper at todays close:

ES -568.75

NQ +670 (net gain including 1 unprofitable short and the todays paper gain)

ER2 +2,080 (entered 788.30 today and it closed 813.80 - paper gain of +$2,550, less the -$470 loss on initial short)

 

Net Standing as head into tomorrow:

+2,181.25

--------------

Outstanding positions:

Short 1 ESZ07

Long 2 NQZ07

Long 1 ER2Z07

--------------

(note, since NQ is only $20 per point, assume using 2 NQ contracts vs 1 each for ES and ER2)...

[waveslider 10]

Dogpile,

I am getting different results with your strategy. What is the max bars the strategy will see?

[waveslider 10]

ok nevermind. I see you changed the code. I notice it works best when the market is not trending (majority of the time). That's a good way to operate if you can cut losses in the trends.

[Dogpile 10]

I am using the minimum -- 6 bars -- since it only needs to reference the highest high and lowest low of last 5 bars. yah, I added the volume code just to only create signals on an active contract.

 

another way to cross-check it is do it is to confirm it using the ETFs since they don't rollover (SPY, QQQQ, IWM) and the futures are based on these ETF's as the underlying.

 

Range expansion off opening price seems to be very valid idea for the long side. The short-side is tougher -- it is profitable and adds to returns but you can take a ton of heat along the way -- this could be because in a multi-year bull market -- or have something to do with just market dynamics where market falls fast and rises in relentless fashion (take a lot of pain but rewarded occassionally in a big way on short side).

 

Breaking down the strategy results you can see the effect the short-side drawdowns have on the overall result. It is still profitable but painful at times.

[Dogpile 10]

<<That's a good way to operate if you can cut losses in the trends.>>

 

waveslide, I can see in the strategy orders that some of the orders fly right in the face of a few core principles I abide by. ie, tomorrow, it looks like we will very likely get short based on the strategy rules. I would not want to get short BELOW the opening price given the set-up going in. Above the open might be a short tomorrow -- but I do not want to short on weakness with all the likely residual momentum in the market. I have noticed this type of order --- in the face of a potential breakaway gap --- being a problem for this strategy. So already I can see that this short-strategy as it is written just isn't right for me. Perhaps I will only take the longs and rely on my own existing short-scalp strategies to make money on the short side.

 

The long-side strategy looks good. I actually went long RUS today and thought of keeping 1 on in line with this strategy. I wimped out and scratched the order -- big mistake.

[Dogpile 10]

Plugged formula into a 'ShowMe' function in Tradestation Radarscreen to automate the calculations and can post them after the opening price prints.

 

Today we will get 'sell-long and go-short' signals at the following prices:

 

NQZ07.D Short 4 Contracts on Sell-Stop @ 2066.75

ER2Z07.D Short 2 Contract on Sell-Stop @ 819.90

 

note:

ESZ07.D - Already short and no buy signal can trigger today

 

(since already long, the NQ and ER2 trades are actually 'Stop And Reverse' trades whereby you would short 4 NQ and 2 ER2 -- first to sell long, then to go short)

-------------

 

I really don't like these potential short orders -- not to say there isn't downside, there will likely be a good flush down today or tomorrow... -- its just that we have no prospect of covering these orders until we close low in the 5-day range -- which won't be for a while most likely --- seems like a tough trade but I will follow them here anyway. Meanshile, I will be working and thinking on some filters for the short-side code. Have a couple of ideas -- need to work out the code for them.

[waveslider 10]

FWIW, though there is a lot of emotion in the market right now, there is a good case for calling this market a ranging one. For one, the major down move in August retraced the entire up leg that Mar-July achieved.

This is the kind of thing the system doesn't see - it is stupid, it just does the same thing over and over again. Still it makes money! That's because it is often doing the contrarian action. Contrarian actions make money, but it takes balls to be a contrarian - and you have to have a good strategy.

[waveslider 10]

"The book uses a 9-day range for its triggers. I prefer a 5-day range."

 

Why 5? The nine looks smoother. Coincidentally, Gann and Demark seem attracted to the number nine (the beatles too).

[Dogpile 10]

why a 5?

 

good question. I feel like a 5-day buy system will catch those hard 1-2 day pullback sell-flushes that occur during an uptrend.

 

5 is a fib number. I played around with 8 too (another fib number).

 

9-days works well too. 5 works well. 8 works well. the concept is just plain powerful on the long side (range expansion off opening price).

 

my preference on looking at all of them was that following the 5-day lines up best with my existing strategies. My current strategies are all about 'going-with' the intraday trend but being more careful if it coincides with a 'daily trend'. I believe in choppy markets if looking at them on a 2-4 day basis... but those 6-8 day consecutive moves do happen so I have built-in rules so I don't get run-over when those do occur.

 

the strategy just needs the recent close to be closer to the bottom of the range than the top of the range -- so it fits in with my 3-5 day buying/selling cycle thesis. it doesn't have to make a 5-day low or high, it just has to be closer to that side.

 

I am just trialing the 5-day --- I have also thought about doing some combination like:

long 9-day ES

long 8-day NQ

long 5-day ER2

 

and leaving the short-selling to my own scalping -- I do extremely well on the short side as is...

 

your comments are greatly appreciated.

[Dogpile 10]

well, new short orders triggered (NQ & ER2) so short everything now... tough location IMO to short but I will see this through...

 

Existing:

Short 1 ESZ07.D

Short 2 NQZ07.D

Short 1 ER2Z07.D

[bh_trade 10]

The shorts triggered today always struggled. Too many short coverers and wannabe longs providing bidside support into the partial fill of today's opening gap.

 

Curious, are you paper trading these or actual $$?

[Dogpile 10]

I do not live trade this strategy yet... I am in study mode on this one. I am working out the mechanics of it while I also have to get used to the pain it often brings in the short-run. Mechanical trading is a lot mental -- as pointed out in Art Collins book. Good systems work but traders heads can screw them up. I may get started on a few shares of IWM/SPY/QQQQ just to have a little skin in the game and start to get used to the swings.

 

I am also working on enhancements. ie, I converted the code to intraday charts and going to see if I can improve results by using closing VWAP rather than closing price in the set-up. I also have idea around using 'weekly MACD histogram' as a filter to help screen out some short trades. finally, something calendar-based might add some value given the tendency of hard corrections to occur between May and September. ie, if weekly MACD histogram has not been below zero in 6+ months then turn short-sale strategy on and turn it off when weekly histogram turns up. Or, turn on short-sale strategy if haven't touched lower daily -2.5 keltner channel in 13 weeks and turn it off 8 weeks later.... something like that.

[Dogpile 10]

Update:

 

Yesterdays short entries looking better as of today -- but a long way to go before a long trigger is possible. Close will have to be in bottom 1/2 of 5-day range to potentially get a long entry (no orders possible for tomorrow since we are already short and todays close was high in the 5-day range).

 

Status: Realized + paper gain(loss) as of todays close

 

ES: -43.25 pts = -540.63 (1 ES contract per order)

NQ: +32.75 pts = +1310.00 (2 NQ contracts per oder)

ER2: +30.70 pts = +3070.00 (1 ER2 contract per order)

 

Net: +$3,839.38

 

Existing positions:

Short 1 ES

Short 2 NQ

Short 1 ER2

[Dogpile 10]

thumbnail attached

[TinGull 10]

Maybe I'm confused...but how is being 43 ES in the hole giving you only 540 bucks loss?

[Dogpile 10]

uh, no I am confused... was doing it x 12.50 (tick) rather than $50.00 (point)... will update

 

too caught up in details, trees for the forest thing.... you know what I mean

[Dogpile 10]

ES -43.25 pts x 50.00 x 1 contract = -2162.50

NQ +32.75 pts x 20.00 x 2 contracts = +1310.00

ER2 +30.70 pts x 100.00 x 1 contract = +3070.00

 

net = +$2217.50

[Dogpile 10]

got a buy signal on RUS today at 818.40.

will get buy signal on ES today if touches 1543.50.

 

I have done a lot of studying and created some new strategies based on the same underlying principle here which I will be trading in a real account starting 10/1. I have 2 filters and set some limit orders to exit trades which greatly increases the expectancy. Also, I will only be trading the long side with this strategy. For the short-side, I will rely on my primary trading account to make money there. I just don't think this concept is valid for the short-side -- or I should say, it is less valid on the short-side and there are better strategies for the short-side. For the long-side, 'range expansion off of opening price' is awesome. I have my own concepts which I believe in for the short-side, which I will trade on a discretionary basis. Thus, I am going to stop this thread here as I have changed the strategies and this forum has served its purpose for me.