Trading Silver with PSAR SUPERTREND EA plus ZigandZag & PSAR Oscillator Confirming - page 6

New comment
 

Hi JJ,

I think you have a very good point. There is a balance between stops being too tight and getting stopped out early to too loose stops which can produce good profits but potentially can wipe out all gains when occasionally hit. On most trades, I see that nearly 75% or so of the profits are lost due to no Take Profit. That is why Zazzy has a Max Profit Number.

I am considering two solutions to increase profits. First to issue up to 3 orders, the first two with TPs. Move the SL to Open when the first TP is hit for the remaining 2, creating a "free" trade for the last 2, then move the last SL to the 1st TP when the 2nd TP is hit. Then if there is an extended trend, you catch it with 1 order but if the trend reverses, you perhaps have either a break even or average to good profits. However, the big problem is that you can potentially have 3X losses if the first TP is not hit. Trying to implement this is where all the errors in Zazzy are occurring; it's not ready for testing yet. I have not considered supplemental orders in the same direction although conceptually it is probably easy to do. Will need to implement money management to insure control of the funds.

The other thing I am thinking about is using the SAR for stops by creating a "Super SAR". This would use the SAR calculations but modify them in one of several ways:

A stop based on the distance between the SuperTrend and SAR values.

Creating a dynamic SAR Calculation so that the exponential Step is adjusted based on the Trend Slope or some other rate of change of the prices.

Using the Support/Resistance to reset the SAR start.

The objectives of the three alternatives is to make the SAR a better SL and to get it to "flex" into the prices sooner after a long trend so that less Potential Profits are lost. A side benefit is that the order identification would be more timely and accurate, especially for trends that have an abrupt reversal.

Have not done anything to the two EAs, so don't expect a release this weekend.

Cheers to all,

Tzuman

 

I like the way you think, Tzuman!

Don't know which solution you've described will prove best. Only way I know to find out is to test them.

If you can use v202 as the test platform, since it seems 'lighter on it's feet', we can test each variant over the same time periods and see how they react to different market situations.

These can be used to build ZAZZY, though I personally would also like a ZAZZY-LIGHT for use with "speed-challenged" PCs. :-)

I've read some threads where the EA used different strategies in trending & ranging markets. Don't recall how well that worked out.

Of course, money management is critical. Some EAs allow you to partition the account between different currencies and set max drawdowns to keep from totally blowing up accounts. Leaves something to start over with. That is good for catastrophic loss protection, but doesn't overcome a poor investment strategy.

One thing that makes me nervous (and this is not a criticism, as I see it in myself) is having a 'wish list' of goodies that I want on the EA, only to see it become a dinosaur that can hardly get out of it's own way. Ah well, human nature!!

Finally, I've attached a snap-shot of the Feb 4 to Feb 22, 2011 period from the visual test I ran last night. It gives a good visual of what I described earlier, where the EA would only close the positions when a trend reversal was signaled.

Meanwhile, have yourself a great weekend, Bob. Its always good to see someone keeping a balance in life!! I'll be dinking around with the EAs we've got and helping a friend remodel his step van into a BBQ truck.

Joe

 

To everyone:

As you read through the earlier postings and look at the test Reports, you will see that we still have a lot of room for further improvement. In particular, note that with 'perfect' trend reversals the SILVER test gave a theoretical $38,000 return on a $1000 investment in less than 6 months.

Mind you that this is THEORETICAL!

Our best tests using the !$_PSar Supertrend EA v202 achieved profits (not including the original $1000 investment) of $1,022.78 on the EURO and $5348.19 on SILVER.

Tonight I'm watching the !$_Psar Supertrend v202 that Tzuman worked on trade in Strategy Tester using the month of February, 2011. And I'm wondering:

What would it take to get the EA to trade LONG or SHORT after the ZAZ signal setup and the PSAR trigger (or visa versa) and KEEP THE POSITIONS OPEN until the reverse signal was triggered? The EA could ADD positions in the indicated trend direction based on Supertrend or the ATR indicator or whatever, but wouldn't close them until the reversal was signaled. At that point it would close ALL open positions.

Any ideas?

JJ Johnson

 

Hello together,

last night i posted a message to this forum but i can`t find them here. It will be loosed?

O.k. Below a short summary of all posted EA-Versions. What EA-Version should be used for testing? What currency, metals, etc. and TF should be tested? Include Zazzy1.04 the Version202-lebmode (OrderStrategy 1 or 2)?

#1 PsarSuperTrendEAv202

#9 PsarSuperTrendEAv202-leebmod

#12 PsarSuperTrendEAv203

#22 ZAZ PSAR V1.00

#23 ZAZ PSAR V1.01

#33 ZAZ PSAR V1.01 (same as #23?)

#37 Zazzy v1.03.zip

#39 !$ PSAR SuperTrend EA v.202

#39 !$ Zazzy v1.04

#46 Zazzy v1.04.zip

This are generaly question for to organize the work for my self and understanding your principles by following the forum.

Nice Weekend!

 

Hello Trade4Gain,

Good question. Use the post 46 Zazzy v104.zip. The experts directory in that file contains both Zazzy v104 and PSAR SuperTrend 202a which is an updated version. Use the Feb 4th to Feb 22nd interval for testing as JJ did. It will be very interesting to see the results on 2 different vendors.

Cheers

 

Hello,

it`s not realy easy to understand the ea-versions.

I attach 3 Reports by using following ea`s:

1. !$Zazzyv1.04 Strategy 1

2. !$Zazzyv1.04 Stategy 2

3. !$PsarSupertrendv202a

The TF and Currency is the same.

Number 3 i understand the settings and working.

But Number 1+2 i don`t understand. By reading the post should they open the same order but different closingways by handling SL+TP. The ea should use the ZAZ ActivePeriods with a cross of the psaroszillator in the way of the arrow. But it opened different orders.

Can you pls give a short explanation how the ea works in strategy 1 and strategy 2?

Currently i will read the previous postes again, perhpaps i may find my thinkingmistake.

Cheers

 

Hello Trade4Gain,

I looked at your Zazzy runs and here is what comes to mind

I have no experience trading with Silver as a contract on any EA. If I look at the test runs, it appears that it has a pip value =0.01. 20 pips may be too large a SL value. Also I have not thoroughly tested it with 4 digit brokers.

PSAR Supertrend has MaxOrders=100. and you set Zazzy to the same value for a consistent comparison. Within Zazzy, this will result in up to 100 orders being placed, depending on the available margin. This value of 100 is incorrect and should be 1. For PSAR Supertrend, it appears that the value different interpretation as your #3 run looks closer to what I would expect. I suspect it is only used in Money Management calculations.

Strategy 1 uses the last close 2nd last and compares to the corresponding values for the SuperTrend and PSAR and also checks that the PSAR oscillator crosses zero. Strategy 0 does not use the PSAR oscillator as I believe the conditions Close1>PSAR1 and Close2<

PSAR2 are identical to the oscillator passing through zero.

Strategy 2 uses the SAR oscillator crossing zero and the ZAZ buy sell indicators occurring within the LookBackPeriod, 4.

JJ & I have found that Strategy 2 does not produce any winning results. Your test #2 also confirms this result.

Your tests were very helpful as they identified two areas for correction and testing. I will evaluate the MaxOrders in PSAR Supertrend to determine its use.

 

6/24 1am post

Hello JJ,

I somehow missed this post. Yesterday I spent an hour responding to it and shut down Firefox before posting it. Got so mad, I shut the computer off LOL.

Anyway, if we find an EA that is profitable but doesn't trade for weeks, I would be extremely pleased. I believe that a trader needs several different profitable EAs that complement each other.

On trading multiple currencies, Zazzy has this already built in. The NumberofEAs is used in this condition in conjunction with Money Management to allocate free margin across several ea's in one account. In addition, Zazzy calculates a unique Magic Number for each time period so you can run it the same currency with two or more different time periods and each EA will control only its own time period.

If you do not use any SLs or TPs, both EAs should keep orders open until there is a reversal signal.

On "speed challenged cpus" I hope you are not using a Pentium 5 with Window 2000, LOL. Seriously, this is a problem, I find indicator code is extremely inefficient. Compounding the problem is that we generally have two separate instances of the indicator running, one on the screen and one within the EA, so our inefficiency is 2X. When I find improvements, I generally tweak the code and put it into the !$_ category for differentation. On top of this MT4 is extremely inefficient, I've read that MT5 is 10x faster.

There are a number of good indicators that are cpu intensive, iVAR for example, see the Digital ASCTrend thread. This seems to be good at identifying when not to trade but it will stall a slow cpu. However, we must recognize that this is a business that does require a lot of calculations. if iVAR substantially improves an EA's profitability, it's a no brainer trade off requiring a new cpu and better profits vs lower returns with an old cpu. I just bought a Asus laptop with last year's cpu, I7 740qm and it handles everything very well. BJs has one on sale through June for $750, it is not the latest but it is great!

Cheers

 

Hi Tzuman!

Hope you had a good weekend, despite any frustrations!! :-)

As to computers, I've got a Compaq desktop with Windows XP. It came with a Celeron 2.80 G/Hertz CPU with 1.24 G/B RAM. Seemed pretty fast at the time.

Don't know how it compares to the new ones; haven't seriously done any comparison research, but doesn't surprise me if I'll soon need to upgrade. Are they effectively using 64 bit dual processors or is this more of a marketing gimmick?

JJ

 

To Tzuman:

Bob, you're not the only one who has missed postings. After reading over your replies to T4G, I realized I had missed your Zazzy 1.04/v202a zip file.

Took your suggestion and ran a test using v202a without the Trailing Profits or Trailing Stop. Got a good return on the EURO, though the BUY/SELL signals didn't exactly match the trend change signals. Didn't take any short positions from Feb 22 to Feb 25, but I'm OK with that in such a short period. Taken all together, quite encouraging!

On SILVER, it's clear the v202a is still selling during LONG ONLY periods. Not sure why. The return was great, but the drawdown was painful. Without those errant sell orders, it wouldn't have been half as bad.

Can you take a look at them and let me know what you think?

Joe

12345678910
New comment
Reason: