OmniTrader Forum OmniTrader Forum
forums calendars search
today this week
 
register logon control panel Forum Rules
You are currently browsing as a guest.
You should logon to access more features
A Self-Moderated Community - ALL MEMBERS, PLEASE READ!
Vote for Members who contribute the most to your trading, and help us moderate content within the Forums.


  Current location        Thread information  
OmniTrader 2017 Upgrade Forums
Portfolio Simulation Mode
Portfolio Simulation Results -
Last Activity 6/13/2018 11:40 AM
26 replies, 2566 viewings

Jump to page : 1 2
Now viewing page 1 [25 messages per page]
 
back reply
Printer friendly version

^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : Portfolio Simulation Results -
Posted : 11/20/2009 5:16 PM
Post #18647

So I turned $10k into $2,264,400 dollars in 12 months using Portfolio Simulation - not too shabby, huh?!! hahaha..

well, I know its a bit unrealistic b/c of the trade plans.. so I wanted to get some others views or any ideas on what they would do? as if they used this strategy, what would they(you) wanna do to make it more attainable and realistic?

I understand I could keep the same trade plan as if i were trading it myself. But, let's say I did keep the current trade plan - and began making live trades, Meaning, if I were using Integrated Brokerage and I left the TradePlan order block as is which is default settings, would money be lost b/c settins are at MR/NPP? curious..

What i did was take the "reversal(stocks)" strategy that is available to everyone of us -
1. I added additional systems provided to us for those that have OT Pro.
2. I added a confirm block

(**See attached images so you can see which systems were enabled in the confirm block(some may only be from CPS3 plugin) and systems block (KC and Envelope systems).

[Edited by julesrulesny on 11/20/2009 5:20 PM]

Attached file : confirm.PNG (209KB - 287 downloads)
Attached file : Systems.PNG (214KB - 254 downloads)
Attached file : PS.PNG (178KB - 278 downloads)
Attached file : ReversalTrends.pdf (36KB - 283 downloads)
Attached file : ReversalTrendTrades.pdf (46KB - 290 downloads)

^ Top
Jim Dean

Idol
2000500100100100
Posts: 2820

Joined: 9/21/2006
Location: L'ville, GA

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/20/2009 8:54 PM
Post #18652 - In reply to #18647

Hint:
"Market Reversal" (MR) systems and signals are another name for Pivot Points.
Crystal balls offer superb reruns, don't they?
^ Top
Buffalo Bill

Veteran
1001002525
Posts: 258

Joined: 10/3/2006
Location: Stafford, VA

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/20/2009 8:58 PM
Post #18653 - In reply to #18647

Jules

I am guessing you "asked" this tongue-in-cheek, but I'll answer as if you didn't know the answer anyway (for anyone else who might be wondering)

1) survivorship bias - the "smooth movers" list is a technical filtered list of stocks. You wouldn't know they would become smooth movers last year but if you did trade one back then (20-20 hindsight) you could make some coin. Use an index (SP 100/500, R1.0) or all optionable stocks/NYSE and NASDAQ (IMHO you could filter this down using some simple technical filters like V(22)>100000 and not bias the list enough to invalidate the test)

2) MR/NPP is not an exit for trading. It's a back-testing tool only. You don't know when the pivot point was until **after the fact** - and that is when OT says you exited.. A MR exit is a perfect exit because of 20-20 hindsight. Put a simple tradeplan on instead - a 1-2 ATR fixed loss stop and a 1-2 ATR (or even 3%) fixed profit exit and then see.

MR/NPP is good to compare HRs of changes in your strategy or to compare strategies.

FWIW Nirvana doesn't use MR/NPP when designing systems anymore. Gianluca said they are using real tradeplans so as they develop and test they have real, tradeable results

Do a search on NPP/MR exits. Jim Dean did a good write up on it not too long ago. Also search "survivorship bias"

We all know turning that 10K into something that looks like an international phone number in 1-2 yrs is not realistic.

[Edited by Buffalo Bill on 11/20/2009 9:04 PM]

^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/20/2009 9:39 PM
Post #18655 - In reply to #18653

Bill,

Thanks!! if it wasn't for that discussion on NPP/MR's, i would've jumped on this strategy thinking its legit and i hit the holy grail!! this is the second time in recent weeks where i've posted a similar thread about my 20,000+% ROI. LOL!

my source smooth movers - this is why i post these questions b/c i was actuallly thinking "Smooth Movers" was random enough.. altho smooth moving, i thought it would be a random 'enuff' set of symbols.. yet your idea of just using an Index or Optionable Stocks with your filters for Volume is simple and obviously the most realistic!!

All this talk about NPP/MR, I had NO inkling that it was for 'back-testing' only.. errr, what's the point of having it available in backtesting if its skews the results so much!

with a simple thread like this, it becomes very empowering b/c i learn such helpful, simple hints that go a long way. now I can narrow down my scope with a more simple realistic trade plan.. and i will look forward to hearing more ideas.
So, in saying that - if i were to use a "Trailing Profit Stop" / "Stop Market", that won't allow OT to trade in hindsight will it? meaning, its ok if i were to use that as an exit without skewing the results or trading with a crystal ball? (after posting, I already added your idea for the 1-2 ATR Fixed Loss Stop to my trade plan- again, i thank you, and i'll post my results!)

Also, i'm going to do a search now on 'survivorship bias'. and maybe i should read that 'next pivot' point discussion again which i found here: (if anyone is interested..) http://www.omnitrader.com/omnitrader/proforum/thread-view.asp?threadid=2525
^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/20/2009 9:41 PM
Post #18656 - In reply to #18652

ggrrrrr I hate NPP and MR's!

See Jim, look how much you've taught me in such a short span? i don't think any of my teachers from elementary school thru college have been able to teach me as much as i've learned from you and in these forums.. i guess b/c i'm actually interested in this.. lol.
^ Top
Greg Winch

Member
25
Posts: 42

Joined: 1/28/2007
Location: Phoenix, Arizona

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/20/2009 10:29 PM
Post #18658 - In reply to #18656

Jules,
Don't feel alone.
Earlier this year I took a similar trading strategy live.
It was only after my account was down about $7,000 that I started scratching my head and studying the strategy further.
NPP got me good! Losing real money is a good but hard lesson. If I'd paid more attention to some of the great posts in these forums, it's a lot less expensive lesson!
Make sure you check N-Bar performance and good trading!
Greg
^ Top
Jim Dean

Idol
2000500100100100
Posts: 2820

Joined: 9/21/2006
Location: L'ville, GA

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/21/2009 9:48 AM
Post #18661 - In reply to #18658

Remember:

NPP/MR functions ARE useful for "development" of alternative ENTRY modes, but should NEVER be used for actual trading

=============

Also Remember:

Backtesting survivorship-bias is a problem only when there is significant filtering (of any sort) in your INITIAL symbol list.

The easy way to avoid this is to:
1. use a pretty big, raw list like Russ3k or Optionable or SPOR (my term = SP1500+Optionable+Russ3k)
2. no Fundamental or technical filters that affect the initial list
3. include ALL Technical filters that you WILL BE using for real trading into a Filter Block in the strategy(s) you are testing

By putting the tech filters into the Filter Block for backtesting, you are telling OT to RE-EVALUATE the filter with each new bar ... just as if you'd run a fresh OmniScan that day.

Downside is that the actual backtest may take a lot longer to run - first, because you are using a huge starting list, and second because the filters (if complex) take time to run.

The slowdown is COMPOUNDED by the fact that the strategy has to run ALL the PRIOR BLOCKS (notably Systems) BEFORE it checks the Filter block to see if the symbol should even be considered. A huge waste of time, IMO. I've asked N many times to permit a Filter Block to be placed BEFORE the System block, for exactly this reason (to simulate OScan Tech filtering without Survivorship bias).

However, this is an IMPORTANT effect (filtering as you go), so be sure to follow steps 1-3 in ALL cases when you are trying to get a feel for real-life-trading potentials.


[Edited by Jim Dean on 11/21/2009 9:50 AM]

^ Top
Greg Winch

Member
25
Posts: 42

Joined: 1/28/2007
Location: Phoenix, Arizona

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/21/2009 10:44 AM
Post #18662 - In reply to #18661

I agree with you, Jim. Placing a filter block before the system block would save alot of testing time for those of us that like to tinker with different scenarios in the Port Sim and also inside the Strategy Wizard (I think).
Has Nirvana ever responded to you as to why we are not allowed to do this?

Also, I like your Steps 1-3. The only filter I have ever used on lists in Omniscan is a volume filter (SMA(V,200)>1000000 or something like that). It tends to filter out stocks that do not have alot of trading activity. That is just my personal choice and again, it saves on backtesting time.

Greg
^ Top
Jim Dean

Idol
2000500100100100
Posts: 2820

Joined: 9/21/2006
Location: L'ville, GA

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/21/2009 11:32 AM
Post #18663 - In reply to #18662

I have discussed this with Ed several times over the past few years, both via email and in person.
I've formally requested it in Feature Requests, and through personal contact with other staff.

The anwser I have been given is that it is not necessary, since it can be done in the Filter Block already, now.

I've tried to argue the "speed" case but it seems not to matter. I'm not sure if Nirvana cares much about the Survivorship bias thing. Certainly the filters that they typically use are very simple, so they don't take much time to calculate.

OTOH, I am a proponent of robust pattern-based filtering, to find good-personality stocks. These filters take a lot more CPU cycles than the more typical ones that people use in OmniScans.

I've tried to illustrate it this way for them:

Presume some SIMPLE filters, but ones that are very restrictive, which would normally narrow-down the Russ3k to a filtered Focus List of only 300 symbols.
Also presume that the Strategy uses 20 SIMPLE systems to search for signals, and votes them, prior to the filter block. OR, presume the Strategy uses 3 COMPLEX systems to search for signals.

In either case, for REALTIME trading, the strategy only runs the 20 (or 3) systems on 300 symbols, since OmniScan has prefiltered out 2700 symbols first ... sensisble.

But if you properly protect yourself from survivorship bias for backtesting, you must use the full Russ3k as the FL, and build those simple filters into the Strategy.
So, for backtesting, each of the 20 simple (or 3 complex) systems must be run, EVERY DAY, on 3000 symbols, and then voted, before the filters come into play.

So, we have AT LEAST TEN TIMES as many CPU cycles PER BACKTEST DAY than we would by using OmniScan in real life.

This seems absurd to me.

So, I've proposed that they create a simple PreFilter or PreScan block that permits the same Technical Scans that OScan does. It cannot do fundamental scans, since OT does not know about historical fundamentals (yet). And it cannot re-eval the source-lists, since OT does not know about how the Russ3k changed over time. But it CAN do the Technical filtering, which as active traders is at the heart of our concerns anyway.

I've had no success.

It would help ALL OT USERS ... would not require OTPro or even OmniScan ... just Formula Builder.

Maybe if some other folks request this (reference this thread or copy my example if you want), it will appear to be more than just a "wierd Jim Dean idea".

The good news is that it IS DO-ABLE now, but you need a fast machine and some patience.


^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/21/2009 3:28 PM
Post #18667 - In reply to #18647

Jim, Greg, Bill:

I never once thought of adding a filter block before 'systems'.. that's an interesting take. i can see how right off the bat we could filter out a bunch of symbols with some simple technical formula's. i really like that.

obviously we can't, but why can't we? I understand Jim talked to ed about this but you'd think its something easy to do, no? However, from what ive learned, we have the best product on the market b/c OT is different from other trading and charting applications.. On a different subject, i have learned with this whole "API" thing.. as in we can't just integrate with any old broker we please or get any data feed we please b/c of the coding involved. i understand the way our OT processes symbols is quite different from the rest.. so this brings me to a separate opinion on my point that throwing in a filter block before the systems block maybe alot harder to do than it sounds.. but to me, all it sounds like is getting an omniscan function/criteria to get first dibs to be the first part of a strategy.

JIM - your ideas are limitless and i am going to try and add your 3 steps/points into my TP. many gracious and wonderful thanks as always for your feeback!
^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/21/2009 3:35 PM
Post #18668 - In reply to #18663

Originally written by 210921 on 11/21/2009 12:32 PM

It cannot do fundamental scans, since OT does not know about historical fundamentals (yet). And it cannot re-eval the source-lists, since OT does not know about how the Russ3k changed over time. But it CAN do the Technical filtering, which as active traders is at the heart of our concerns anyway.

I've had no success.




I'm confused by this comment Jim - that Oscan can not do Fundamental scans? well, i know we can actually but can you elaborate? b/c i know if i put in Financial information (or even the Nirvana list 'Low P/E with Earnings Growth' this is a Fundamental scan itself, etc) i'm confused on how OT does not know about historical fundamentals, (yet).. my guess is that you are referring to formula builder in a filter block, in a strategy. (that we can't access fundamental information in a filter block?).
^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/21/2009 3:58 PM
Post #18670 - In reply to #18658

Greg,

If you don't mind I'd like to ask if you can elaborate on how i can check my 'N-bar' performance? what do you mean by that? .

(i do know of it being a function in a strategy(filter/confirm blocks,etc)

this is something i now should know, but i don't as i'm still learning. in the attached images, i circled the areas i don't quite understand, nor how to use them properly. (parameters and performance).

thanks,
marc
Attached file : Parameters.PNG (172KB - 269 downloads)
Attached file : Performance.PNG (216KB - 209 downloads)

^ Top
Jim Dean

Idol
2000500100100100
Posts: 2820

Joined: 9/21/2006
Location: L'ville, GA

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/21/2009 4:09 PM
Post #18671 - In reply to #18668

"It" was referring to the Filter Block.
OT is limited to current info about fundamentals (incl limited items re a few quarters ago).
OT does not know what the fundamentals-data was that would have been reported at the end of the day, Nov 20th, two years ago.
Any fundamental data known today is "future" to any backtest.
Similarly, the composition of a given Index such as the Russ3k two years ago was diff than today - it probably included a considerable number of symbols that are not traded today, in fact.
Worden does provide historical fundamental dataseries. You can actully plot how the most recent quarter earnings has changed over the past twenty years. You can do MA's of that data, make it part of backtests, etc. For all the morningstar data.
I'm hoping that will become a feature offered by OmniData.
Afaik there are no cost effective sources for the historical composition of the major indexes.

^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/21/2009 4:37 PM
Post #18672 - In reply to #18671

thanks jim!
^ Top
Buffalo Bill

Veteran
1001002525
Posts: 258

Joined: 10/3/2006
Location: Stafford, VA

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/21/2009 9:06 PM
Post #18675 - In reply to #18670

Jules

A good technique to check the validity of your system is to run some backtests with N-Bar exits only (exit at day N) like 3, 5, 7 ,10, 20 - whatever you think your holding time would be plus or minus some time. If your entries are good the stock should move in your direction pretty quickly (we are traders not investors after all) so the strategies should be profitable even exiting at day N. Do several runs with different days not just "3" or "10" A "real" tradeplan should do better than an N Bar exit, so if it's profitable for N-Bar exits it should be better wit real exits (if they are well thought out and meet your trading goals)


^ Top
Greg Winch

Member
25
Posts: 42

Joined: 1/28/2007
Location: Phoenix, Arizona

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/22/2009 8:00 AM
Post #18684 - In reply to #18670

Jules,
When you set up or modify a strategy, click on the systems block.
In the lower right hand area of each system, there is a Performance Metric that gives you a choice of either N-Bar or Next Pivot Point. Always check the N-Bar metric before running an analysis of your strategy. This will ensure closer results to real life trading.
Also, Bill's suggested method of using a Trade Plan in your strategy that initially incorporates an automatic exit after being in the trade for X number of bars is a good idea. Different strategies require looking at different number of bars, depending on your time frame. I usually start with 5 bars for my shorter term EOD swing trades.
Greg
^ Top
Jim Dean

Idol
2000500100100100
Posts: 2820

Joined: 9/21/2006
Location: L'ville, GA

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/22/2009 8:43 AM
Post #18685 - In reply to #18684

A more robust flex-N-bar approach is to write code that exits if the rate of price improvement has been very slow in recent bars, or if the overall rate since entry has been very low.

Usually exit logic of this type would not come into play UNTIL some arbitrary number of bars had passed.

There are many theories and ways to implement this ... but it's paradigm is similar to N-bar ... the idea is to give the trade all the time it needs, as long as it is moving in the direction you want, at a reasonable speed.

Pure N-bar is useful for testing. I think modifications are in order for real trading, however.

I also feel that NPP is useful for measuring the efficacy of different entry methods. It generally exits at the same points each time (for alternative entries that are fairly close to one another)... N-bar exits at different points, depending on which bar you enter on.

NPP can also be used for testing EXIT strategies ... if you hold your ENTRY method constant, then NPP provides an "ideal exit" measure against which to test differing real-life trade plans.

So, NPP is not "eeee-vill" ... it just needs to be used wisely. The only "problem" is that it is a "default" in too many of the canned systems, etc. IMO that should be changed. But NPP (aka Market Reversal) *is* a useful TOOL for strategy development purposes.

^ Top


This accout has been deleted
 
Subject : RE: Portfolio Simulation Results -
Posted : 11/23/2009 11:31 AM
Post #18698 - In reply to #18653

Buffallo,

Just a quick thing to note.... Nirvana still uses the NPP - all strategies on the Money Zone use a NPP
^ Top


This accout has been deleted
 
Subject : RE: Portfolio Simulation Results -
Posted : 1/29/2011 4:32 PM
Post #23102 - In reply to #18647

jules a cheeky question to ask but for some of us who dont know how to edit strategies,,if you got goot results could you share by putting the strategy on the forum instead of screen prints?
would be very much appreciated by a novice newbie
^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 1/30/2011 6:06 PM
Post #23124 - In reply to #23102

Ok, no worries. Though, its been a while so I have to check to see if I still have that strategy. No big deal though. If you don't want screen prints, how would you like it then? Just the name of the strategy? name of the indicators, systems, types of trade plans used?? b/c most of my strategies are edited from the original so if you want to use what I have used, you would need to make the same changes yourself or I can upload the strategy for you to save as a file directly into your ...Nirvana/OT**/Strategies file.
All of which is very possible.. Editing is far easier than you may think. If you understand how to combine indicators and systems in a chart, then its relatively the same concept when combining them in a strategy.. be a little more specific on how you'd like to view them and I'll see if I could get it out to you..
^ Top


This accout has been deleted
 
Subject : RE: Portfolio Simulation Results -
Posted : 2/20/2011 6:11 PM
Post #23439 - In reply to #18647

for those of us that dont have the brains lol
could you please please just post the strategy on here for us to download

also

ive read on forum dated years back about the holy grail strategy

does it work on ot2011 and can you post that too for us to download?

i have ot2011 pro

many thanks guys
^ Top


This accout has been deleted
 
Subject : RE: Portfolio Simulation Results -
Posted : 2/20/2011 6:30 PM
Post #23440 - In reply to #18647

can i have your opinions please
are these results ok good great or what????
what should i change?
any advice welcome
ive strategy report and port sim results

Attached file : portsim.JPG (517KB - 208 downloads)
Attached file : report.JPG (552KB - 206 downloads)

^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 2/23/2011 12:34 PM
Post #23473 - In reply to #23439

FYI - There is no such thing as a "Holy Grail Strategy" and if there was, it may only work out for that person who found it. what works for one trader, most certainly may not work for another.. It could've just been a title that the trader had made up.
So, there's no Holy Grail strategy to download or post. YOu need to find a strategy that suits your trading style and personality. I didn't quite understand this concept either when I first started out.. i just always thought I can get a strategy from someone else and use it to make money. what you can do is get a strategy from someone else and then make it your own to develop your own ways of utilizing any particular strategy. Finally, you will also find out one day that Trade Management and a Trade Plan is by far the most important aspect of trading. It will save you and make you money in the long run. A good trade plan can make a poor strategy profitable! (in my opinion of course..)

Any strategy from previous versions should work on OT, especially if you have Pro.

And, which strategy were you referring too? (besides the Holy Grail strategy) The one in my original post in this thread? If memory serves me correct, I believe this was one of the default strategies that came with OT. If not, I'll have to check my OT to see if I still have it saved. I should though.. Remember that this Portf.Sim. run was not accurate. It may have tested this strategy and turned $10,000 into over $2million but was skewed and incorrect. Without getting into specifics, it had to do with NPP (Next Pivot Point) in the "Trade Plan / Order" block of my Strategy. this told OT to exit out of the trade on the NPP. Do a search here in the forums for NPP and you will get more information. What happens is when testing and you have a NPP as an exit, its like OT knowing the future when exiting the trade. If you are good and fully understand NPP, it can be very useful and profitable.

Also, check out SWL (Signal Watch Live) from Nirvana.. i ahven't used it in a long time but its a chat room where (most OT users) traders gather thru out the trading day to discuss trading, strategies used, particular trades that they may be in, etc.. Its free too and a great way to get acclamated to trading and OT as well..
^ Top
Mel

Veteran
10010025
Posts: 233

Joined: 12/30/2003
Location: Kensington

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 2/23/2011 4:38 PM
Post #23474 - In reply to #23473

Actually, there is a Holy Grail strategy, from the Street Smarts book by Laurence Conners and Linda Bradford Raschke. Its basically a bounce form the 20MA in the direction of the trend. It is posted somewhere in the pro forum, I think. If uou can't find it under that name with a serch, let me know and I will dig it up for you.

It requires that you have the Pro version of OT.

Mel
^ Top
julesrulesny

Elite
500100100100100
Posts: 900

Joined: 8/15/2007
Location: NYC

User Profile
 
Subject : RE: Portfolio Simulation Results -
Posted : 2/23/2011 5:00 PM
Post #23475 - In reply to #23474

Oh. Ok, I misunderstood. I was thinking it as being the "Holy Grail" of all strategies..
which, a bounce off from the 20ma is probably as good as you can get in simplicity to use for a reversal.

Thanks for that follow up..
Jump to page : 1 2
Now viewing page 1 [25 messages per page]
back reply

Legend    Action      Notification  
Administrator
Forum Moderator
Registered User
Unregistered User
E-Mail this thread to a friend
Toggle e-mail notification


Nirvana Systems
For any problems or issues please contact our Webmaster at webmaster@nirvsys.com.