OmniVest User Forums
General Discussion Topics
How Omnivest Works
42 replies, 7356 viewings
|
OmniVest User Forums
General Discussion Topics
How Omnivest Works
|
Last Activity 4/15/2019 7:43 PM 42 replies, 7356 viewings |
|
| Printer friendly version |
Mel Member Posts: 20 Joined: 3/4/2014  User Profile |
Mel  Attached file : Omnivest.pdf (22KB - 591 downloads) | ||
| Lompocian New User Posts: 3 Joined: 3/24/2014 Location: Lompoc, CA User Profile |
| ||
| Jim Dean Elite   Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
Not much point in them spending lots of hours documenting it till it settles down. [Edited by Jim Dean on 5/9/2014 5:17 PM] | ||
| BrianD Legend  Posts: 302 Joined: 2/23/2013 Location: Grand Rapids, MI User Profile |
I sense frustration from you and Lompocian. I’m also thinking many users are looking for guidance on ‘why’. OV and TP present tremendous feature/function for power users. Nirvana also understands there has to be simplified OV & TP for more basic users, or for Nirvana to provide ‘pick and go’ portfolios/account settings/TP’s – like a ETF or Mutual Fund. Stay tuned. For users who want to understand and control the why, keep it simple as you start. Take small steps. Create an Account. Think about your comfort in risk. Pick a few strategies by looking at charts and performance. Make a portfolio. Review portfolio over some time frames through account settings. Review risk/reward through changing account settings. Paper trade with a broker through TP. At the end of the day, trading is just a bloody tough business. From a personal perspective, after running OV for 15 months, and live trading with OV/TP for 11 months, I still feel 'lost' at times. But there are many supportive people on this wagon, so do not be afraid to ask for help. At the end of the day, Nirvana's success is predicated on our success, and ours on theirs. | ||
| Mel Member Posts: 20 Joined: 3/4/2014 User Profile |
Thanks for validating my understanding so far. Jim, This took me an hour to draw, but considerably longer to understand. As a hierarchaeolgist, this is my mental approach to understanding complicated processes. What I had to do was go through innumerable forum posts, and play with the various web pages to figure this much out. Videos tell me what Ed believes is possible to do with OV, but I personally need a mental model of how it works to be able to do something with it I can understand and trust and possibly contribute something to improving it. No such model is available unless I make one. Perhaps there are others like me out there. Mel | ||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
| ||
| Barry Cohen Icon  Posts: 1844 Joined: 10/11/2012 User Profile |
| ||
| Mel Member Posts: 20 Joined: 3/4/2014 User Profile |
OV's complexity for me comes from trying to image how changes in various parameter would affect things. I don't like to randomly tinker with things - my old physics professor told me that "every calculation should be preceded by a thought." To do meaningful experiments with OV, for me, it is necessary to have a white box view, not a black box view. Then I can construct experiments that allow me to learn, not aimlessly wander. The model now allows me to understand the basis for how to get to the final allocation to a position, to understand that portfolios are not yet object with their own properties that can balanced with each other - all orders are are currently summed across portfolios. The model says lists are confined to strategies, and putting strategies into a portfolio does not cause them to operate on all the lists in the portfolio the other strategies use. This kind of nuance is not obvious to the new user. I believe it says that equity curve conditions apply only to strategies, not portfolios, since portfolios are not yet objects. I am now at the state that I do not know what I do not know. As I find those things, I intend to learn them, and add them to the document, for myself, if not for others. Mel [Edited by Mel on 5/10/2014 8:27 AM] | ||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
A. an OVest "STRATEGY" is comprised of FOUR components: 1. an OVest "SYSTEM" which is in fact an uploaded OT "Strategy" (refer to OT for what it's strategy looks and works like) 2. specification of the trade DIRECTION(s) that the OVest Strategy will use, from the various L+S trades generated by the OVest System 3. assignment of an OVest Symbol LIST to be used with that OVest Strategy - can be static or dynamic - distinct from lists used by other Strat's 4. optional use of a CONDITION that acts like a filter for the OVest Strategy, blocking or passing Entry signals according to a Boolean rule - the Condition is an OScript formula based either on a symbol (typ an index) of your choosing, OR upon the Equity Curve of the OVest Strat thus far (unmodified by the Condition). B. An OVest Portfolio is a "basket" that holds one or more OVest Strategies, each with an Allocation percent that weights the sizing of trades from multiple strat's in the port, with the starting point position size being established by the OVest strat itself (that value is blackbox). This is btw one of the hardest-to-sort-out-and-follow calc's within the core OVest product. Just think of the percent as a weighting factor for the strategy and that will probably be a good enough handle. ** Currently, the OVest Portfolio is just a "basket" that conveniently groups strat's (you can define as many different Port's as you like). There are currently no Conditions that you can define relative to a particular Portfolio that impact how its component strat's are handled. This will change very soon, with the V3 addition of Portfolio Switcher (similar to Strategy Condition, but acting on the entire basket), and that Switcher will be dramatically enhanced/automated by Portfolio Balancer, a soon-to-be-released "OVestPro" tool. C. An OVest Account holds any number of OVest Portfolios &/or OVest Strategies (ie strat's not in a portfolio "basket". You can create an unlimited number of Accounts (for testing), and up to five live accounts for live(inc.paper trading) (to different brokerage accounts). The OVest Account has a huge number of "Settings" which I won't describe here ... look at the "Account Settings" panel. Of these, the one that has the most sweeping impact on your historical-simulation modelling is the date range at the top of the panel. All of the other settings affect both simulation and live trading with that account. The settings act as "filters" to control which trades presented to the Account by its component Strats (sep or part of Ports) are actually used, and also can dramatically impact the position sizing of those trades. If you change even one of those settings, it can have major impacts on your overall account performance. Be sure to test in historical simulation before going live! D. The Trade Processor is a "final step", applying only to LIVE accounts. Currently it is a standalone engine that you run on a local PC or in a personally-leased Cloud partition (many have successfully been running it in Amazon, and a few in Azure - lots of discussion threads about that). The TProc grabs info from OVest, grabs info from your broker (re avail funds), and tries its best to get as many if not all of the OVest trades for each day to be processed (exits then entries). Since OVest is not currently tied into the brokerage, it doesn't know how much money is avail, so it transmits its sizing for each trade to the TProc as a PERCENTAGE. This keeps it generic ... but also means that the TProc interpretation of the actual trade sizes differs somewhat from OVest's presumption when it calcs the ongoing equity of the account ... that is, the OVest equity at the HRE may not match your brokerage. *** This going to change significantly, fairly soon, as Nirvana brings up a Server-based TProc ... ie running on their servers, without you having to set it up or manage it re installation etc. From that point forward, many new "channels of communication" will be available between the OVest analysis/director functions listed in A-C, and the TProc execution logic. The TProc logic is going to be made much more "intelligent" as a result, and there should be little or no discrepancies between OVest tracking and TProc tracking, once that process is fully integrated. I think that covers all the big stuff. Of course there are a lot of howto details, and a lot of that is still in flux. I hope this helps! I'm reposting it in the knowledgebase (click here) so it doesn't get lost, and so that amendments and additions can be collected together with it as time goes on. [Edited by Jim Dean on 5/10/2014 10:02 AM] | ||
| Mel Member Posts: 20 Joined: 3/4/2014 User Profile |
1)Is looking at strategy equity curves a good way to compare strategies? Different strategies have different capital investments for different % of time. Are the curves we look at really a normalized comparison, or should I compare something else? 2) It worries to mix numerous RTM trades in the same bucket as low hit rate, high return trend following trades. If a high allocation RTM trade eats the capital the day the big winning trend trade should have been entered, I am destroying the equity curve of the trend strategy. To know if this is going to happen, it seems if I am trading liquid lists, I would need to divide the lists into non-intersecting sets of liquid symbols to assure that one set of strategies doesn't destroy the equity curves of the others. I don't know how to create non-intersecting lists in OT. Only non-intersecting list tests can answer my worry, I think. Has this been addressed. 3) Is there an easy way to make strategies that generate both long and short signals to only take short signals? Or vice versa? Mel | ||
| JimB Member Posts: 36 Joined: 3/7/2014 Location: Rogers, MN User Profile |
| ||
| SteveL Veteran Posts: 189 Joined: 10/11/2012 Location: Boulder, CO User Profile |
Good summary. I've decided to delete the remainder of this comment, because after more experimenting, I'm not certain how allocation is done. deleted [Edited by SteveL on 5/10/2014 11:48 AM] | ||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
1. equity curves ... give a very good "big picture" 2. summary statistics (TPM, CAR, MDD, etc) ... offer fast overviews 3. excel export ... everything behind the scenes, for you to play with ==== The "worries" of mixing various strat's together are actually "blessings" ... OVest allows you to easily bake your own cake. Yes, it takes time and thought and can be frustrating (or entertaining). Hey, that's trading! The upcoming OVestPro Portfolio Wizard (probably 2-3 months out) will automate that entire process, allowing you to define the metrics/goals you wish. In the meantime, for really detailed and testing, dump the data to excel and go to town (that's how Portfolio Balancer and Portfolio Wizard were born - done by dedicated, brilliant users ... now being implemented by Nirvana with their assistance - Steve Mayo and Mark Holstius are the stars of that epic - thanks, guys! ----- You can create your own lists ... either by uploading specific symbol sets, for static lists used as is through the entire test range and at the HRE (you can change the list whenever you want, manually) ... OR by creating custom dynamic lists that rebuild a new set of symbols every new day, based on OScript filtering rules - DL's are imho one of THE most powerful features of OVest. An infinite number of filter rules are possible. ------ You can turn individual Strat's off and on with Strat OScript conditions, and once the PortSwitcher comes online, you can control which Port's in an Account are active, or even "meld" them with weighting factors (eventually) ... that process will also use OScript conditions. OR, if you have OVestPro (a verreeee good idea), Portfolio Balancer will help you determine the best switching algorithm, and control it for you! ====== My orig description, part A.2, explained how to make Strat's use only Long, only Short, or Both. I thought that was clear. The raw Systems that the Strats are built from produce signals in both directions ... the modifier allows you to ignore some or none of them. **** General comment: I fully agree that it's best to know the theory before throwing money at the thing, or throwing a lot of time at it. BUT, some things are best learned by a LITTLE bit of info (I think I've laid that out already), plus spending time to PLAY with things to see how they work. Some of your questions thus far would easily have been answered from a little "learning playtime" ... other questions truly do hit on the heart of the difficulties ... and point to the cool solutions that are already in the works. [Edited by Jim Dean on 5/10/2014 10:25 AM] | ||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
I tried to present the description as I understand how OV is "supposed" to work, at least as originally envisioned. I am unclear, myself, on many of the algebraic magic that occurs to determine allocations and sizing ... and that all is CHANGING soon with the mod's to how Port's work, and a bit later, to some degree, with the tighter integration of TP. So, I tried to keep my explanations fairly general. Note that I said it's best to think of the strat alloc %'s as "sort of weights" and did not try to focus too much on it. I didn't talk much at all about Port alloc's since that will all be changing, too. I tried to sidestep the original question by deferring it till OV3 was in place ... to some degree, right now, this stuff not fully understood (ie guessed-at) by most of us. [Edited by Jim Dean on 5/10/2014 11:56 AM] | ||
| Mel Member Posts: 20 Joined: 3/4/2014 User Profile |
Thanks. I had my browser open and was writing during the time you posted your stuff. I found it after I posted mine. I think I have to go outside OT and OV to create lists with non-intersecting symbols. True? My worry about mixing is creating combinations that do better in testing than in real trading. I am just beginning to think of ways to isolate and quantify the problems. Mel | ||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
So, no, you don't need to go outside OV or OT to create lists. Both tools include that capability right now. Again ... I suggest that you start, like the rest of us have, by playing around with things to discover these basic functions. Think of it as an "apprenticeship" process. Once you become a Journeyman, try building some serious accounts. After doing that a while, you'll Master the process, and discover and disseminate new knowledge about the details along the way. And btw - this tool is still growing. NO ONE is a "Master" yet ... not even the Nirvana developers! Ed repeatedly emphasizes how he is seeing and learning new things every day! It's a process that's been proven over 2-3 millennia. :~) [Edited by Jim Dean on 5/10/2014 10:32 AM] | ||
| SteveL Veteran Posts: 189 Joined: 10/11/2012 Location: Boulder, CO User Profile |
[Edited by SteveL on 5/10/2014 11:48 AM] | ||
| Steve Mayo Legend Posts: 414 Joined: 10/11/2012 Location: Austin, TX User Profile |
The "default" trade sizes for Nirvana's "canned" strategies are fixed. For example, on R10-L-ELS is is 10%. For R10-L-SP it is 25%. You can determine each one easily by looking at the strategy details popup (when viewing a longer timeframe) and rounding it off. What is shown there is a backtest calculation (not the true default) that is averaged after generating the equity curve -- it takes into account the "allow settings to reduce trade size" setting at the account level...and maybe the share rounding (I've never used that one) so you have to round it off. Point is that it's a fixed default that gets modified! On strats YOU create using Strat Builder, you set that default trade size using the % per trade field. The "Alloc" fields at both the strat and port levels then simply modify that default trade size. For example, if you set the ALLOC for R10-L-SP at 200% and then use it in a portfolio having the ALLOC set at 50%, the result will be every trade from R10-L-SP (with rare exception) will get 25% of the available equity that day (25% default x 200% @strat level x 50% at port level = 25%). Again, in the current version, that DEFAULT trade size is fixed (in this example 25%); albeit, things may change in a few weeks as the Pro tools are implemented. The only exception is when on a day the trade is next in the daily candidate list and OV is trying to fit that trade into the available equity under the situation where "allow settings" is checked. [Edited by Steve Mayo on 5/10/2014 1:24 PM] | ||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
You said "On strats YOU create using Strat Builder, you set that default trade size using the % per trade field." Do you mean that the %/trade input overrides both the raw OT-strategy-TradePlan value, AND all/any effects of the OVest Account Settings? If so, is that a "permanent spec" for OVest? Seems to be pretty heavy-handed, to arbitrarily override some (potentially) very sophisticated rules. Is there any plan to allow the percent input in Strat Lab to have a "default" setting that doesn't override anything? I do understand that there's a need for doing apples to apples when testing many strats vs one list. But it sounds like you are saying that AFTER the lab session is over, the "custom strat" that's created includes this override forevermore. If so, I gotta object to that. Especially since that adds yet another characteristic to the strat that is NOT contained in the name. The rest of what you said is what I understood to be the case. Important to keep in mind that all of N's current OVsystems=OTstrats use fairly simple Trade Plans. Some of us think that scaling-in and out of trades is important ... TradePlans have that capability. I wonder what will happen when an OTstrat with a scaling-out TPlan gets uploaded to OVest? I hope that the "power" of that logic is not stripped off. [Edited by Jim Dean on 5/10/2014 12:34 PM] | ||
| Steve Mayo Legend Posts: 414 Joined: 10/11/2012 Location: Austin, TX User Profile |
And, it's easily testable. Here's a comparison of the "Nirvana Canned Version" of R10-L-SP with two I created. On one, I set my % per trade at 25%, on the other, I set my % per trade at 10%. Notice that the trades are exactly the same, only the Qty% is different -- and it's roughly 25% and 10% as per the "default" established for that strategy. EDIT: Re-reading your question. The default is for that strategy, it still goes through the "adjustments" at the strategy, portfolio and account level, but all are a function of the default. 2nd EDIT -- I meant to say the DEFAULT value is fixed - the ultimate trade size gets modified by the strat/port/account settings. Dang Jim, I'm glad you don't work for the FDA. He He. [Edited by Steve Mayo on 5/10/2014 1:37 PM] Attached file : https___www_myomnivest_com_Trades_Strategy_aspx_AccountID_23647_ID_1405_Popup_1.png (98KB - 443 downloads) Attached file : https___www_myomnivest_com_Trades_Strategy_aspx_Popup_1_Mode_all_ID_716732_HistoricalPositions.png (86KB - 414 downloads) Attached file : https___www_myomnivest_com_Trades_Strategy_aspx_AccountID_23647_ID_716732_Popup_1.png (93KB - 402 downloads) | ||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
Restated: 1. does the strat lab input override the default OVsystem (I think it's "yes") 2. does the strat lab input override the Account Settings rules (I hope it's "no") 3. is there a "use the default" planned, esp for ET uploaded strats where the sizing might be less arbitrary? | ||
| kmcintyre Elite Posts: 890 Joined: 10/11/2012 Location: Portland, OR User Profile |
Mel (and Lopocian), who appear to be very intelligent people, post their concern about how to use OV. Just basic usage questions... And we drill down into the minutia of how all the knobs and levers that must be navigated to use OV. (Because without understanding the minutia, using OV is kinda like throwing darts, or pulling numbers out of a hat...) I just got off the phone with my daughter and her husband. An hour + explaining the basics of OV. He's a PhD in bio chemistry. My daughter is a successful entrepreneur. They both looked at the site prior to our conversation. It's too complicated for them to embrace without my hand holding and assistance. I'm looking forward to OV3 because I think it will enable me to tweak and tune easier. My kids don't want to tweak at all. My kids are going to need a lot fewer knobs and levers before they can embrace OV on their own. (The knobs and levers can be there, but they need to be hidden and either default to good performing static settings OR auto-tune to provide good performance over time.) I know that is part of the vision. Until the vision becomes reality, I think OV is relegated to experienced traders who have a significant amount of time to experiment in order to embrace the product. Cheers Keith Keith | ||
| Geoff Veteran Posts: 180 Joined: 12/4/2012 Location: Byron Bay NSW Australia User Profile |
Originally written by 112582 on 5/10/2014 6:35 AM I am now starting to focus on Omnivest, and find it difficult for the new user to understand. I have built a little model that describes what I think I know so far. If those of you who are expert users would read it and help me fix it up, it could turn into a introduction for new users. I could use some help, since I have yet to try the trade processor. Mel Hello Mel, welcome. I think I know where you are coming from, I like to see things 'graphically' too, I find it can help to understand how the various components are related and how they work together. While I have been looking at and using OV for 12 months now, I'm still no expert but, FWIW I have attached my own flowchart of how it all works and hope this will help you get started. It doesn't tell you how the individual components work or are used, but will hopefully show you where they stand in the 'scheme of things'. I'm also hoping it will show you a typical 'workflow' for using OV too. If this flowchart is thought to be useful, or anyone can see an error or has a suggestion to improve it, please let me know and I will amend it. AMENDMENT. NEW VERSION 2 UPLOADED. [Edited by Geoff on 5/12/2014 6:24 PM] Attached file : OV Basic Use Flowchart Ver 2.pdf (129KB - 591 downloads) | ||
| Mark Holstius Elite Posts: 744 Joined: 10/11/2012 Location: Sleepy Hollow, IL User Profile |
Thanks for taking the time to put it together... Mark | ||
| JimB Member Posts: 36 Joined: 3/7/2014 Location: Rogers, MN User Profile |
|
|
|
|
Administrator
Forum Moderator |
Registered User
Unregistered User |
 E-Mail this thread to a friend(un)/Freeze thread | |
Toggle e-mail notification |
