Is fully automated trading a fantasy? Your opinions here
- Thread starter ybfjax
- Start date
I am 80% sure I understand what the automated systems at RT do, extrapolating from Simmons background.
The part I don't get is how the algorithm decides what direction on the manifold (with a Ricci like tensor at each point, so the manifold has geometry) is the direction the given underlying is going to take. That may be based on statistics though.
So all they may be doing is forming a Ricci tensor (or a mathematical structure like it) with each point being from a probability measure set, instead of a traditional topological set. The set (manifold) then has both a geometrical and a probability structure imposed on it. They then hedge with futures or ETFs to smooth out the risk on fill in the equity.
In order for this to work, you basically have to be able to trade and borrow at close to "zero" rates...
I plan to write something on this, but I don't know if I will...If I do, it will be here, http://www.synctrading.com/phpbb
nitro
The part I don't get is how the algorithm decides what direction on the manifold (with a Ricci like tensor at each point, so the manifold has geometry) is the direction the given underlying is going to take. That may be based on statistics though.
So all they may be doing is forming a Ricci tensor (or a mathematical structure like it) with each point being from a probability measure set, instead of a traditional topological set. The set (manifold) then has both a geometrical and a probability structure imposed on it. They then hedge with futures or ETFs to smooth out the risk on fill in the equity.
In order for this to work, you basically have to be able to trade and borrow at close to "zero" rates...
I plan to write something on this, but I don't know if I will...If I do, it will be here, http://www.synctrading.com/phpbb
nitro
i never heard of simons until someone just posted a link http://en.wikipedia.org/wiki/Renaissance_Technologies
i worked for a similar secret firm call regal asset management who traded the who's who of people like Wal-Mart, Jimmy Dean, Mary Kay, and on and on. Regal is no more and the top trader Stan Finney (who i worked for as his quant) is bigger and more profitable than ever with SPYGLASS TRADING, L.P.
while there i was trained by ned davis ndr.com in the use of their platform. i also had an unlimited "practically" budget to find the holy grail. i feel that i did and i left. via many poor partnering decisions i have not owned the world yet, but feel i could one day "kidding". i feel my research is that solid. the application of that research required a trading platform which didn't exist and that has delayed some things.
i had a start up cta in 1998-2000 which was very touted and partner greed destroyed that. i had another partnership and ego destroyed that. i now have a partner who we both just want to beat it. the money is not the issue with either of us. its all about achieving perfection which we feel we are close. both application wise and system wise. we have exhaustively gone over every detail and prepared for failure.
so i would say we spend 75% of our time now on what if's could go wrong and scraping for every small edge whether that be clearing cost, network speed etc. the other 25% is spent on refining the systems.
mark brown
Anyone who has a fully automated profitable system for sale should email me 
Interested in having one. I have a tradestation program for er that is up 2000% 7 months but on the cash market. Need to program one on futures with related signals so it can trade automatically. Anyone interested in working on this can email also.
Interested in having one. I have a tradestation program for er that is up 2000% 7 months but on the cash market. Need to program one on futures with related signals so it can trade automatically. Anyone interested in working on this can email also.Quote from TraderRobb:
Anyone who has a fully automated profitable system for sale should email me
i dont understand you say you want to buy a system? then you say you have one that makes 2000% in 7 months?
i mean no wonder the trend line guys and the brokers just laugh out loud when they hear the word system. its about like ever guy who tells his wife "i could have been a race car driver"
mark brown
<i>"i dont understand you say you want to buy a system? then you say you have one that makes 2000% in 7 months?
i mean no wonder the trend line guys and the brokers just laugh out loud when they hear the word system. its about like ever guy who tells his wife "i could have been a race car driver" </i>
#1: Cash RUT and futures ER2 will not cross over system-wise. I can tell you from personal experience that writing systems in cash stock index markets and trying to apply logic / execute in the futures market will not come out the same... probably not even close.
#2: Fully mechanical systems are possible, but not easy to find or construct. I wrote dozens of systems for FX markets using all manner of tech analysis tools. Oscillators like RSI, etc or moving averages absolutely will not hold up over time. Neither will anything too heavily tweaked or multiple-rules based.
Attached is an example, one of two "sister" GBPUSD systems results I wrote in mid-2004 that are walking forward untouched, no modifications since late 2004.
Factored in -$40 per trade on the bid/ask spread of -4pips slippage, no commission cost reflected. One full-size FX contract straight thru, no adding size or pyramiding.
*
It just so happens that GBPUSD tests out best = most profitable in almost every example of systems I wrote. EURUSD and AUDUSD were next, but probably 30% less profitable. Anything related to Yen can be profitable overall, but the drawdown curves are usually gut-wrenching.
*
There are numerous commercial systems for sale that hold up over the course of time... with periods of months to years sometimes unprofitable or b.e. at best.
I'm aware at least one fund manager who earns seven - eight figures annually with managed FX accounts using nothing but commercial-based systems purchased and run verbatim (or slightly tweaked at most) in basket portfolio fashion.
i mean no wonder the trend line guys and the brokers just laugh out loud when they hear the word system. its about like ever guy who tells his wife "i could have been a race car driver" </i>
#1: Cash RUT and futures ER2 will not cross over system-wise. I can tell you from personal experience that writing systems in cash stock index markets and trying to apply logic / execute in the futures market will not come out the same... probably not even close.
#2: Fully mechanical systems are possible, but not easy to find or construct. I wrote dozens of systems for FX markets using all manner of tech analysis tools. Oscillators like RSI, etc or moving averages absolutely will not hold up over time. Neither will anything too heavily tweaked or multiple-rules based.
Attached is an example, one of two "sister" GBPUSD systems results I wrote in mid-2004 that are walking forward untouched, no modifications since late 2004.
Factored in -$40 per trade on the bid/ask spread of -4pips slippage, no commission cost reflected. One full-size FX contract straight thru, no adding size or pyramiding.
*
It just so happens that GBPUSD tests out best = most profitable in almost every example of systems I wrote. EURUSD and AUDUSD were next, but probably 30% less profitable. Anything related to Yen can be profitable overall, but the drawdown curves are usually gut-wrenching.
*
There are numerous commercial systems for sale that hold up over the course of time... with periods of months to years sometimes unprofitable or b.e. at best.
I'm aware at least one fund manager who earns seven - eight figures annually with managed FX accounts using nothing but commercial-based systems purchased and run verbatim (or slightly tweaked at most) in basket portfolio fashion.
Hey Nitro,
Why do you believe that Simons et al. are using the Ricci curvature tensor? I ask because I consider myself something of a Simons follower/watcher. I've had discussions with 1) funds of funds that were at one point invested with RT; 2) a couple guys who were literally among the first five or six employees at the firm; and 3) a couple of the PhDs who left the firm for greener pastures (if they exist!). And they all swore up and down that there is no geometry involved in what RT does. And we're talking about individuals who really don't have a reason to lie (i.e., they're not protecting the strategy, as they no longer work there or have money invested with them).
polr
Why do you believe that Simons et al. are using the Ricci curvature tensor? I ask because I consider myself something of a Simons follower/watcher. I've had discussions with 1) funds of funds that were at one point invested with RT; 2) a couple guys who were literally among the first five or six employees at the firm; and 3) a couple of the PhDs who left the firm for greener pastures (if they exist!). And they all swore up and down that there is no geometry involved in what RT does. And we're talking about individuals who really don't have a reason to lie (i.e., they're not protecting the strategy, as they no longer work there or have money invested with them).
polr
Austinp:
Thanks for the # 1 and 2.
The cash er and er2 are difficult to cross over, I agree. However, the er2 should be working to follow the cash market except overnight. Your comment is one of experience, but a system on the futures should incorporate daytime cash movement.
Nice work on the 2004 system. Difficult to trade perhaps on a personal account.
Presently, I use the signals on the cash markets as earmarks for entry on futures on a manual "system".
Thanks for the # 1 and 2.
The cash er and er2 are difficult to cross over, I agree. However, the er2 should be working to follow the cash market except overnight. Your comment is one of experience, but a system on the futures should incorporate daytime cash movement.
Nice work on the 2004 system. Difficult to trade perhaps on a personal account.
Presently, I use the signals on the cash markets as earmarks for entry on futures on a manual "system".
Austinp:
Appears you program easylanguage Tradestation. Curious at to what time frame you might use. Are you interested in writing code and how would you determine pricing?
Appears you program easylanguage Tradestation. Curious at to what time frame you might use. Are you interested in writing code and how would you determine pricing?