You trade futures only? Us? Globex / CME?
That would be my markets.
That would be my markets.
Strategy(ies) exchange
- Thread starter dom993
- Start date
Despite the lack of apparent interest in this offer, here is an update of where I stand:
- Validated both strategies' real-time behavior using MarketReplay (re-read this ... it is the real-time behavior that I validated, not the performance of the strategies! Ninja's fill engine for Sim101 & MarketReplay is giving so wrong fills it would be plain laughable if it wasn't such a pain then to validate anything by comparing trade logs between Backtest & MarketReplay).
- Moved to real-time execution on my IB sim account for both strategies (which are running live in my real account on Ensign). I plan to carefully monitor the order management for a few weeks (the only thing I couldn't test at all until then), already hit a serious issue with Ninja
... hopefully they will be able to fix it.
- I am working on an "extended hours" version of the DF20 strategy (it currently only trades 9am to 12pm, a bit more on Mondays & Fridays). That "extended hours" version shows about +25% P&L in backtesting, quite consistently across the almost 3 years I use for backtesting. See attached.
- Validated both strategies' real-time behavior using MarketReplay (re-read this ... it is the real-time behavior that I validated, not the performance of the strategies! Ninja's fill engine for Sim101 & MarketReplay is giving so wrong fills it would be plain laughable if it wasn't such a pain then to validate anything by comparing trade logs between Backtest & MarketReplay).
- Moved to real-time execution on my IB sim account for both strategies (which are running live in my real account on Ensign). I plan to carefully monitor the order management for a few weeks (the only thing I couldn't test at all until then), already hit a serious issue with Ninja
... hopefully they will be able to fix it.- I am working on an "extended hours" version of the DF20 strategy (it currently only trades 9am to 12pm, a bit more on Mondays & Fridays). That "extended hours" version shows about +25% P&L in backtesting, quite consistently across the almost 3 years I use for backtesting. See attached.
Another update (already 5 weeks since last one ?!) :
- I developed a fairly large workaround for the NinjaTrader issue with "native" IB OCOs ... since I am only trading 1 contract per target at this time, I decided to cut a corner at this stage and constrain that workaround for 1 contract per OCO (it is always possible to run multiple instances of a strategy if more contracts per target are required). Did very extensive testing of all sorts of order-feed errors, and I am very pleased with the results.
- While I was at it, I addressed a number of other failure scenarios: UPS Suspend, System Shutdown, Internet-service interrupt. I am pretty proud of the behavior on Internet-loss followed by reconnection (any strategy w/ an open trade is left alone, all open entries are cancelled, each chart is reloaded (data) then its strategy(ies) restarted, and I take care not to reload data for a chart/strategy if that would impact another strategy w/ open trades on same instrument - on same or different chart).
In the process, I realized Ninja's driver for Kinetick doesn't work properly through most Internet deconnection / reconnection cycles ... luckily, using the IQfeed driver works just fine (Kinetick is just a rebrand of IQfeed for Ninja).
- I also fixed a bunch of memory-leaks (ha! there isn't such thing in C# thanks to GC aka Garbage-Collector, right? wrong), which means I don't have to restart Ninja every-time before doing a backtest.
- For the reversal (DF20) strategy I exposed all filters parameters, and made sure the Time-of-Day filters do work regardless of the timezone in which Ninja is run (these parameters are expressed in Nymex local time, ie. EST).
I finally decided this week I was ready to switchover my live trading from Ensign to Ninja, and I have been running both strategies on Ninja in my live accounts (2 accounts, 1 per strategy) since Wednesday (I continue monitoring for any real-time discrepancy with Ensign).
My idea of "exchanging" live & profitable strategies still hold (see 1st post of this thread) ... PM me if you are interested.
- I developed a fairly large workaround for the NinjaTrader issue with "native" IB OCOs ... since I am only trading 1 contract per target at this time, I decided to cut a corner at this stage and constrain that workaround for 1 contract per OCO (it is always possible to run multiple instances of a strategy if more contracts per target are required). Did very extensive testing of all sorts of order-feed errors, and I am very pleased with the results.
- While I was at it, I addressed a number of other failure scenarios: UPS Suspend, System Shutdown, Internet-service interrupt. I am pretty proud of the behavior on Internet-loss followed by reconnection (any strategy w/ an open trade is left alone, all open entries are cancelled, each chart is reloaded (data) then its strategy(ies) restarted, and I take care not to reload data for a chart/strategy if that would impact another strategy w/ open trades on same instrument - on same or different chart).
In the process, I realized Ninja's driver for Kinetick doesn't work properly through most Internet deconnection / reconnection cycles ... luckily, using the IQfeed driver works just fine (Kinetick is just a rebrand of IQfeed for Ninja).
- I also fixed a bunch of memory-leaks (ha! there isn't such thing in C# thanks to GC aka Garbage-Collector, right? wrong), which means I don't have to restart Ninja every-time before doing a backtest.
- For the reversal (DF20) strategy I exposed all filters parameters, and made sure the Time-of-Day filters do work regardless of the timezone in which Ninja is run (these parameters are expressed in Nymex local time, ie. EST).
I finally decided this week I was ready to switchover my live trading from Ensign to Ninja, and I have been running both strategies on Ninja in my live accounts (2 accounts, 1 per strategy) since Wednesday (I continue monitoring for any real-time discrepancy with Ensign).
My idea of "exchanging" live & profitable strategies still hold (see 1st post of this thread) ... PM me if you are interested.
I though I would give an update on the performance of the reversal system ... I spent a few weeks in August & early September analyzing the behavior & performance of that system from January-2007 to October-2009, and came-up with a small number of improvements.
The performance over that period of time remains less than stellar, but given the nature of CL back then (the bubble up followed by the crash down), I consider it a reasonable test.
The attached performance summary & P&L graph cover Jan-2007 to Oct-2012.
I am happy the system closed October on a new P&L peak - it has certainly not performed as good in the last 6 months than in the prior years, but at least it is grinding upwards.
Anyone interested in the idea of a strategy exchange, PM me.
The performance over that period of time remains less than stellar, but given the nature of CL back then (the bubble up followed by the crash down), I consider it a reasonable test.
The attached performance summary & P&L graph cover Jan-2007 to Oct-2012.
I am happy the system closed October on a new P&L peak - it has certainly not performed as good in the last 6 months than in the prior years, but at least it is grinding upwards.
Anyone interested in the idea of a strategy exchange, PM me.
Hello,
I can see that people are having a hard time sharing profitable strategies. Maybe a better approach is to co-develop a few ideas, narrow down and develop just a few of them. This way no one looses.
I also believe in running multiple, if possible, uncorrelated strategies with different logic , instruments and style each. I am currently running 2 and I am starting up my third soon. They are all stock- EOD, though.
I am looking to diversify into intra-day strategies but the testing and set-up involved is much tougher than EOD. I would be forced to run in the cloud as I live outside the US.
If throwing around ideas is interesting let me know.
btw, running 2 strat on one commodity sounds quite risky to me...
I can see that people are having a hard time sharing profitable strategies. Maybe a better approach is to co-develop a few ideas, narrow down and develop just a few of them. This way no one looses.
I also believe in running multiple, if possible, uncorrelated strategies with different logic , instruments and style each. I am currently running 2 and I am starting up my third soon. They are all stock- EOD, though.
I am looking to diversify into intra-day strategies but the testing and set-up involved is much tougher than EOD. I would be forced to run in the cloud as I live outside the US.
If throwing around ideas is interesting let me know.
btw, running 2 strat on one commodity sounds quite risky to me...