Favorite Futures: Futures - bonds, currencies, index
Posts: 288 since Oct 2010
Thanks: 70 given,
Agreed. I generally use limit orders for that reason. Ninja has an unmanaged orders mode (which simply means Ninja does less behind the scenes and you can program all the raw orders yourself.) It's not as bad as it sounds, but it meant that I can handle partial fills and still maintain stops etc. It was a bit of work, but it's all done now and in a separate library. So, I'm happy relying on limit during the night (and for a bit of added safety, I usually check the liquidity ,spread and reduce the contracts if necessary)
This means I often wake up with a position in something like AUD, JPY, EUR, or the bund. It also means I get possibly 15-22 hours in a position for what is effectively still a day trade. This makes it more viable to trend day trade on a 30, 60 or 120 min bar, for example, which reduces noise/volatility risk.
I trade with Interactive Brokers, and they keep stop/target orders on their servers or at the exchange, but very few brokers do that. For everyone else, your stop/target will be managed by your software on your pc. So, a power outage, internet disconnection, your pc running slow, freezing, crashing, etc., all of these leave your vulnerable. Even if you have code that detects lost connections, during that time, there's nothing your strategy can do other then notify you, if the internet connection was not the culprit.
For me the important phrase of the majority of the entries is: "OCO (One Cancels Other) functionality is simulated on your local PC", this is where you are going to have problems if you loose connection to your pc. Only Interactive Brokers, and TD Ameritrade say: "OCO (One Cancels Other) functionality is natively supported on their servers." With IB, the moment my stop/target bracket order is sent, and I get no errors back, I can leave the trade unattended, the worst that can happen is that my stop loss is hit, with no other unintentional behaviors, whether my local pc is on or not.
The following user says Thank You to monpere for this post:
Well, this is why I think you have to have a particular type of development philosophy when automated trading. You need to work with your software and broker and understand and work around all the things that can go wrong. For example, I don't use OCO very much, since running Ninja in unmanaged mode it's not to difficult to check orders statuses and act accordingly. But when I do, for things like bracketing, I don't put on the stop until I receive conf that the OCO bracket is cancelled. If the connection goes down and OCO doesn't work, I know that the bracket will act as a decent stop and the stop doesn't go on. That's what I mean by working with what you've got. It's not perfect, but it means you're happy you've covered as much as you can, and you're not totally relying on third party software to sort things out.
I think there is a fairly reasonable fear of what happens with your positions when your connection goes down. Since my strats have been designed with this in mind, and I always get notified about it, and can call my broker and check positions and tell them what to leave or close, I don't let this constantly worry me.
That's the side I haven't seen. All I see is all the functionality I lose with TradeStation (as compared to the Ninja guys) and it leaves me envious. I have skimmed enough threads to get the sense that Ninja has it's drawbacks in reliability though.
I've found NT7 to be much more stable then NT6.5. Where I use to get daily crashes and hangs in NT6.5, in NT7 this occurs less then once a month now. Hopefully it will get continually better with future releases. My main draw to NT is because of the power of C#, and the fact I was able to user the platform for free while developing and proofing my concepts. What platform do you use for automation?
To veer back toward the original thread intent.....
For all you discretionary guys....
How many times have you said to yourself..."Man I wish I could have gotten in earlier" or "I sure wish I could have gotten out when the signals told me to."
My experience with discretionary was that by the time I had checked, rechecked, TRIPLE checked to ensure my entry criteria were met, the opportunity was either diminished or gone altogether.
The same was true for anything other than basic bracket exits. If I had exit signals based off market action or anything other than a simple OCO.....I always ended up saying....man I wish I'd have gotten out sooner, I'd have kept more profit (or reduced my loss).
If you multiply that concept by the incident rate, times the number of trades, the concept quickly adds up.