Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
Auto trading is only thing that conquered my darkside
I like the idea of having the strategy poll itself every half hour and let you know its status. I am going to implement something on that order.
"The day I became a winning trader was the day it became boring. Daily losses no longer bother me and daily wins no longer excited me. Took years of pain and busting a few accounts before finally got my mind right. I survived the darkness within and now just chillax and let my black box do the work."
1) What is the trading hours? Is this RTH or round the clock?
2) What is your calculated slippage per round trip?
I find that strategies that net around $30-$40/trade are profitable on a few contracts but quickly "saturate" when adding positionsize (due to the slippage).
rpm123 I have never used them so I don't have an opinion.
"The day I became a winning trader was the day it became boring. Daily losses no longer bother me and daily wins no longer excited me. Took years of pain and busting a few accounts before finally got my mind right. I survived the darkness within and now just chillax and let my black box do the work."
I've said it before, and I'll say it again: a truly successful trader will not divulge his approach. The results I posted are a backtest. The result represents 1 contract trading the CL (RTH) with limit orders.
It was my first automated system, but I retired it quickly. The approach is still valid, but now I use a discretionary filter also. It has greatly reduced the number of trades, while keeping the R/R fairly intact. It just makes more sense...
I kind of regret posting it. I was trying to overcome my paranoia, but I see now that I would rather keep it.
The whole point of the post was to show the benefits of dynamic targets and letting winners run.... Win% is not everything... I was never smart enough to be able to program the dynamic entries though, so I guess I will continue to spend my days in front of the computer...
As long as we're on the subject of ATS risk management strategies.....
1) You can now access any windows XP or later computer using XP or Win7 from any computer. This also includes windows7 phones. There's yet to be a 4G win7 phone, but I'm anticipating it out by Christmas. This virtually eliminates the need for leaving your system "unattended."
2) Depending on your broker/trading platform, different strategies will add or reduce 2nd and 3rd order risk. On TS servers, running IOG strategies usually violates their "15 second" rule with respect to order placement. So you either have to use static profit/loss values, OR run the risk that if it violates the 15 second rule, the order will not be maintained on the trade server.
3) Use of limit orders obviously improves price certainty (and reduces slippage). But it comes at a cost with ATS. Limit strategies GREATLY increase the complexity of strategy execution. Partial orders, orders jumped during high velocity surges, etc....can all present challenges.
In essence, I would feel better about leaving a strategy "unattended" if it was running on a dedicated server, with market orders and natively maintained orders on the tradeserver. Obviously that severely limits your strategy development options. Once you start getting more complex than that, the risk increases.....in the event of a positive connection failure, anything but the most simple strategies pose a SERIOUS threat.
When you combine that with the concept that Liquid posted in the other thread about geometric increases in positionsize, the risk becomes even greater.
In the end...ATS is not meant to be "auto pilot." It's meant to be more "cruise control." where you're awake, and monitoring to ensure it doesn't drive you off the road.
Noteworthy is if any of you have ever studied complex risk management concepts....ponder this.
Even the most robust dedicated servers will not "guarantee" more than 99.99% uptime.
That's a 1:10,000 failure rate.
Over the course of 6 months, there are 126 trading days (roughly) at 23.25 hours (electronic) for instruments like CL. That's roughly 10.5M seconds of live trading. at a 1:10k failure rate, that's over 1000 seconds of permitted downtime. Ask yourself what 17 minutes of downtime could do on an instrument like CL. Combine that with the fact that you're data feed has a historical failure rate, your platform server has a historical failure rates, etc......
In essence, unless you're running the most BASIC strategy, running unattended isn't very wise.
I don't mean to belittle your point, which is a very good one, but I'd like to go into it in a bit more detail. I am someone who does leave live strats unattended, at least still in touch with me by text/email. I guess I would never be away from them for more than about 4-8 hours.
What you're describing is the worst case, and of course we should always consider the worst case. So, in your example (and leaving aside data feed, platform etc for now), there are some extra things that need to happen.
The downtime needs to be contiguous - i.e. all 17 mins at once. Something bad needs to happen to your position in those 17 mins. If you have a stop on at the exchange, which I always do for this very reason, you'd have to be very unlucky for something bad to happen.
Most of my strats would do very little during a position. Just trail the stop and look for an exit signal. It's not so bad if that doesn't happen for a short while.
My strats typically detect lost connections and act accordingly.
I agree that if your trading is very short term you might get some problems, but you still need a stop and your trades are likely to be lower total risk anyway.
You also mention that unattended isn't wise unless strat is very basic. Not sure I completely agree. Most complex strats I run have their complexity in spotting a setup, and managing a stop. As I said earlier, it's not a huge risk for me if they can't manage a position for 17 mins.
In summary, it's about knowing the worst case, the typical case and deciding whether you accept the risk.
I'll leave you with this. Let's say I have one of these events every six months, and it's a bad one that loses me 200 points. For the rest of those six months, my strats trade while I'm asleep or not working, from UK 11pm to 7am, and 6pm to 10pm. That's a lot of hours over six months, and my strats wouldn't be trading then if I couldn't accept the unattended risk. So, they need to make more than 150 points in all those hours to make the unattended risk worthwhile. Well, if they don't do that easily they're not worth running unattended. Actually the same argument would apply if you spent each afternoon at the beach. You're paying a risk cost to be there.
Risk tolerance is a personal choice. One woman might be a beauty queen to me and might look like the lunch lady to you.
I already said, TS does not maintain stops that are updated more than 15 seconds apart. That means, if you're using anything not static (like a trailing stop, etc) then you run the risk of having an open position with no stop in the event of a connection failure.
Secondly, you don't need 17 minutes of downtime to see a catastrophic move. Depending on your instrument, this risk can vary.
I've PERSONALLY observed CL move 50 ticks in 5 seconds. I've personally watched it move 200 ticks in one direction in 60 seconds and then watch it reverse back the other direction 150 ticks in the next 60 seconds.
I don't know what you're trading setups involve, but mine don't involve 150 tick stops.
As I said, if you're running a different tradeserver/broker, different data feed, then these risks could vary greatly.
You don't have to lose internet access in order to lose connection to the tradeserver (or the data stream). So when you combine those things....and you're trading on a fairly leveraged instrument (like CL or GC), then these risks are substantial.
In the Army, when we did risk analysis, you use a 2 dimensional relationship between frequency (likelihood, expectancy, etc) and outcome.
I have a background in nuclear engineering, and the risk analysis is similar.
Even if something is VERY unlikely to happen, but if it does, it's catastrophic, it's generally considered "high risk." If a nuclear power plant has risk management controls that predict a catastrophic incident every 1000 years of operation, you'd say that the risk is small yes? But if/when it happens (a confluence of failures) the result is completely unacceptible....then the risk still gets labeled "high."
If you're conducting an exercise (like jumping out of an aircraft with a parachute) and you know that the total incident rate is very small (1:10,000) but when there IS an incident, it's generally considered loss of life, limb, eyesight, aircraft, etc, then it's still an elevated risk.
Now, if I did indeed have stops maintained on the tradeserver, then the risk decreases dramatically, as the incident rate is extremely low/unlikely and the outcome/effect is minimal......
I would encourage everyone to verify and reverify that their broker/tradeserver actually maintains stop orders and exits in the event of a failure.
I guess my background makes me plan to the extreme and a bit paranoid. I have a dedicated server, I have 2 computers up and running (with internet access). In the event I lose my cable connection, I have an aircard backup available. I have my account number and the tradedesk number written on the whiteboard above my desk (to reduce time accessing the tradedesk). I have all that information on speeddial on both my cell phone and landline and I have my account information available on my cell phone as well.
The final phase of risk management is mitigation....i.e. what steps have you taken in order to either reduce the likelihood of an incident or reduce/minimize it's effects.
For me, running anything other than static stop orders presents a significant risk. But I cannot be tied down to simple bracket orders. OR, I simply have to ensure that I minimize the "unattended" time.
Ah, my broker keeps stops at the exchange until cancelled. I agree, risk is too high otherwise. I don't trade CL automated because of the volatility you describe.
Yes, risk is a personal choice, but I still think it's an important point to balance against the potential for profit while unattended. If I was trading CL automated, I'd expect to be making a lot of ticks in normal times to cover the unattended risk. Even at a few hundred points lost. You see, this is not a simple case of mitigating risk by not trading unattended, but balancing the risk/reward of trading unattended. I simply couldn't trade during my night, so if my strats do and I make 300 points a month from it for each 500 point catastrophe every six months then it's really not a big deal riskwise.
Agreed. Different instruments present challenges and aspects that shape viable strategies.
CL for instance....is usually MUCH more tame during off hours. I find that the optimal time for discretionary trading (with a smallish account) happens to be during the OpenEurope to OpenAmerican period. There's enough volume to move it, but it's not the wild west that it becomes during RTH.
Having said that, I gave up on trading off hours....the slippage is way too unpredictable. You can see several ticks of slippage on market orders and as I said earlier, I've tried automating with limit orders.....it's difficult with anything more than a single share/contract. Partial orders become rather difficult to address.
So when you overlay those aspects....most CL traders trade from about 8 a.m. EST to about 2-3p.m EST. Which basically lends itself to "attended" trading. Trading off hours is really a viable solution if you're only trading 1 or 2 contracts OR you're going after larger moves. Chasing 10 ticks with 2 contracts off hours you might see most or all of your move soaked up in slippage. That, and even though it "wonders" aimlessly sometimes.....it also presents a set of risks in terms of sudden volume spikes....(essentially, the trend during off hours even long term, isn't always reliable....you get professionals bleeding off position and taking profits, amateurs dabbling, etc....and then all the sudden, someone comes in with a hammer and ruins your day with a large move).
I really wish TradeStation would maintain stop orders....the more and more I go down this road, I regret not going with a different language/platform, like Ninja. It would be a little more difficult, but the capabilities seem much greater.
as it were, If I'm going to babysit my strategy all the time, then I limit my trading hours to when that's feasible....which is day job hours.
