Hi Everyone,
Naturally I am now curious to hear about your good and bad experiences using Collective2, as vendors and as customers. I think I found maybe two or three profitable strats in the population I sorted and the rest were scary at best.
The following 2 users say Thank You to Dragon for this post:
for my first idea, c2 is a kind of a playground for sys-freaks + sys-developer-beginner + trader who think about playing bigger one day.
the one who are succesful with their stuff playing there a bit move forward to some more seriously / bigger places and/or create their own funds / managed accounts.
but nothing to say against such a not too expencive place as a playing / testing-area to make open trackrecords + stuff to see how all that works.
max-td
The following user says Thank You to max-td for this post:
Interesting. Thanks Max. You mean there are better places than C2? Do you have links to other sites? Or do you just mean those that create their own funds?
The following user says Thank You to Dragon for this post:
At C2, I sorted the group by the guys who don't have over a 3% drawdown and had over 1000 trades. Those guys had some pretty interesting things going on. I then added a sharpe ratio greater than 1 (I looked this up on Wiki and I don't fully understand its relevance). It seemed there were two guys on there if I remember that were the only guys that didn't have more than 2 losing days. They do about $100 a day if I remember correctly. I am soo ok with that. But I don't know if it's legit.
Overall, as a developer, I think C2 does more things good than bad given the lack of alternatives.
My biggest gripe is in the lag between the timestamp--and price--of real trades and what NT reports to C2. Today for example, there was a trade that had a 9 tick profit and C2 reports it as 5. Now this can cut both ways and work in my favor if the trend continues after excecution and reporting to C2, but I wonder when it's all said and done if the net effect hurts the system's real from reported results.
Do you think C2 should report the actual price of the trade or at market at the time it receives the "confirmation" from NT?
Site Administrator Swing Trader Data Scientist & DevOps
Manta, Ecuador
Experience: Advanced
Platform: Custom solution
Trading: Futures & Crypto
Posts: 50,002 since Jun 2009
Thanks: 32,468 given,
98,269
received
That's just because of the poor way NT handles it.
If you write your strategy properly (I posted code on how to do it in EasyLanguage, C# would not be a lot harder) then you will send buy stop and sell stop orders ahead of time, with limit fills and OCO orders, etc. This is assuming your strategy can make a decision ahead of time, which I think most can. All you need is an extra 5-10 seconds so there is no slippage at C2.
The issue with NT's built-in support for C2 is there is no limit orders, no buy/sell stop, no OCO. Everything is market, and as such there is going to be a delay for sending that to C2 and then C2 adds an additional couple second delay to provide for a 'realism' of the end-user getting fills.
Site Administrator Swing Trader Data Scientist & DevOps
Manta, Ecuador
Experience: Advanced
Platform: Custom solution
Trading: Futures & Crypto
Posts: 50,002 since Jun 2009
Thanks: 32,468 given,
98,269
received
I am saying don't use NT's built-in C2 support. Just code C2 support directly into your strategy, bypass NT's handling of it. This gives you all the control and flexibility you need.
Another trade fired and this time it was 1 tick positive slippage. I wouldn't be so concerned if the strategy took oversized gains, but for a scalper (win 9, lose 5), C2 may not be suitable if they take a cut on both sides.
To my knowledge, C2 has no mechanism to throttle the number of contracts a subscriber trades in a system. If the strategy instrument is thinly traded, a large group of orders coming to the market at the same time could negatively impact results. Limiting the number of subscribers does not solve the problem.
If you start to think about this kind of potential issues, you should find something smarter than C2.
A machine hosting the system, which knows who are its clients, and knows the overall allowed quantity per signal.
Then this machine act like a scheduler, sending signal for x qty to client 1, y to client 2.
With NT on both sides, a not so hard to code client-server application.
Sam, a few brokers are addressing this issue by creating subscriber payment models on a per contract charge and indeed restricting those who try to game the system. I think Attain is one of them.
Just want to let the futures.io (formerly BMT) community know I have put out feelers for this project with for hire NinjaScript consultants from the NT website and have yet to receive a proposal.
What are you going to use, for orders, Stop orders and/or Limit orders ?
The difficulty will be how to manage the unfilled orders in C2 (they have to be cancelled after something is triggered, opposite signal, or after x minutes, or another trigger of your choice), and how synchronize your NT position with the C2 position (partial fill for instance).
And being able to read the xml sent back by C2.
The other difficulty, is the testing process, as C2 allows only 3 signals (from mémory, not sure), then after you'll have to pay for a testing. And just 3 bullets for this looks not enough.
A bit tricky, fun, but not so easy.
Success requires no deodorant! (Sun Tzu)
The following user says Thank You to sam028 for this post:
I've played a bit with the idea, deeper tests are needed (we are limited to 5 signals per system, not 3, I was wrong), but it works.
And a bit nicer than Mike's "old school" method, all is under in the strategy itself, no wget or response file to write.
But Mike has maybe spent 2 hours on it, and I spent much more time on it .
No so good in fact, for trade #3:
- NT: buy 10117, sold 10128 (+11)
- C2: buy 10117, sold 10104 (-13)
10128 was the high of the candle, so maybe it will not be filled in real, ok, why not.
But sold 10104, it's impossible, as we had 10104 only after the end of trade #4.
The cancelled time, on C2, doesn't make sense for trades #3 #4 and #5.
Nor the "Recently Closed Trades"...
Nothing magic.
The signal happens, the strategy executes a SetStopLoss(), SetProfitTarget() and EnterLongLimit() or EnterShortLimit(), then build the correct URL, and sent it to the C2 website. Just like what Mike has done for MC, it's just a bit faster here because all is made in the strategy, no timers and external .exe needed to send the request to the C2 website.
Hi Zoethecus
You are absolutely right.
This is exactly what happened to me back in 2008. And limiting the number of subscribers won't solve the problem (e.g. if one of them is a broker).
is there a system or would people be interested in a system that allowes you to follow in real time and trade a live account or choose to be in sim mode while the main account trades live - you will have a choice to trade the same quantity or a % of it -
?
Im actually trying to put my system on collective 2 right now.... There's alot of things that i don't get. For example, if i understood it correctly, if i enter a limit order and i get filled, it automatically get turned into a market order for the people that are following my signals. Now most of the times thats ok, but there are times i get filled on my limit orders since its been sitting there for a while. That could have adverse affects if your trading the ES, since your market order might get a bad fill, and your results would be different. Also, half the trades i did today didn't post for some reason. I don't understand why, and when i reported it, they didn't respond back to me yet....
I heard of something similar to collective2 called covestor. I'll have to check it out later on, still trying to get collective2 to work properly. I don't think they have the same connectivity to ninja like collective2 has, and im not even sure you can day trade futures on there. Will have to do a bit of more research.
The following user says Thank You to Michael.H for this post:
We have decided that the best thing is to execute is in-house in our own brokerage, as we find 2nd party execution a bit cumbersome, at time expansive and does not allow us to see the real picture from a risk management perspective.
Our procedure:
1) All systems are executed in house on a server. Most of the time the developer delivers the code protected.
We sign all non-disclosure agreements.
2) We test the system up to 3 months with forward testing, to see if the hypothetical match the signals. We have learned the hard way that with Range bar you can some incredible results while the reality is far from it.
3) All signals have to run on TradeStation, because after all is said and done we have found it is the best platform to generate the signals from.
4) Execution is done via NinjaTrader/Rithmic or CQG/TradeMaven.
5) All orders are executed as MIT on TS then sent as market orders.
As far as using 2rd party system executions:
Sadly some developers are not on top of their system execution. They can shoot an order that might not hit the server and no execution while a liquidating position might get triggered. It is very frustrating to have developer of this kind because you need to start closing positions manually. While customers assume that orders get executed right, this is not always the case. Technology is not perfect. When we have a system that shoots and no execution is done due to data failure or anything else, we know what is the order is and we can act accordingly and in a timely fashion.
The cost of execution in house is small as compare to the execution of systems with C2 and Strategy Runner. I don’t have to pay expansive routing fees.
Consider:
C2 wants $1.99 per side that is besides broker commissions. They say 1.99 on their site and I assume it's per side.
Strategy Runner wants $0.8 RT, which is very reasonable, and I would choose it over C2 any day. They do have a $75/Month minimum. But, They are very negotiable when it comes to volume.
Also, they do have a good real time platform to watch positions.
Strategy Runner has a number of systems developers that are truly on top of things, and watch their trades and see if execution was done right. I personally reached out to them and discussed execution.
Back to in-house execution…
In-house execution allows us to execute without paying on a per contract basis. Sometimes the developers make a lot more than the broker that way, and that should not be the case. I feel brokers and customers take more risk, not the developers.
Above all, if we have a system with a lot of contracts, we can say enough!
Systems can only carry that many contracts, and if someone cares about the integrity of performance, then at some point they close it. At the very least, they have slight modifications for entry in terms of time so not all contracts are just thrown into the market. Slippage control is the key to performance.
The system execution technology is not perfect, but with the release of MultiCharts and the ability to run easy language and execute it on one machine using a low latency solution could solve lots of headaches for us as well. All will be done on one machine. With all the system execution difficulties, I am still going to execute most of my business in house, because that is the best risk management you can do and so far proven to be the best.
I hope I have brought to light things that would help people make the right decision when it comes to systems selection and execution.
Trading futures and options involves substantial risk of loss and is not suitable for all investors. Past performance is not necessarily indicative of future results. You may lose more than your initial investment. All posts are opinions and do not claim to be facts. Please conduct your own due diligence. Use only Risk capital when trading Futures.
1 800 771 6748 local 561 367 8686 email support@OptimusFutures.com
The following 3 users say Thank You to mattz for this post:
we typically wait for the results as they accumulate for 3 months.
However, we use our discretion for less time, depending on the experience of the programmer, etc.
Most of the systems in house are purely mechanical.
There is another solution we use for discretionary trading, and that is SAS by Strategy Runner.
As a broker I will see your trades on CQG Trader and will be able to confirm as trades are allocated to accounts.
I have attached here a link to the brochure: https://www.strategyrunner.com/Content/Documents/SAS_UserGuide.pdf
Trading futures and options involves substantial risk of loss and is not suitable for all investors. Past performance is not necessarily indicative of future results. You may lose more than your initial investment. All posts are opinions and do not claim to be facts. Please conduct your own due diligence. Use only Risk capital when trading Futures.
1 800 771 6748 local 561 367 8686 email support@OptimusFutures.com
its almost one o'clock on East Coast...even the Dog is sleeping, and I am still on futures.io (formerly BMT)....I need to get a life.
M
Trading futures and options involves substantial risk of loss and is not suitable for all investors. Past performance is not necessarily indicative of future results. You may lose more than your initial investment. All posts are opinions and do not claim to be facts. Please conduct your own due diligence. Use only Risk capital when trading Futures.
1 800 771 6748 local 561 367 8686 email support@OptimusFutures.com
The following user says Thank You to mattz for this post:
Maybe they stopped? I will help you with your system if you want to offer it to the public.
Trading futures and options involves substantial risk of loss and is not suitable for all investors. Past performance is not necessarily indicative of future results. You may lose more than your initial investment. All posts are opinions and do not claim to be facts. Please conduct your own due diligence. Use only Risk capital when trading Futures.
1 800 771 6748 local 561 367 8686 email support@OptimusFutures.com
Trading futures and options involves substantial risk of loss and is not suitable for all investors. Past performance is not necessarily indicative of future results. You may lose more than your initial investment. All posts are opinions and do not claim to be facts. Please conduct your own due diligence. Use only Risk capital when trading Futures.
1 800 771 6748 local 561 367 8686 email support@OptimusFutures.com
I linked my trading to C2 for a week, and the only thing I can say is: it doesn't work.
No one trade was filled at my price, the majority of trades were filled at completely different prices. Many loss trades where profitable on C2 and vice-versa. Some trades were left open on C2, even if I closed it, and others were rejected with no valid reason.
And this was only from me to C2, I can't imagine what mess my subscribers would have received.
Reliability is zero.
To be honest, I tried also other automation services (tradency, zipsignals, rentasignal, zulutrade) and no one was reliable enough to trust.
The following 3 users say Thank You to jagui for this post:
The only good solution, for a good execution on the subscribers machine, between the machine which is sending the order, and your subscribers machine.
A good old tcp client/server can do the job, and the whole process won't take more than 200 or 300 ms, vs 5 or 10 s for C2.
C2 is good enough for swings, but not for sclaps.
I send orders from NT to C2 and from time to time there is a small discrepancy because of the lag from NT to C2, but it has worked as much in my favor as against me. In other words, I think it's a wash and not much different than a subscriber should expect from taking a trade from a third party signal.
My main complaint with C2 is the cut they take (30%) from the developer--it is too high in light of the service provided. A a result, I am in discussions with Stryker and World Cup.
Im gonna have to agree with jagui and others about collective2... I feel bad for people following it. Even if the system is profitable, C2 unfortunately has too many problems.
Its not just 30%, the ask subscribers to pay an extra $4 per r/t..... thats asinine. i start scaling at 1 point.... since they don't accept limit orders, subscribers will get 1 tick worst price than me because i trade the ES, since it goes market when i get filled, thats $12.5... Then, they take an extra $4 per r/t...
I make $47 after commission( i have a good commission schedule), subscribers make around $30 after commission and cost( assuming they pay only $4 commission to their own broker).... so essentially, anyone following me will under perform by about 37-40%
You guys are probably gonna think this is really funny. I sent an email to their tech support as to why some of the orders were not properly being sent to collective2 via ninjatrader ...ie entry was long, never got sent. Exit was 2 points above, that got filled on collective2. Therefore, I made 2 points, however, the poeple that would have been following me just got short..lol
I never got a reply, until today.... A MONTH LATER....lol
I would stay away from them IMO..
The following 4 users say Thank You to Michael.H for this post:
I used to have a system on Collective 2 and had the same execution problems that you describe. I am a scalper and was looking to document my track record. The support I received from Ninjatrader and C2 were very weak. I think the ONLY way to have a system on C2 if you use Ninjatrader is to manually put the orders directly into C2, which is cumbersome. Their system does go down from time to time also.... The subscribers would have been put through hell if I would have allowed subscribers, which, I wisely didn't until I fixed the problem. However, no one could fix the problem and I closed the system.
However if you do it directly onto the C2 site, you can verify if the order went through and was, indeed, executed. Otherwise, it is a nightmare, in my opinion.
Just my 2 cents,
ZENTRADER
The following user says Thank You to ZENDINA for this post:
Thank you for sharing your deep experience with the C2. True their system is quite problematic. I tried them on demo and what struck me was that the providers available (and not all of them) were there for only 7 days.
I have a system on C2, to supplement my trading with subscriber revenue. I can tell you, I am definitely rethinking my decision based on my experiences since i've been on the site in the last several months. Today was likely the nail on the coffin when C2 autotrade service was down for a limited time, but nonetheless had me fiddling with my VPS and such for hours. Apparently they made a conscious decision to take their database down to do some required maintenance, I would let that pass if it wasn't that THEY DIDNT TELL THEIR CUSTOMERS. The only mention of it was a forum post after the fact.
Arrived here researching Collective2 after watching a K. Davey webinar this evening.
Reminds me of Interactive Brokers (scheduled TWS shutdown and reboot of servers in the wee hours every morning) and HitchHiker's Guide to the Galaxy (Vogon explanation for destruction of Earth without apparent warning).
The common thread among IB, TWS, Hitchhiker and "annoying" is provision of service. Up to and including the destruction of the planet in the case of Hitchhiker.
I like Douglas Adams' ironic sense of humour (he wrote HitchHiker's Guide) and his other stories, and am thoroughly satisfied with IB even if TWS' need to shut down every 24 hours (avoidable by use of IB Gateway) and IB's need to reboot their servers at midnight EST drives me nuts because it potentially stresses my own homemade strategies that run continuously.
I enjoyed your webinar tonight enough to consider running the 3 pair Forex strategy since glimmers of the "expert follower" still exist in me .
Since it's served by a third party sought to research the 3rd party, the only unknown in the equation. I was interested in @gonzofist's remarks as another service provider (the apparent lack of warning in his post the trigger for my remarks, since even though scheduled I consider the interruption in IB service I've learned to live with an annoyance), wondering if you have any complaints about Collective2 as a platform.
I've been a member at Collective2 since 2005, nearly the beginning. The service has certainly come a long way since then. It still has issues from time to time, and sometimes accessing the website is annoyingly slow. Once in a while, I could not get in at all. But that has not happened recently.
I auto trade one system on there (my own), and in 7 months or so of doing so, there was only one time when my system went out of sync (due to a holiday open, where Tradestation forex opened a few hours before my C2 forex broker FXCM opened). Other than that, it has worked pretty well. I've never noticed significant delays between when I send an order, and when C2 broker fills it. But, I haven't checked every fill to see if this is the case.
Notably, with that particular system, I do simple market orders in, and market orders out. Very simple. Trades last days, so that reduces the opportunities for error. I'd be a little more scared running a high frequency strat there.
I have a couple other systems there that use stop orders, etc. They have more issues, but that is from developer to Collective2, not from Collective2 to end user broker.
All things considered, I think Collective2 does a good, but not perfect, job for the market it is in. I've used other signal provider services, and I find them comparable on number of issues, performance, etc.
I hope this helps!
P.S. I do recall the recent planned "unplanned" outage, and people complaining after the fact, which they rightly deserved to. It did not impact me. Years ago, I had a case where I had to hold an ES position over the weekend, because I could not get into their website. So, yes, these things do happen, and if the trading strategy was high frequency or relied on precisely timed entries/exits, your performance might be negatively impacted.
The following 2 users say Thank You to kevinkdog for this post:
It does help, even if after I posted my reply to you it dawned on me given its trading frequency the 3 pair strategy likely does not require continuous connectivity. Even the bank I trust to hold most of my assets seems to have outages occasionally the moment (every few months) I want to access it.
My issue in the prior post had not been with frequency of service disruptions, but their process of handling them or rather lack of process. It gave me insight into how the company functions and the worrying disconnect between entities within the organization.
This and several other subjective reasons led me to leave the site entirely. The standard needs to be very high for sites that have access to not only your broker logins but also your credit card data, which is retained on their servers.
A little more research will lead you as it did me to more objective reasons not to trust this site with your financial future. In 2009 their database was breached and apparently logins,cc data were compromised Internet trading site collective2.com hacked | Office of Inadequate Security the e-mails users received (I was not a member then) were not explicit whether the data obtained was in plain text or hashed.
While no business is immune from a breach, I have come to expect a level of redundant security from other entities that hold this same level of information brokers,banks etc. that simply was not implemented before or after the breach. Your free to your own interpretation.
I get the feeling there are actually very few vendors and subscribers there. Maybe I am wrong about the subscribers, but good systems worthy of following are sure in short supply.
Now here is a moral question: Let's say I have found a good system with low fees and DD and decent returns, would I tell it to anyone? Probably not....
The following user says Thank You to Pedro40 for this post:
Sorry for the delay, I had to look up "necromancy."
Anyhow I still do use Collective2. Like any of the signal services, though, you do have to be careful with it. There are good looking systems out there that are actually terrible, and vice versa. It can be hard to separate wheat from chaff...
The following user says Thank You to kevinkdog for this post: