hi.I have red disscussions on coincollector and I can say a few words because I subscribed for this system.The problem is that those orders are LIMIT,and system is construed such way that its analyses hourly or 40 minutes charts and try to pick up highs and lows.{nothing wrong about it,its correct way to trade}BUT those limit orders very often makes fills impossible as the market just touch the price and reality is that in vast majority of cases you are not going to get filled.C2 ,however,displays those trades as filled and this is the most misleading part of this system.So, I think instead of limit orders they should substitude it with MIT orders.In this case,you are pretty much assured that at least if market touch your price, you going to get filled although ,of course, you should deduct 0.25 points on enter and 0.25 points on exit slippage.At least,you going to be in right direction with the system.Otherwise,if you not going to get filled and market going against you,C2 shows that trade has been filled and its makes money but you ,actually,lose money.So, I think this system make sense,but to get REAL results you have to deduct half a point from each roundturns,but ,inthis case,system is not going to be in the top 5 or 10 systems.
Igor, I have exactly the same problem with another system (Extreme-os). Although I can get the orders in within 10 s of receiving the ITM signal, I don’t get filled on at least half of the orders, and when I do it is at worse prices than C2 lists. Also, I have hit some orders easily that C2 shows as canceled 30-40 minutes later. If you miss an entry order you simply miss the trade. But if you miss an exit limit order (which C2 shows as filled) then you are stuck in the position and make your own decision about exiting (usually at a much worse price than C2 lists, so the trade shows up as a winner on C2 but may be a losing trade in a real account). So I agree with you that many of these systems are so dependent on getting exact limit fills which usually can’t be realized (maybe with Autotrade they can … anybody care to comment on that?). In this case the C2 table results will be far better than a real account could realize as I have experienced so far with Extreme-os (although I am subscribed to Bris and have only missed one trade there out of the last 7, so that is a different story at least with my broker (Scottrade) and the 10 s delay I have between receiving the ITM signal and getting the trade off). I’m not complaining, but it is certainly necessary to “calibrate” a system to your own trading situation before plunking down a lot of money trading every signal from any given system. Bris is working well for me, Extreme-os (so far) is not. I’d love to hear other real world experiences with various systems on this same issue.
I am not sure if Matthew does this - but in my own testing of the systems I design, I do not allow the system to “execute” a limit order unless the price has passed the limit order level by at least 1 tick. That way my results are realistic.
In addition, if, in realtime, price touches but does not break the limit price, I give it 30 seconds and if I have not yet been filled, I go market and take my lumps.
Mauro Grabinfeld - Koan Artist
While AutoTrade is still a work in progress, it has been designed to help you deal with this problem. (There’s no way to eliminate it, since the reality is that if a system trades using limits, you may or may not be filled – there’s no way to know.)
With AutoTrade, you can set a “Convert to market order after XX seconds…” parameter. This is effectively a Market-If-Touched (MIT) order. Let’s take an example. A system vendor issues a BUY at limit 99. If the price trades down to 99, then C2 will decide that all buy orders will be filled at that price and will assume everyone in the world gets that fill. In real life, you might not (depends how liquid the market is, etc.)
Once C2 reports that a trade is “filled” in its hypothetical best-case world, it will tell your TradeBullet to change the order from a a limit 99 to a MARKET order if the limit price is not filled in your brokerage account within X seconds. (X is set by you.) The bigger X is, the more slippage you are likely to get. Thus, you might get filled at 99.25 or worse.
Note that you can specify whether or not to use a “convert to market” separately for futures, stocks, options, and forex.
This feature is currently available in C2 AutoTrade.
MK
It may be impossible to do this automatically, but it would be very interesting to compare a C2 trade results table using the limit prices given by a system (like Extreme-os which trades stocks), to that same table where the average bid-ask spread on the symbol was added to all buy orders and subtracted from all sell orders (ie. treat all the limit orders as MIT orders).
Matt, what would you guess is a typical time delay between when the C2 software gets its fill information (however it does that), and when the ITM window alarms for a new trade (on a typical broadand connection)? I seem to miss (or hit) many of the trades due to timing alone, even if I did not have my 10 s delay between the ITM alarm and the order entry. For example, the one trade today from Extreme-os that shows as canceled on C2, I got filled at actually a better price than the given limit. It appears that various time delays can be at least 10 s or more, which is enough to miss many of these limit orders and of course skew the table results considerably.
Maybe it is only possible to trade the better day trading systems through Autotrade, and it would be very informative if someone who is doing this could comment on their results. I’d move my accounts to IB in an instant if anywhere near the table results could be expected.
Igor;
You have brought up a very relevant subject.
The issues you mention are just the reason for me starting development of own index trading systems some 5 years ago. At that time I subscribed to a self-proclaimed high-performance system which could issue signal e-mails at any time of the day. It was trading about once per week, but the signals issued reflected trades that they actually (from the QQQ-prices they logged in the statistics) had executed some 5-10 minutes before issuing the mail. When receiving the e-mail on the cell-phone, wherever I happened to be, I dialled in my QQQ-position change to Ameritrade and nearly always got quite a different price from what they had in their log.
Their signals seems to be triggered when the market passes through some set limit, whereafter the market normally shoots off in the “wrong” direction before any subscriber could have a chance of getting any order placed. Decent records on their web-site - but impossible to match real life. In particular their break-even situations turned out with significant losses for a subscriber.
Thus, if I ever was to have my own workable system, it had to be simple to manage. It may not produce optimum scores, but at least a follower should be able to match more or less 100% the performance. The solution; All signals are sent out well ahead of market open with a clear message that the trades (both for the user and for the C2 portfolio) shall be made either at the market open or at the market close. Absolutely no doubt or ambiguity and everybody gets the same score. Liquidity is no problem either when QQQQ always is the traded stock.
Some seem to claim that C2 is about simulation. Well - it may be to start with. But as soon as there are subscribers to a system it is no more a simulation game. Then it certainly is real life; real investments for the subscriber and real commercial activity for the system developer.
Christian … do you have a system on C2 that trades QQQQ?
"Some seem to claim that C2 is about simulation. Well - it may be to start with. But as soon as there are subscribers to a system it is no more a simulation game. Then it certainly is real life; real investments for the subscriber and real commercial activity for the system developer."
Well said, Christian! Well said! It most certainly is not a game once people subscribe, and those system delevopers who think otherwise are liable to get hurt.
Randy;
Although not excelling, the “QQQQ-crude” is one. The first 8 months (up to April) in its history represent a predecessor system.
Chris
Please see Autotrade Forum, 'CoinCollector – Cancelled Trades?'
for more info on limit orders filling.
Arie
