This may be the dumbest thing for 2006, but...

Discussion in 'Order Execution' started by davidrousseau, Jan 12, 2006.

  1. I set a stop 4 points down on the Russell 2000, and if it hit 66, it would close out 60 small cap positions. Day's low was 68. I got up at 8 and all my position had been sold. I looked at a very odd tick on the graph. Other than blatent stop running, it seems like a fair and orderly market would not allow a quick 4 point dip. A few seconds later, it was back to 68.

    Anyone know the best way to investigate this ?
  2. I remember something very weird happened in the futures market a couple days ago in the very early morning.

    I had a bid to buy at a very low price and somehow it got filled overnight.

    Charts showed a huge spike down that only lasted a minute.

    Don't have any idea what happened. Was probably a fat finger incident.:p
  3. What platform does that? I mean close 60 positions based on 1 entry?

  4. Cybertrader alerts allow you two different ways to close all positions based on any number of variables. I had set it via an alert that close all orders so that all positions were closed if the IWN hit what I believe was 68 when it was trading at 70. The following morning, the days range said it never hit that price. I usually set the alerts so that they take effect at 9:30AM and end at 4:00 PM, but often start an alart at 9:45 to avoid running stops right at the open. In this case, the tick was either 9:45 or 10:45. I have to go back and look at te chart as well as some other records.

    Check out Cybertrader alart templates. They work well, but not when specialists move the price several points, and I don't see how they can do this, but don't dispute that it happens. I just want them to prove how they came up with the price.

    What I am getting at, is if a specialist created a keying error, and the market is presumed to be fair and orderly, did they violate a rule ? Let's say you are holding 500 shares of Google, with a stop at 400, and the accidentally key in the price (keying ?) of 44.6, and sell all your shares at a 400 point loss, who holds them accountable for this accident ? In my case, the loss was less than 1000, but the price did not show on Inet time and sales, and was not shown in the days range.

    Date Open High Low Close Volume Adj Close*
    12-Jan-06 69.28 69.53 68.91 69.13 1,540,100 69.13
    11-Jan-06 69.67 69.67 68.93 69.49 863,400 69.49
    10-Jan-06 68.70 69.56 68.51 69.50 1,166,500 69.50
    9-Jan-06 68.50 69.20 68.46 68.87 994,900 68.87
  5. davidrousseau,

    That's a nice feature for a trading platform to have (I don't believe IB has it). I've often thought of using such an approach, especially to stops but always was bothered by the possibility of little freak accidents and bad data spikes causing problems such as you just encountered. It's one of those things that you go thru, get upset with and hopefully learn from and come out with a better way around in the end.

    One thing I've learned with etf's is not to trust the 1st 5 min of data (9:30-9:35). You might want to filter that out.
    Another thing would be the actual algorithm that sets the whole thing in motion. Is it based *only* on last price or maybe bid-ask data too? ETF's often have very erroneous bid-ask in first 5 minutes.

    Good luck!
  6. I should be called risktaker, but if I really was, I would use even tighter stops. I just started enabling the stops at 9:45, and you can also specify a volume to be reached for the stop to trigger.

    I'm stillk an amatuer but do understand that there is frequent manipulation. I have a couple of simple theories.. first, I have a list of pretty strong stocks compiled from various sources. Two categories.... great fundimentals, and in particular, energy. It's high now and most of the stocks I look at trade around a 10 point range. When they go low, I'm back in.

    The second is a momentum approach where I buy in when the stock is rising on high volume, and place tight trailing stops. Entry is often a limit below the current price, so when they run stops, I'm in.

    One trick I learned at Cybertrader, is to own one share of a stock you are watching. That way you can set an alert to buy on increase or decrease, set a stop, and set a profit taking stop. The reason for one share is that you can't set stops on shares you don't own. One share is good enough.

    I might set a buy of SPY at 129 for tomorrow, set a stop to take profits at 131 or 132, and a stop at 128.5. A nice shot out of the hole at the open could trigger the buy, and if it opens down, nothing happens. I think CNBC and the other "influences" may try to drag the DOW over 11,000 so they can create a selling spree after IBM and INTEL report.

    Or, I might sleep in. A missed gain is better than a missed loss to me. Look at DNA..... good results and they got hammered. They will most likely be back in the 90's.