Discussion in 'Order Execution' started by ZBEAR, Feb 14, 2003.

  1. Pabst


    Never before has a mechanical "fat finger" circuit breaker been implemented on Globex. The previous rule of thumb has been that the high or low in the pit during a Globex FF move will also be the electronic H/L for that period. IMO the problem or test of the circuit breaker will be not when an FF occurs but what will the price action/order flow resemble during a bona-fide panic.
    #11     Sep 14, 2003
  2. Pabst


    Elite Member

    Registered: Dec 2001
    Posts: 1009

    07-31-03 05:42 PM

    #12     Sep 14, 2003
  3. vikana

    vikana Moderator

    This is the biggest reason I don't trade the dows. Illiquid and the CBOT seems to muck with the trades every other day. I have never had that happen in ES/NQ.
    #13     Sep 14, 2003
  4. Cutten


    Because the "algorithm" only states that the exchange *may* bust trades beyond a certain level, not that it has to. A large spike could be a legitimate trade or series of trades, so the exchange can't pass a rule saying that all large spikes get busted.

    However, busted trades are a terrible idea IMO. What happens is the exchange sets a "realistic" low price, and busts everything below it. Instead, they should set the same price, and then all trades below it are treated as though they were executed at that price. That way you avoid getting busted on a leg you closed, and you know that ALL buys and sells will be honoured. If the orders which caused the spike were legit, then you get the price you wanted. If the orders were not legit, then you know you will still be filled at the best possible level of the day. You know your position with 100% accuracy and certainty, and can immediately offset it. Surely this is a better solution that busting out trades, leaving people with huge risk, and disincentivising bids and offers during panic situations.
    #14     Sep 19, 2003
  5. It seems pretty simple...if you get an execution far away from where "the real" contract traded in the pit, you can expect an execution. Obviously the real market NEVER traded there.
    #15     Sep 19, 2003
  6. Pabst


    Cutten I made the same argument in a different thread a couple of months back. But I now have reservations.

    Elite Member

    Registered: Dec 2001
    Posts: 1029

    07-31-03 05:42 PM

    Of course the massive irony is that by canceling trades the exchanges are penalizing the buyers who step in to stabilize the market. Often these busted trades leave buyers naked short and benefit those directly responsible for the dramatic declines. Price changes don't occur in a vacuum. The reasons for the snapbacks is because risk takers stepped in and made purchases, in many cases not knowing whether the sudden breaks were the first reaction to horrific news.

    For the exchanges this will be a tough problem to fix. While many participants choose to have stops away from the market residing on exchange servers, there is little incentive to leave a size bid or offer "live" away from the market. Why work a bid for 500 ES ten handles lower when the only way you're going to get filled is if something bizarre happens as the market trades off 50 points! Thus at any given time, i.e. all the time, the aggregate number of stops far exceeds the number of limit orders in the book. Don't let the size you see in your "market depth" deceive you, that is only a sample of what is close to the market and thus somewhat "marketable". Further away from the current bid/offer the number of limit orders severely decreases and the number of stops hugely increases.

    The CME's secret plan to counteract the avalanche of stops is noble, BUT, what if a legitimate terror, assassination, resignation, or Fed action justifies a sudden revaluation of futures beyond a previous defined yardstick. There were a few rate cuts in the last few years where being filled 10 pts worse on a stop would have been a great fill!! For lack of a better idea perhaps there should be no new circuit breakers but instead could the exchanges offer some inducement (waive fees) to institutions who leave resting orders scaled away. I don't know but maybe a great way to "pay" these institutions for liquidity would be to not bust their bargain basement trades when they're actually winners.

    My concern about the stated solution that we have both expressed is one of obviousness. How can a limit order thats filled during a "fat finger" be filled worse than the specified price. And how can a stop that's not really triggered, i.e. is now out of the new range, be filled. While most participants monitor their positions with some timeliness, many other traders, often have an order working away from the market while away from a computer. I'd be pretty pissed if I found out that my limit was filled 5pts worse and now later in the day the market's right back at my original level and instead of being able to scratch the trade I'm out a fair amount of money. I'd like to argue that in a true auction market, no prices be changed, no trades busted, but I haven't been on the losing (or winning end, though close) of an FF. In the pit you'd go to Federal prison for changing a price. Perhaps it's time for the electronic marketplace to "grow up' and say, we're supposed to be bigger and better than open outcry, we'll live with the consequences of "unjustified" sudden volatility.
    #16     Sep 19, 2003
  7. the no bust zone is really 150 points in YM vs ZD ?

    sheesh ... some lucky SOB's and some unlucky SOB's

    tonight due to fat finger in ZD ??

    YM trades 48 points over ZD

    meanwhile the SP / ES , ND / NQ did not budge !
    #17     Jan 7, 2004
  8. nkhoi

    nkhoi Moderator

  9. jrs3


    Are you referring to this 80pt spike?
    #19     Jan 8, 2004
  10. nkhoi

    nkhoi Moderator

    yes, that's the one.
    #20     Jan 8, 2004