Page 1 of 1

Finessing MOO orders

Posted: Mon Feb 27, 2012 10:34 am
by LeviF
Does anyone attempt to get a better price than the open by monitoring market action and waiting for a dip, etc? Do you think it helps your slippage in the long run?

Posted: Mon Feb 27, 2012 11:44 am
by sluggo
I know people who do this and, yes, they feel it reduces their average slippage over the long term.
But the histogram (distribution) of slippage_reduction has got some very nasty outlier trades,
where waiting to execute turned out to have been exactly the worst thing to do. Wait for a dip
before buying, only to watch the price march steadily upwards. Wait for a rally to sell, however
the price drops linearly. It could happen, it does happen. So it might be prudent to have a plan
(or even a rule), what to do when things go bad while you are waiting.

Posted: Mon Feb 27, 2012 12:18 pm
by marriot
RedRock wrote:Chose to move towards less tuning and greater robustness. And in the process, eliminating an issue re complicated execution sequence and position monitoring requirements.... ie Simplified my life.

Posted: Mon Feb 27, 2012 12:54 pm
by Wisdom
I think a lot of people have attempted this. When you're right, you'll keep doing it. When it costs you a lot of money, you might stop for awhile.

Posted: Tue Feb 28, 2012 4:42 am
by marriot
Having intraday data test should be something like : buy today at limit, then if not executed buy on close.
But i wonder if it is not " easier" just trade with stop order.

Posted: Tue Feb 28, 2012 11:54 am
by sluggo
If you split your order in half, executed half at market on open, and executed the other half using some kind of slippage_reduction finesse, then
  • You would be able to precisely measure the benefit of your slippage_reduction; the MOO half serves as a benchmark, an experimental control;
  • You would cut your maximum remorse in half. If slippage_reduction turns out to be a huge disaster, your remorse for trying it is only half. On the other hand, if slippage_reduction is a wild success, your remorse for not fully embracing it is only half.
One way to buy lower than the open is to use a limit order, as marriot suggests.

Another way to buy lower than the open is to follow the intraday trend and use an intraday trailing stop: Buy at (the lowest tick of the day so far) plus (V ticks), where V is some measure of intraday volatility. Perhaps, V= the average height of yesterday's 30 minute bars. Follow the trend down and when it finally reverses by V ticks or more, buy.

A third way is to use both a limit order and a stop order (one cancels the other).

You might study your past trades and decide that, for this particular market, my average slippage is X ticks. (or, the +1 standard deviation point of my slippage distribution for this market is Y ticks). If my finesse "works" and the market is offering me negative slippage of 2X ticks, I execute now and take the money! That is, I place a limit order to buy at (Open minus 2X ticks) and pray I get filled. Or, if negative_slippage at this instant of this trade is 2X ticks, I might tighten up the intraday stop. If I have an opportunity to cancel out the slippage of 2 other trades (and this trade too!) then: do it.

(I wrote "ticks" above, but many people measure slippage in dollars rather than in ticks. Others measure slippage in ATRs. Do whatever you think is appropriate.)