Pair Trading Lab Helpdesk & Forum
Pair Trading => Models, Algorithms & Approaches => Topic started by: vladimir vaynshtok on September 04, 2016, 02:54:50 pm
-
Hi All:
Let us say we have a good pair with 100bp return per trade and everything else looks great.
But with low VOL of 100K the BID/ASK spreads are 100bp.
As far as I understood Back Testing assumes Market on Close IB order and getting Close prices for both stocks. However there are couple problems with that. First MarketOnClose order might not be executed for low VOL stock.
Secondly if executed we do not know were it will be executed. Thus we might loose spreads on open bucket and the same on closing bucket.
Do we have any protection against loosing too much BID/ASK spreads via bad execution for Market On Close order or order not filled?
There are some smart execution strategies for bucket that offers some grantees ... just not quite sure if PTLab Trader uses any ?
Thank you very much
-
Dear Vladimir,
at this moment, PTL Trader uses simple market orders to enter positions. For retail traders trading liquid stocks it is enough.
To protect the trader from high bid-ask spreads, the software has a simple protection - internally it uses two z-scores, one calculated from ask/bid and the other calculated as bid/ask. Then it uses the "ask z-score" as a signal to open long pair positions and "bid z-score" to open short positions. So, it basically compensates the bid-ask spread by entering at slightly more advantageous z-score. It looks at bid-ask at real-time and waits for the best moment.
But of course, if you are really unlucky, it may see no opportunity at all. So in general for illiquid markets, real trading brings less trades than backtesting, which is quite simple.
We don't use MoC or LoC orders...
In future we will also have "enterprise" version of PTL Trader with advanced execution algos (based on limit orders fired to order book, etc...). We are consulting potential options with a previous HFT trading expert. But these won't be available for retail users.
For now, you need to stick with liquid stocks to have real profits.
-
I see.
I guess there is one simple strategy that may make sense when 1 stock is not liquid and another is liquid.
Put a limit on illiquid and wait for exe. Once that exe then exe liquid at market.
In many cases that make sense.
-
Makes sense. I will keep this in mind while designing the advanced execution algo.