Slip and Commission

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
Algonquin
Roundtable Knight
Roundtable Knight
Posts: 158
Joined: Sun Jan 28, 2007 9:38 pm
Location: Greenwich, CT

Post by Algonquin » Fri May 16, 2008 10:16 am

Hello all,

It has been two years since the last post in this very important (and helpful) thread regarding slippage estimates. And during those intervening years the markets have gotten progressively more volatile and noisy.

However, commission costs have dropped dramatically, and electronic trading has become more prevalent.

In the interest of refreshing our collective thoughts on this topic, I thought I would pose the following question to forum:

What changes, if any, have occurred with regard to overall transaction costs (slippage and commission)?

Have transaction costs increased overall due to increased market volatility? Or, has there been an overall decrease due to increased electronic trading and lower commission rates? Or has nothing changed?

What slippage + commission estimates are being employed out there?

For testing purposes on a long term trend following system, I have adhered to a 50% slippage rate (which includes commission costs), and 10% slippage rate on rolls. Having never actually traded a large basket of instruments, I have relatively limited slippage experience, and thus have opted for a conservative approach to my estimations. I would, however, be interested to know what others are doing in light of the many markets changes that have occurred in the past two years.

Cheers,

Algonquin

sluggo
Roundtable Knight
Roundtable Knight
Posts: 2986
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo » Fri May 16, 2008 11:42 am

Algonquin wrote:I have relatively limited slippage experience, and thus have opted for a conservative approach to my estimations.
Bravo! I congratulate you on an eminently reasonable idea. Lacking hard data, you are forced to make a guess about what constitutes a "conservative" amount of slippage. This is unavoidable. When you do start trading, when you begin to track your real-life, real-money slippage, you will have the enjoyable task of deciding how to boil that mass of data down to a figure you can use in subsequent simulations. Will you choose
  • the mean value of observed slippage
  • the median value (50% point of the distribution) of observed slippage
  • a bit more conservative than median, such as the 60% point of the distribution
  • use "trade weighted" slippage, or "number of contracts weighted" slippage
  • and so forth.
Rejoice! You are following Ted Anneman's recommended path to righteousness (link)

Algonquin
Roundtable Knight
Roundtable Knight
Posts: 158
Joined: Sun Jan 28, 2007 9:38 pm
Location: Greenwich, CT

Post by Algonquin » Fri May 16, 2008 12:47 pm

Hi Sluggo,

Thanks for yet another thought-provoking post. And congratulations on recently surpassing the 1,000 post mark. I, like many, have benefited greatly from your generosity of spirit.

With relatively little real world trading experience in the area of large basketed trading systems, I was faced with a quandry of how to estimate slippage across multiple instruments given that all of my available information was third-party hearsay. Naturally, I turned to the many posts across cyberspace that address this topic; from an interview with Charlie Wright of Fall River, to Attains's research, the posts of our own forum members, Ed Seykota's comments on the subject, and so on. I was especially struck (and frightened) by the only real empircal evidence I had; namely, audited track records of numerous CTAs which, over time, had trading results which clustered around MAR ratios of .40 - .70, accompanied by sharpe ratios clustering around .50 - .70. This, to me, spoke volumes about the real world of trading, and the potential impact of transaction costs and liquidity (I assumed that these were highly proficient system developers who could doubtless fashion rather good backtested results).

Ultimately, I decided that I would rather be safe (leaving open the possibility of an pleasant surprise) than sorry, and chose what I surmised to be a conservative approach to slippage estimation (though, in all candor, the decision of what is or isn't conservative is also educated guesswork).

Does this bias my testing in favor of longer term systems? Unquestionably. It has had a very real impact on where I place my entry points, stops, exits, etc. But as Toby Crabel once noted, when starting out trading the primary focus should be to survive, not to maximize returns. Over time, we all improve as traders, so risk adjusted returns will continually improve through the natural learning process. But initially, when lacking hard information on something as critical as slippage, I made the personal decision to maximize my chances of survival, rather than maximizing my returns through higher frequency trading.

Was it an educated decision based on what information was available? I sure hope so. But the basis of the decision was, in all honesty, as much a personal decision as anything else.

All this being said, I would appreciate hearing from any forum members who would like to share their experiences with slippage, especially over the past couple of years.

Algonquin

orange
Full Member
Full Member
Posts: 21
Joined: Sun Dec 30, 2007 5:19 am
Location: Singapore
Contact:

Post by orange » Mon Jul 28, 2008 5:02 am

zacharyoxman wrote:Now, with brokerage costs being lower, I think a $75 c/s figure is fairly accurate, and with more and more markets moving towards electronic execution, the slippage cost should, in theory, decrease as time goes on.
Hi,

I am curious why there is no mention of a bid/ask spread cost or has it been included in the $75 c/s figure mentioned by zacharyoxman in the quote above?

Suppose I buy 1 contract of CL at 124.11/124.15, my spread cost will be $40. I also assume that slippage plus commission is $75 per roundturn. So the total trading cost that I should assume in my backtest for trading 1 contract of CL is $75 + $40 =$115?


Regards,
Feng

Post Reply