Comments on Max Dama on Automated Trading: Market Impact

It looks interesting Barry.

2010-03-09T09:12:28.577-08:00

It looks interesting Barry.

Hi, this is a nice post. On a similar vein, you mi...

2010-03-09T03:08:31.827-08:00

Hi, this is a nice post.
On a similar vein, you might find my book 'Algorithmic Trading and DMA' - website www.algo-dma.com interesting,
cheers
b

Nizar, I haven't settled on heuristics such as th...

2009-11-16T04:13:46.327-08:00

Nizar,

I haven't settled on heuristics such as the one you suggest. Perhaps others have useful opinions on that.

There are many services and software packages to estimate transaction costs but they are primarily targeted toward institutional investors. For example prime brokers all provide transaction cost analysis (TCA) and so do some companies like ITG: http://www.itg.com/offerings/measurement-analytics/itg-post-trade-analytics/

Regards,
Max

Max, Solid post. I was actually doing a bit of w...

2009-11-13T17:23:23.643-08:00

Max,

Solid post.

I was actually doing a bit of work on this just a few weeks ago.

The effect your order has on the market is something fascinating because I don't know of any software that can simulate this.

Maybe yourself or one of your readers can enlighten me?

Backtesting and simulation, as far as I'm aware, assumes that your order would not affect the price, and that everything takes place as if your order wasn't present.

So to determine approximate scalability of systems, I have estimated that if you are buying anymore than >5% of the volume then you will move the price.

This estimation is not based on anything from than market observation.

eg. if you were trading a system based on 1-minute bars, and you have flexibility to scale into your position at anytime throughout that bar, if the average bar liquiditiy is $1million, then maximum position size is $50k.

Max, do you think this is a reasonable assumption, and what has your research shown in this regard?

And thanks for those links Max and Nik, I'll be taking a look at them over the weekend.

Nizar.

Thanks again Nik, I registered. This is really use...

2009-11-05T12:51:48.405-08:00

Thanks again Nik, I registered. This is really useful

Regards,
Max

http://events.unisfair.com/index.jsp?eid=455&seid=...

2009-11-05T12:41:30.932-08:00

http://events.unisfair.com/index.jsp?eid=455&seid=32&code=finvirtual

dont know if you can log in, if you dont have a account specifically for this.

Best
-Nik

Thanks a lot for those references Nik. Exactly w...

2009-11-05T12:04:35.812-08:00

Thanks a lot for those references Nik.

Exactly which webinar is it you're referring to? Is it here? http://www.mathworks.com/computational-finance/demos.html?show=recorded

Regards,
Max

Thanks Chintan. Either look at the order book and...

2009-11-05T11:53:39.312-08:00

Thanks Chintan.

Either look at the order book and see how much volume is available at the best bid/ask or empirically measure it from past executions.

m, Look for George Sofianos at GSET. For example,...

2009-11-05T11:51:43.370-08:00

m,

Look for George Sofianos at GSET. For example, here's some of his writing: http://gset.gs.com/gset/education/thought.asp

as an example if an instrument has a daily volume ...

2009-11-05T09:09:36.679-08:00

as an example if an instrument has a daily volume of 1000k then how many % of contracts of daily volume can be traded without any slippages?

Any formula to calculate minimal participation rat...

2009-11-05T09:07:56.574-08:00

Any formula to calculate minimal participation rate at which institutional investor can enjoy privileges of retail investors like no slippage etc...?

great post! do you have additional reference docu...

2009-11-04T14:59:32.008-08:00

great post!
do you have additional reference document/link??

excellent article! Glad to see you posting again ...

2009-11-04T06:46:03.242-08:00

excellent article! Glad to see you posting again

What I found quite interesting in the past was how popular liquidity seeking algorithms such as Dagger, Guerilla or Sonar are detecting icebergs. The reason being that apart from lifting the icebergs for your own execution, one can sort-of front-run them. Turned out it's quite easy: http://www.uni-graz.at/socialpolitik/papers/Frey.pdf

The immediate replenishment with a new peak size after a transaction provides the market with a signal that the price level is likely to contain hidden liquidity. We construct an iceberg detection algorithm that uses the order book dynamics to make predictions of whether or not the bid or the ask side has an iceberg order. For a typical stock, the detection algorithm correctly detects approximately 70% of the books with iceberg orders and when the algorithm generates an iceberg signal it is correct for approximately 90% of the books. These results suggest that marginal liquidity providers can, on average, respond to iceberg orders by changing their order submission strategies.

There is also a 1 hour presentation on this using Matlab, which was featured on their (Mathworks') last quantitative finance virtual conference and can be downloaded in their archive.

Great Article with deep insights!

2009-11-04T04:57:32.009-08:00

Great Article with deep insights!