Advanced/Professional volatility trading

sle · Aug 30, 2012

Quote from quatron:
<<many things>>

There is more to market-making then making markets to retail punters PMM-style. I just resigned from making markets for a large IB and our platform was a variation on Murex. So, yeah, you can and people do. I would even venture as far as to say that internal platforms at most places are actually worse then these commercial products. Sure, most people rewrite the models and upgrade connectivity, but otherwise it's perfectly reasonable. Low-grade MMs like the guys making markets in Russia or India don't even need too modify anything and they are the bulk of Orc/TBrick/Actant customers.

People in volarb funds e.g. who trade relative vol/skew etc (like me), usually use whatever platform their PB provides and most execution is done via voice anyway (it's pretty hard to find sufficient liquidity electronically). Why would I pay institutional commissions for crappy electronic execution, no matter how snazzy the platform is?

1234 is perfectly right - once I am done, people will do stuff on the back of my trades, but that's part of doing business. Truth is, nothing is secret and you should expect people to to figure out what you are doing sooner or later.

quatron · Aug 31, 2012

Quote from sle:

There is more to market-making then making markets to retail punters PMM-style. I just resigned from making markets for a large IB and our platform was a variation on Murex. So, yeah, you can and people do. I would even venture as far as to say that internal platforms at most places are actually worse then these commercial products. Sure, most people rewrite the models and upgrade connectivity, but otherwise it's perfectly reasonable. Low-grade MMs like the guys making markets in Russia or India don't even need too modify anything and they are the bulk of Orc/TBrick/Actant customers.

People in volarb funds e.g. who trade relative vol/skew etc (like me), usually use whatever platform their PB provides and most execution is done via voice anyway (it's pretty hard to find sufficient liquidity electronically). Why would I pay institutional commissions for crappy electronic execution, no matter how snazzy the platform is?

1234 is perfectly right - once I am done, people will do stuff on the back of my trades, but that's part of doing business. Truth is, nothing is secret and you should expect people to to figure out what you are doing sooner or later.

Well, in asian markets a PMM can make 30% of all the market volume. A PMM usually trades broker markets (i.e. dark pools) as well. Don't know much about US, their role might be different there.
Most successful PMMs I know started with some commercial product like Orc and then built their own. Not just models/connectivity but entire trading cycle. Only risk/backoffice is "outsourced". I'm talking about IMC/Optiver/Tibra/Liquid etc.
Your statement contradicts itself - if their systems are worse than popular commercial ones then how come they can compete and take large market share?

sle · Aug 31, 2012

Quote from quatron:
Your statement contradicts itself - if their systems are worse than popular commercial ones then how come they can compete and take large market share?

Well, that would imply that it's not systems that play a key role in profitability of a market-making desk, especially institutional market making desk. It's the ability to leverage upon chunky client flow (primary) and good risk-taking skill-set (secondary and of diminishing importance these days) that makes money, not super-fast execution or most advanced volatility model.

Quote from quatron:
Well, in asian markets a PMM can make 30% of all the market volume. A PMM usually trades broker markets (i.e. dark pools) as well. Don't know much about US, their role might be different there.

There are no official dark pools for options that I know about, at least in the US.

Quote from quatron:
Most successful PMMs I know started with some commercial product like Orc and then built their own. Not just models/connectivity but entire trading cycle. Only risk/backoffice is "outsourced". I'm talking about IMC/Optiver/Tibra/Liquid etc.

Sure, except as I said, there is more to being a market-maker then being the guy who is able to trade a 100-lot and move the market the fastest. It was not unusual for me to take down a couple million dollars of SPX vega in a single trade - try executing that on the screen with these guys. Their business model is a niche "true-MM" model, highly dependent on "stupid" flow (Timberhill being the extreme example) and lack of concentrated positions.

In any case, trying to play their game on your retail platform and size is simply silly. There are so many great capacity-constrained opportunities in options waiting to be harvested, without any need for 10k-month "volatility-arbitrage" systems, low-latency co-location and proprietary volatility models.

quatron · Aug 31, 2012

Quote from sle:

Well, that would imply that it's not systems that play a key role in profitability of a market-making desk, especially institutional market making desk. It's the ability to leverage upon chunky client flow (primary) and good risk-taking skill-set (secondary and of diminishing importance these days) that makes money, not super-fast execution or most advanced volatility model.

In asian markets MMs have to win the trades by being fast because brokers place retail orders on exchange. So if you are slow you don't get any flow. This might be different in the US.

Quote from sle:

Sure, except as I said, there is more to being a market-maker then being the guy who is able to trade a 100-lot and move the market the fastest. It was not unusual for me to take down a couple million dollars of SPX vega in a single trade - try executing that on the screen with these guys. Their business model is a niche "true-MM" model, highly dependent on "stupid" flow (Timberhill being the extreme example) and lack of concentrated positions.

Guess this was still a vega neutral trade, so you are talking about vega on individual legs.

Quote from sle:

In any case, trying to play their game on your retail platform and size is simply silly. There are so many great capacity-constrained opportunities in options waiting to be harvested, without any need for 10k-month "volatility-arbitrage" systems, low-latency co-location and proprietary volatility models.

Totally agree. But the only point here against professional-grade system is 10k-month. Otherwise such a system would benefit an advanced retail trader too.

sle · Aug 31, 2012

Quote from quatron:
In asian markets MMs have to win the trades by being fast because brokers place retail orders on exchange. So if you are slow you don't get any flow. This might be different in the US.

Nice. Here, a lot of the flow is internalized first (read - the retail traders get raped, which is why retail commissions are so much better then institutional).

Quote from quatron:
Guess this was still a vega neutral trade, so you are talking about vega on individual legs.

Sometimes a Vega-neutral trade, and sometimes just a chunk of risk. You price it according to your current lean, try to lay off some and carry some with you.

Anyway, my sims have finished running, I am off to bed. Talk later.

sle · Aug 31, 2012

Quote from quatron:
Totally agree. But the only point here against professional-grade system is 10k-month. Otherwise such a system would benefit an advanced retail trader too.

If you feel like starting an open-source project, I will contribute in any way I can

Yes, a professional-grade volatility trading system (or futures trading system, or stock trading system) that was free or near free would definitely benefit a retail trader.

quatron · Aug 31, 2012

Quote from sle:

If you feel like starting an open-source project, I will contribute in any way I can Yes, a professional-grade volatility trading system (or futures trading system, or stock trading system) that was free or near free would definitely benefit a retail trader.

I have difficulty in defining what near free is for a retail trader. Every trader is different. How would you price it?

cdcaveman · Aug 31, 2012

i actually did some exploration into this... freaking bam looked at a few of the places mentioned and realized how fucking valuable it is to know how to code such things.. this why quant programmers are getting the fucking big bucks..

cdcaveman · Aug 31, 2012

Quote from quatron:

I have difficulty in defining what near free is for a retail trader. Every trader is different. How would you price it?

i've talked to the guys at Ninja trader... explaining to them the market for options traders... Ninja trader you just write C# scripts to the thing and you can auto trade backtest optimize walkfoward etc.. i would love to have a engine to write scripts to related to trading options i'm not there yet with my knowledge but even just dynamic hedging at N - time lazy style with parameters to purchase based upon gamma assumptions is something more complex then i can do in any platform... there surely is no shortcut..

1245 · Aug 31, 2012

Quote from sle:

There are no official dark pools for options that I know about, at least in the US.

I know of two option dark pools. Instinet and Knight. You can get access to the Knight option dark pool though Option City even as customer or pro customer.

1245