What techniques to market makers use to manage operational and 'massive' adverse selection risks?

[IAS_LLC]

When you say "After canceling, i immediatly will go back to the end of the line (usually)", can you help me understand your logic behind this? It may be a good idea, but I can't really see why immediately looking at it. My thought is that if you're last in the Q, you have really high adverse selection risk, because the guy who took you out just got in on the new BBO without paying the spread. Furthermore, you've lowered your probability of capturing the spread (or whatever your strategy is, because I'm sure it's not to see prices move against you), because the probability of the Q behind you holding up against opposing order flow pressure has become lower than it was at your initial Q position.

I'm an engineer, not an academic so I welcome respectful questioning of my methods.... feel free

When I cancel and go back to the end of the queue, its not because I got taken out by anything/anyone, its because when my estimated position is nearing the front of the queue, I didn't think a fill would be profitable, so I canceled. The "usually" ( i.e not always) caveat I issued in the statement you quoted covers the situation when bbo (or whatever level i'm targeting) is about to move. So if there are only X number of contracts at BBO, and X is less than my strategies threshold, I won't pile on just to have my order filled and watch the market move away from me.

I just want to reiterate, I'm not making markets... I'm scalping for 1-2 pips, and trying to do so very quickly. My efforts to maintain a Q position has more to with exposing myself to more opportunities to capture small market wiggles than it does with trying to employ a market making strategy to capture the bid-ask spread.

Did I answer your question? I ramble sometimes...and don't always speak the same technical language as others..

Cheers

 

[cjbuckley4]

I'm an engineer, not an academic so I welcome respectful questioning of my methods.... feel free

When I cancel and go back to the end of the queue, its not because I got taken out by anything/anyone, its because when my estimated position is nearing the front of the queue, I didn't think a fill would be profitable, so I canceled. The "usually" ( i.e not always) caveat I issued in the statement you quoted covers the situation when bbo (or whatever level i'm targeting) is about to move. So if there are only X number of contracts at BBO, and X is less than my strategies threshold, I won't pile on just to have my order filled and watch the market move away from me.

I just want to reiterate, I'm not making markets... I'm scalping for 1-2 pips, and trying to do so very quickly. My efforts to maintain a Q position has more to with exposing myself to more opportunities to capture small market wiggles than it does with trying to employ a market making strategy to capture the bid-ask spread.

Did I answer your question? I ramble sometimes...and don't always speak the same technical language as others..

Cheers

Thanks for your answer. I've considered something that's probably similar. I'd still personally be a bit concerned about joining the back of the Q at the BBO, but if you know the alpha gained by doing so in the context of your strategy is greater than the adverse selection risk, I really can't argue with that.

 

[IAS_LLC]

Thanks for your answer. I've considered something that's probably similar. I'd still personally be a bit concerned about joining the back of the Q at the BBO, but if you know the alpha gained by doing so in the context of your strategy is greater than the adverse selection risk, I really can't argue with that.

Yup, i'm confident my expected value is > 0 and my std of returns is favorable using the method I am... atleast in sim mode. As you've suggested, I may do better if I'm more selective in where I join the book... just haven't gotten around to evaluating that yet... There isn't enough time in the day, unfortunately. I'm sure you've ran into the same conundrum I have: I know the thing is working... but I also know it needs a lot of improvement... When do you release it into the wild?

 

[Gambit]

yes or in similar markets and with rfqs

Are you connected to CME direct through an API?

 

[cjbuckley4]

Yup, i'm confident my expected value is > 0 and my std of returns is favorable using the method I am... atleast in sim mode. As you've suggested, I may do better if I'm more selective in where I join the book... just haven't gotten around to evaluating that yet... There isn't enough time in the day, unfortunately. I'm sure you've ran into the same conundrum I have: I know the thing is working... but I also know it needs a lot of improvement... When do you release it into the wild?

Extremely loaded question: what are you doing to model Q position for your SIM? We can take that discussion private if you'd prefer. If you're doing something even remotely fancy it will totally derail this thread.

 

[brownegg]

I admit that I am wholly out of my element here. The latency numbers I quoted are straight from vendors. Let's suspend disbelief for a moment. Could liquidity provision in a single market be viable? How about trading the soybean crush?

Single market just straight-up bid/ask is almost impossible to do. In fact I'm not sure I've ever seen it done outside of pit traders in futures and the old NASDAQ equity market makers. But in those cases spreads and liquidity are such that even the tiniest bit of info or time/place advantage could be levered into great money.

Crush is a different story. Are you talking about arbing into the three legs vs the listed contract? Last I knew that trade would line up about once a month--meaning a single instance, not a day or something (I'm not joking). If you're talking about legging into the crush at an advantage of sorts, that's very doable. In fact, I think that's where the HFT houses that really know what they're doing make their money. There are still people running those algos, and a lot of what HFT is about is occasionally capturing free alpha. So if you're a guy who can trade the crush without HFT, yes, you would be very happy. Otherwise, meh.

 

[IAS_LLC]

Extremely loaded question: what are you doing to model Q position for your SIM? We can take that discussion private if you'd prefer. If you're doing something even remotely fancy it will totally derail this thread.

Nothing too crazy, but I'll PM you.

 

[Sig]

Single market just straight-up bid/ask is almost impossible to do. In fact I'm not sure I've ever seen it done outside of pit traders in futures and the old NASDAQ equity market makers. But in those cases spreads and liquidity are such that even the tiniest bit of info or time/place advantage could be levered into great money.

Crush is a different story. Are you talking about arbing into the three legs vs the listed contract? Last I knew that trade would line up about once a month--meaning a single instance, not a day or something (I'm not joking). If you're talking about legging into the crush at an advantage of sorts, that's very doable. In fact, I think that's where the HFT houses that really know what they're doing make their money. There are still people running those algos, and a lot of what HFT is about is occasionally capturing free alpha. So if you're a guy who can trade the crush without HFT, yes, you would be very happy. Otherwise, meh.

I see huge bid/ask spreads in option spreads, even liquid ones like SPX if you're just a couple strikes outside the underlying. I'd be interested in your thoughts on the viability of straight-up bid/ask in that kind of market? I've played around manually and never get hit absent a market move, so I'm assuming there's just so little market order volume that the bid/ask is relatively meaningless? If nothing else I've thought it would be a viable strategy to try to catch the other side of all the Interactive Brokers autoliquidations, although I'm guessing a certain counterparty always somehow manages to offer just inside the bid/ask on those.

 

[Gambit]

@ Sig. I am guessing the pros shop their order around upstairs and retail isn't interested in large index contracts. And they who shall not be named don't leave crumbs behind.

 

[winnertakesall]

I know the thing is working... but I also know it needs a lot of improvement... When do you release it into the wild?

Well, you can just run it in the wild for 1 hour, 2 hours, 4 hours, or even a day, and then compare your real results with your SIM results. If the results differs drastically, you want to look into the difference. But if the result is profitable, try it again next day.