wilmott.com

I have been testing my volatility model with paper trading options using real-time data. The trading methodology has been simply to look for mis-priced options and hedge. Just lately I noticed some situations where I made big profits very quickly. This was because the market underwent a major move and not all option prices kept up immediately. So, in one case, in the space of less than a minute I cleaned up about 10% of my total invested capital.I don't have the data saved (and my position was very complex) to verify whether or not these profits would have been made even if I had hung on until the market was stable again, so I am assuming that they were due to speed. (However, I did not trade both ways during this instability, all I did was hedge my existing position.) If the speed was critical, then my model would not necessarily have played a major role. The mis-priced options stood out like a sore thumb because of my real-time graphical skew display -- although my model did give strong confirmation as to which prices were wrong and did indicate what trades were necessary to hedge.Clearly whether or not this would have happened with real trades depends on whether or not my orders were filled. So my question is this, given that I am trading on the internet and assuming I am using an electronic brokerage like IB, and assuming my profits were due to speed, were these opportunities real and would my orders have been filled?Subsidiary questions:1. Can market makers stop offering a market when things are rapidly changing, until they have got their prices up to date, or must they continue to offer their last prices until they change them?2. When there is a fast market, is there any likelihood that my quotes were actually too stale (by the order of 10 seconds or more)?3. My trading methodolgy is pretty primitive. Typically instead of trading several contracts of a single strike it trades single contracts of multiple adjacent strikes. Is this feasible?BTW, the underlying is a highly liquid asset with a huge array of strikes. Market makers who are not well-equipped with computer technology would have a hard struggle keeping their prices up to date.Any comments welcome.

Any comments welcome?How about this: I can't believe you are asking all these questions without telling us which underlying/which options we are talking about!1. Can market makers stop offering a market when things are rapidly changing?Depends on the Exchange.2. When there is a fast market, is there any likelihood that my quotes were actually too stale (by the order of 10 seconds or more)?Depends on the Exchange.3. My trading methodolgy is pretty primitive. Typically instead of trading several contracts of a single strike it trades single contracts of multiple adjacent strikes. Is this feasible? What market?!Jeez man... get a clue. I can take an educated guess that you are either talking about OESX, ODAX, OEX or SPX options... but I can't obviously mind read can I?

QuoteJeez man... get a clue. I can take an educated guess that you are either talking about OESX, ODAX, OEX or SPX options... but I can't obviously mind read can I?Ok. Sorry about that. I wasn't sure whether that information would help or hinder (as in possibly mis-direct) . The underlying was QQQ. I think the time was a little more than an hour before the close (plus or minus about half an hour).Actually there were some other misleading things about my original post -- as I now realize with hindsight. It all happened very quickly and I was so busy clicking on buttons that the sequence of events didn't really register with me until I had a chance to take a break, have my lunch and a cup of tea well after the close.What I think actually happened as I best remember it was this:1. This particular underlying had been in an "off-line" state (not getting real-time quotes) for about half-an-hour, when I re-opened it and saw, as the quotes were refreshed, the value of my position undergo a sudden huge increase (the amount of the profit). This suggests that the profit perhaps was already there in my position and was indeed earned by my model.2. I opened up the skew graphs and ran my calibration fitter to adjust the model. I then realized there had been a major adjustment in the market because several vols were dramatically out of line. 3. Guessing that this mis-pricing opportunity might be short-lived, and not knowing whther or not it was critical to my value gain, I immediately generated potential new trades based on mis-pricing and saw the projected gamma and vega rise from a small positive amount to a considerably larger positive amount. I pressed the execute button (which included delta-hedging with the underlying itself) and saw 90% of the value gain locked into realized profit (because of positions in individual contracts that were closed) and delta-hedging resulted in my underlying position being less short.4. I checked the skew graphs again to re-calibrate my model, saw that they were re-arranged and some different prices now out of line, so I repeated #3. This time there was no new immediate profit.5. I then spent the next 10-15 minutes re-hedging my position (reducing gamma and vega) manually with multiple small trades.6. Overall, after hedging, my delta-balancing position in the underlying underwent a major reversal from very short to moderately long.I realize this is a different story to what I originally wrote, because I was rather over-taken by the drama of so much happening in such a short time and seeing such visible evidence of major mis-pricing, even if each mis-pricing may have lasted for only a few seconds or (maybe as long as a minute) and started wondering about all the questions regarding quotes and market makers I asked in my previous post. I am still not sure whether this profit was really attained by my model or whether it was due to freak prices that might not have been genuinely available. So I would still be grateful for answers to my questions.Further info regarding the extreme short-term mis-pricing (as opposed to the more stable mis-pricing that I believe my model took advantage of): IIRR, I think it was mainly puts on the high-strike (OTM) wing being over-priced and maybe having BBO bids significantly higher than BBO asks (presumably because of exchange differences) I didn't look at the actual prices as this was all automated, but I did notice that the bid IVs were about twice or more what they should have been. Unfortunately I cannot recover this information easily as my position was very complex and the trade search and resulting trades all automated, but not logged.I think this means (check my reasoning) that my profit came from the overall market-re-adjustment rather than the short term mis-pricing because:To have profited from the over-priced puts I would have had to sell them. This would have decreased gamma and vega and made my short position in the underlying shorter. But my opportunistic trades actually increased gamma and vega, and made my underlying position less short -- presumably because I chiefly bought other puts that I had previously sold short!. This suggests that the puts I previously sold had already been re-priced, locking in my profit, and the ones that were seriously over-priced for a few seconds were just late catching up and didn't generate new trades, probably because I was already short them. However, this may be all hogwash as it all happened so fast, I can no longer be sure about what really happened.Once more, I invite comments from those with more experience than myself who may have had similar experiences.

i think fdax has it right on the depends on the exchange. on qqq you will be dealing with 5 exchanges. at certain times the exchange(s) will lift their auto-execution capability. also, you may have trouble with certain routing techniques. in other words, the firm that is executing your business may have an agreement in place with a market maker that may decide not to honor the price you are seeing. having a good trading technique is only 90% of the problem -- the other 90% is execution.i guess the bottom line is that it is going to be difficult to tell what is what until you try it. as a guess, i am going to say that qqq options won't have serious mispricings like what you mention. it is just too liquid and there are too many sophisticated players. that is not to say that there are not opportunities. besides, i would not want to discourage someone from picking up the $20 bill on the floor simply because it could not possibly be there according to theory.

Agree with apine. Mispricings like that don't occur often.First things first, I quote myself: In option markets, at any one point in time, someone is either fucking up, fucking someone or getting fucked. Note that the first and third are not necessarily the same.Fermion: mainly puts on the high-strike (OTM) wing What?! I assume you are just being tired. Leaves the question, did you sell OTM puts or in the money puts? I assume you sold downside puts...Fermion: it was due to freak prices that might not have been genuinely available.I'm quite confused... I thought you traded at those prices? Obviously they must have been if you traded?!I can tell you a hundred stories of similar situations where electronic eyes run amok or people input the wrong parameters in their quote machines.As apine said, gotta try it to see if it works. What suprises me is that you actually executed all this after bringing the system freshly online. You are a braver man than I, for I would have assumed that the model was screwed up, and would have gone to investigate. So your either very brave or very short sighted, you decide which one.Regards.

it sounded to me like he meant "virtual" fills. i.e., he was doing a simulation when this occurred.

QuoteFermion: mainly puts on the high-strike (OTM) wing FDAX: What?! I assume you are just being tired. Leaves the question, did you sell OTM puts or in the money puts? I assume you sold downside puts...This was referring to the options that were seriously out of line for a very short time. Also they were for a long term expiration (March04 and/or later, I think) and, since my model is calibrated over all expirations as well as strikes, it can pick up mis-priced expirations as well as strikes. In this case I think these mis-priced vols seemed about twice what they should have been according to my model and may have been for bids only. As I reasoned on reflection, I don't think I traded these, but my position was so complex it was difficult to tell. I think the profit came primarily from puts nearer the money that I had previously sold at a relatively high price and which suddenly became profitable when the market re-adjusted, while I was off-line and just before I came online again. On futher reflection, I think this all happened at a moment when QQQ, which had been dropping all day, I think, hit the 30.00 mark and stopped.QuoteFermion: it was due to freak prices that might not have been genuinely available.FDAX: I'm quite confused... I thought you traded at those prices? Obviously they must have been if you traded?!.....What suprises me is that you actually executed all this after bringing the system freshly online. You are a braver man than I, for I would have assumed that the model was screwed up, and would have gone to investigate. So your either very brave or very short sighted, you decide which one.No, no. I was only doing a simulation to test my model! That's why I wonder if it could ever have been real. I think I would not have been anywhere near so cavalier if it had been real money! (In fact I don't have enough money to have actually held this position. Even though I was net short, so rolling in virtual cash, I don't think I could have satisfied the margin requirements.)

Quote Fermion: mainly puts on the high-strike (OTM) wing FDAX: What?! I assume you are just being tired. Leaves the question, did you sell OTM puts or in the money puts? I assume you sold downside puts...Whoops! Oh, yeah I must have been tired. I meant ITM, sorry about that. I have occasionally been known to stop at green lights and go on red.

OK, so you simulated. Next time say that. That changes the picture to..... : "Extremly unlikely that you could have made money"... not to say virtually impossible.Regards.

You might benefit from adding a few lines of code to your model which dump any attempted executions to a file so you can analyse this sort of stuff later!

FDAXHunter: "In option markets, at any one point in time, someone is either fucking up, fucking someone or getting fucked. Note that the first and third are not necessarily the same."FDAXHunter,How are you (at this one point in time)? LT.

Better than you with your short vega position, I'm sure...

QuoteOriginally posted by: FDAXHunterBetter than you with your short vega position, I'm sure... That hurt.

I'm sorry... you know I love you, right?

Yikes