Option replication and exotics journal

[stock777]

lol, I'm afraid the answer, while detailed and informative, requires that I understand more then I do.

That's how you survive in this game, you don't mess with things you don't really comprehend.

The thing is, if you are doing so well, it means they are mispricing the product. Now, if you tell me these are not really arbs, but directional bets that require you to actually be RIGHT about some end result, then it would be clearer.

The impression I get from the thread , correct me if I'm wrong, is that you are gaming some pricing edge here, and not placing 'bets' on outcomes. Maybe its a combination of the two?

As some of the other posters have stated, is this something that could be replicated, or does it require the masters touch?

Thanks for your response.

PS At the racetrack , they call them 'beards' , though now you can bet easily wth anonymity. Hire a few to shill for ya.

 

[riskarb]

No, these are anything BUT arbs... honestly; they are garden-variety synthetic straddles, and 25delta strangles traded exotic. The edge for me, when compared to vanilla options, is the lack of risk beyond the barrier strike. It allows me to solve for my risk to strike and lay the risk off symmetrically[synth straddle]. In a very-real sense they are less complex due to the binary condition.

There is no implicit pricing edge, to the contrary...

 

[ssternlight]

Quote from riskarb:

No, these are anything BUT arbs... honestly; they are garden-variety synthetic straddles, and 25delta strangles traded exotic. The edge for me, when compared to vanilla options, is the lack of risk beyond the barrier strike. It allows me to solve for my risk to strike and lay the risk off symmetrically[synth straddle]. In a very-real sense they are less complex due to the binary condition.

There is no implicit pricing edge, to the contrary...

Yes... I was musing on the defined risk point the last few days -- while still not coming up with an answer about the difference in pricing a binary and a vanilla with identical expiry dates.

I actually think the shorter time frame of the trades is also advantageous. To follow this strategy with 1 month options seems to invite many more adjustments to positions and the potential for erroneous judgements. The pricing also seems to be incredibly different -- although as I said I haven't really tracked down why.

If you are taking a 10%+ haircut off the model then there must be a strong dose of "art" to your work -- or the models are just wrong.

I actually pulled up the demo MBRM add-ins and learned that I know nothing about this subject -- it wasn't just humbling it was like being back in kindergarten. The Hoadley stuff is much more intuitive to figure out.

Looking forward to seeing how you do with FX.

As always thanks for the insights. This thread would certainly get my vote for best/most interesting on ET.

 

[IV_Trader]

Quote from riskarb:

It's an OTC transaction. I use 4 primary dealers; all of which happen to be in the EUR.

The longer term structures[vega bets] are hedged with vol and variance swaps, as well as replication with other dealers. Smaller terms/notionals are laid-off with other dealers or gamma traded in spot/futures markets, or with vanilla OTC and exchange options.

There is initially quite a substantial model-edge for the house. In most instances it equates to 10-15% of notional. Fortunately, the edge is bled due to the theta which marks down otm gamma. This is analogous to a contraction in stat volty.

Anyone with access to hoadley's free add-ins could've modeled any of the positions using isolated, strike vols. Someone PMed me and mentioned that on some of my trades that he'd been trading similar strikes/tenors on betonmarkets.com, and found that some of my dealer quotes were overpriced. He was getting better fills than me. A few of the indices aren't available on BOM.

I'll admit it's been tougher to get reasonable pricing recently. When starting out I paid 800-1000basis in edge to initiate, now it's twice that number. Unfortunately, with a dealer market you become "known"

I think one can build the double barrier touch/no touch bets with the Weeklies via strangles, it should be very very close to odds that you getting (considering short time period). And if one can , them the house edge goes down to ask/bid spread instead of 15% . Now if I only can figure the how to hedge it like you do!

 

[riskarb]

Quote from IV_Trader:

I think one can build the double barrier touch/no touch bets with the Weeklies via strangles, it should be very very close to odds that you getting (considering short time period). And if one can , them the house edge goes down to ask/bid spread instead of 15% . Now if I only can figure the how to hedge it like you do!

There are some significant problems with vanilla replication. Consider the risk of outliers with each strategy. Within the framework of the synthetic-exotic you would earn windfall profits on your spot position if it were to trade a large fig beyond the barrier strike. The barrier would go OTB at debit-loss, but the spot/futures would earn large gains beyond the barrier.

In the example of a synth-vanilla straddle; the risk into the strike price doesn't invert-modailty as it can with an exotic. Of course, the advantage is the analog-curvature if the strike is reached on the vanilla. Another advantage with the vanilla is that the hedge is still dynamic. Once the exotic is off the board you're subject to reversal/whipsaw, w/o a corresponding change in PnL on the null-exotic.

Away from the strike, both hedges perform similarly.

 

[IV_Trader]

Quote from riskarb:

There are some significant problems with vanilla replication. Consider the risk of outliers with each strategy. Within the framework of the synthetic-exotic you would earn windfall profits on your spot position if it were to trade a large fig beyond the barrier strike. The barrier would go OTB at debit-loss, but the spot/futures would earn large gains beyond the barrier.

In the example of a synth-vanilla straddle; the risk into the strike price doesn't invert-modailty as it can with an exotic. Of course, the advantage is the analog-curvature if the strike is reached on the vanilla. Another advantage with the vanilla is that the hedge is still dynamic. Once the exotic is off the board you're subject to reversal/whipsaw, w/o a corresponding change in PnL on the null-exotic.

Away from the strike, both hedges perform similarly.

the waiver of benefits (or liability) to go pass barrier strike should be at zero after many trades (I think) when one plays odds for a five days( chance for reversal/whipsaw goes down). Actually , if you can enter the bet for ONE day only , the odds/payout for both exotic and vanilla should be the same ( minus the 15% house edge). Don't have the final numbers for 5days/week yet , still working on it.
Good luck with FX

 

[ssternlight]

Quote from IV_Trader:

the waiver of benefits (or liability) to go pass barrier strike should be at zero after many trades (I think) when one plays odds for a five days( chance for reversal/whipsaw goes down). Actually , if you can enter the bet for ONE day only , the odds/payout for both exotic and vanilla should be the same ( minus the 15% house edge). Don't have the final numbers for 5days/week yet , still working on it.
Good luck with FX

IV, would you be willing to post your numbers when you finish? I'm interested in understanding how people are looking at the valuing of the barrier vs the vanilla -- all other things being equal.

Thanks,

Sam

 

[IV_Trader]

Quote from ssternlight:

IV, would you be willing to post your numbers when you finish? I'm interested in understanding how people are looking at the valuing of the barrier vs the vanilla -- all other things being equal.

Thanks,

Sam

sure , Sam , but only if riskarb don't mind. I don't want to pollute his thread with my opinions (just with questions sometimes)

 

[riskarb]

Quote from IV_Trader:

the waiver of benefits (or liability) to go pass barrier strike should be at zero after many trades (I think) when one plays odds for a five days( chance for reversal/whipsaw goes down). Actually , if you can enter the bet for ONE day only , the odds/payout for both exotic and vanilla should be the same ( minus the 15% house edge). Don't have the final numbers for 5days/week yet , still working on it.
Good luck with FX

Yes, with respect to many trials, but I am referring to outlier risks beyond 3sigma in which the exotic would actually invert PnL modality. The vanilla position goes PnL negative at approx 2sigmas. The exotic position can be trimodal PnL into the direction of the barrier -- positive above strike/negative at strike/positive beyond approx 3sigma.

 

[riskarb]

Quote from IV_Trader:

sure , Sam , but only if riskarb don't mind. I don't want to pollute his thread with my opinions (just with questions sometimes)

Not at all. I'd be interested in any statistical approximations.