Do you see patterns in Random Walks?

[etile]

If Markets go according a Random Walk, then how does one account for market arbitrage. If the price of of a security is random then sure that would perfectly account for a price discrepancy between markets. It would seem that the variance between the two markets could theoretically be quite wide.

However, when the arbitrageurs enter and bring the price back to relative parity, it would be hard to say that the move from relative disparity to parity between the two markets is an entirely random process.

 

[Samsara]

Maestro, quick few questions.

One of the discussions that arose from the flash crash was the tendency for algorithmic traders to amplify unexpected, adverse movements. This "flocking behavior" was uncharacteristic in that it was likely not derived from the psychology of traders, but from the fact that a significant majority of quants shared the same models of normal price behavior, which could not account for the random Waddell & Reed error that caused it, nor for the compounding behavior of their peers. Would you agree/disagree with this read?

Two questions: how did you guys fare during the flash crash? And do you feel that your reliable method for capturing mean-reverting price behavior is able to effectively assimilate events like the flash crash?

 

[SrRuthenate]

If my interpretation is correct, then algos like M's welcome flash crashes, and are nothing like RTM conventional algos.

 

[MAESTRO]

Quote from etile:

If Markets go according a Random Walk, then how does one account for market arbitrage. If the price of of a security is random then sure that would perfectly account for a price discrepancy between markets. It would seem that the variance between the two markets could theoretically be quite wide.

However, when the arbitrageurs enter and bring the price back to relative parity, it would be hard to say that the move from relative disparity to parity between the two markets is an entirely random process.

Two or more stochastic processes could be well correlated. The product of those two processes creates a third one that has less mean deviation than the original two if the correlation is positive. Arbitrage is based on the notion of reduced volatility of that "synthetic" security created via correlation process.

 

[MAESTRO]

Quote from Samsara:

Maestro, quick few questions.

One of the discussions that arose from the flash crash was the tendency for algorithmic traders to amplify unexpected, adverse movements. This "flocking behavior" was uncharacteristic in that it was likely not derived from the psychology of traders, but from the fact that a significant majority of quants shared the same models of normal price behavior, which could not account for the random Waddell & Reed error that caused it, nor for the compounding behavior of their peers. Would you agree/disagree with this read?

Two questions: how did you guys fare during the flash crash? And do you feel that your reliable method for capturing mean-reverting price behavior is able to effectively assimilate events like the flash crash?

If at any time (including the crash) you were able to spot the situation where the second derivative of the mean (acceleration) is significantly less than the acceleration of the price you have an ability for RTM. We were doing very well during the flash crash and in the recent weeks. However, the reaction time of the system and alternative gate ways to the market to avoid illiquidity pockets is the must! Also, by doing so we actually make the flash crash softer as we offer liquidity rather than taking it away.

 

[MAESTRO]

Quote from SrRuthenate:

If my interpretation is correct, then algos like M's welcome flash crashes, and are nothing like RTM conventional algos.

We do not welcome Flash Crashes because at those times there are other factors other than simple price action and book behavior are in play. However, we are well prepared for them.

 

[MAESTRO]

Quote from underwater:

MAESTRO,

Can you please give a rough estimation on how much of the advanced stuff is tradable outright with futures, and how much requires options?

I was playing around with converting normal price series to log-normal with ratio spreads, then exploiting the log-normality. Am I heading totally in the wrong direction and should look elsewhere?

Thanks.

Markets are like chess. There is not just one winning strategy that is a "holy Grail"; there are unlimited number of possible successful strategies that produce good results. Your approach may very well create a winning strategy. I personally feel that Log-normal distribution is too simplistic, however, we had a few good results with it. I do not believe that the markets fit in any predetermined model (Normal, Log-normal etc.). I think the markets are more complex and create randomization of distributions. Having said that, there is always a chance of finding one or two stable patterns that exist during a prolonged period of time.
Good Luck!

Cheers,
MAESTRO

P.S. Options do offer non-linearity of returns that could be instrumental in harvesting some non-linear patterns. However, we have plenty of strategies that only use underlying purchases.

 

[Samsara]

Quote from MAESTRO:

If at any time (including the crash) you were able to spot the situation where the second derivative of the mean (acceleration) is significantly less than the acceleration of the price you have an ability for RTM. We were doing very well during the flash crash and in the recent weeks. However, the reaction time of the system and alternative gate ways to the market to avoid illiquidity pockets is the must! Also, by doing so we actually make the flash crash softer as we offer liquidity rather than taking it away.

Great to hear your insights along these lines, and that the long road to reclaiming your health has improved.

 

[MAESTRO]

Quote from intradaybill:

Apart from philosophy, keep in mind that

"the market can refrain from reverting to the mean for longer than you can remain solvent".

That is only true in case of a static mean. Dynamic means can be designed to "guarantee" the cross.

 

[MAESTRO]

Quote from Samsara:

Great to hear your insights along these lines, and that the long road to reclaiming your health has improved.

Thank you so much!!!! Yes, it has been three long years! But, I guess I will be kicking for some more years! Good to be back!

Thank you!

Cheers,

MAESTRO