Earlier in this thread, I wrote: > Markets are efficient -> (P = NP) > > "But we need to make one other assumption that is currently not standard in modern markets: we need to allow participants to place order-cancels-order (“OCO”) or one-or-the-other orders. These are orders on different securities that automatically cancel the remaining orders whenever one is hit." > > Can you name any market where you can place these orders or are they still in La La Land? >
and > So far *if* the (presumably bug-free) program would fail to solve 3-SAT > problems in La La Land then some markets would not be efficient in the > strong or semi-strong sense. In other words, so far, the alleged strong or > semi-strong inefficiency *might* only be falsified in Lala Land. Raul wrote: > The reason you won't see orders like that in the historic record is > that regulations on the market have not allowed orders of that sort. Can you name a single regulator, the specific regulation, and the jurisdiction where the regulation prohibiting "OCO" applies? On Wed, Sep 4, 2019 at 12:09 PM Raul Miller <[email protected]> wrote: > > On Wed, Sep 4, 2019 at 11:54 AM Donna Y <[email protected]> wrote: > > I think that I found flaws in the paper (how he used Turing > > reduction, whether he properly described market orders as a Turing > > machine and a Language accepted by such TM, other reasoning in his > > argument. It is a bit ridiculous that he proposes a type of order > > that doesn’t exist in the historic record. He also assumes that EMH > > is a N-complete problem, and he tries to demonstrate that markets > > are not efficient without a definition of efficient markets that > > is testable…) > > The reason you won't see orders like that in the historic record is > that regulations on the market have not allowed orders of that sort. > > So if that is the nature of the flaw we find in his reasoning, what we > would essentially be saying is that if EMH is valid, EMH depends on > market regulations to remain valid. > > And this seems like an at least somewhat sensible position, since > regulations limit or eliminate possibilities, and thus can eliminate > or reduce the sort of exponential growth in possibilities that he was > describing. > > In other words, market regulations eliminate NP-Completeness (among > other things). > > That said, it's also worth noting that mathematics is not science. And > it's science which concerns itself with testability, not math (outside > of specific specializations, such as probability and statistics, which > attempt to model tests, with varying degrees of success). > > Thanks, > > > -- > Raul > > > > > > > > If any of my suspicions are true then his argument would be proven false. > > > > It is amazing how much you can learn by investigating claims even if the claims prove to be false. > > > > Many thanks, > > > > Donna Y > > [email protected] > > > > > > > On Aug 30, 2019, at 11:31 AM, Raul Miller <[email protected]> wrote: > > > > > > On Fri, Aug 30, 2019 at 10:25 AM Donna Y <[email protected]> wrote: > > >> arXiv is where Maymin’s paper resides—thus it is not recognized as a proof of P vs NP > > > > > > Also, there's a mathematical distinction between "hasn't been proven, > > > despite tremendous effort" and "has been proven false" which is > > > difficult to express concisely. > > > > > > In colloquial english, the expressions we use to describe "hasn't been > > > proven despite tremendous effort" seem close to "hasn't been proven > > > because no one took it seriously", despite the obvious differences. > > > And, both of them we can and often do treat as "probably wrong", > > > though quantifying the associated probability tends to be wasted > > > effort. > > > > > > Or: even if his paper were published in a prestigious mathematical > > > journal, it still wouldn't be proof of P vs NP. > > > > > > Thanks, > > > > > > -- > > > Raul > > > ---------------------------------------------------------------------- > > > For information about J forums see http://www.jsoftware.com/forums.htm > > > > ---------------------------------------------------------------------- > > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
