Based on a quick skim, this is not a good paper. Computer scientists writing on economics is great, it's helpful to grow new ideas in the field. Unfortunately they sometimes use economic concepts imprecisely at detriment to their question, methodology, and results.
That's the case here. This paper posits a definition of efficiency, but does not explain why that definition is correct or how it relates to other efficiency measures.
A better proof of arbitrage opportunities in markets is Wah 2016, which identifies actual arbitrage opportunities in actual markets.
Separately, what does "Since P probably does not equal NP" mean as a probabilistic statement?
And what is the correct way to concisely and precisely write: "most people familiar with the P = NP problem believe with varying degrees of confidence that P is not equal to NP, but so far no proof exists."
That's the case here. This paper posits a definition of efficiency, but does not explain why that definition is correct or how it relates to other efficiency measures.
A better proof of arbitrage opportunities in markets is Wah 2016, which identifies actual arbitrage opportunities in actual markets.
Separately, what does "Since P probably does not equal NP" mean as a probabilistic statement?
And what is the correct way to concisely and precisely write: "most people familiar with the P = NP problem believe with varying degrees of confidence that P is not equal to NP, but so far no proof exists."