# Can the second law of thermodynamics be violated in a small enough system if tried repeatedly enough?

Second law of thermodynamics is observed in the universe because statistics favors it, right? And in large enough system this statistical tendency becomes certainty.

Does it also mean that negative entropy change in a closed system can occur if the system is small enough and it's tried repeatedly enough? Kind of like if you keep guessing the outcome of a 6-sided dice several times at some point you'll get it right. If yes, has it been demonstrated by experiment?

• $dS=\left(\frac{\delta Q}{T}\right)_\intrev$ This is defined for all size systems. – lucas Apr 28 '16 at 5:16
• In statistical mechanics you take the average over an infinite time... so, no. The short term fluctuations are neither an entropy nor a temperature. They are just short term fluctuations. – CuriousOne Apr 28 '16 at 5:21
• This question has been asked many times on Phys.SE. – Qmechanic Apr 28 '16 at 7:25
• google.com/search?q=decrease+in+quantum+entropy shows a few papers which point out that, yes, at a quantum level, entropy can decrease. – barrycarter Apr 28 '16 at 17:29