Correlation test reliability concerning clone issue.

[Roman Hartmann]

Testing engines on exotic (or even illegal) positions might be less work and should give you a first idea if the engine in question contains borrowed code.

Roman

[michiguel]

"Sorry, I must not have been clear. When I said "this method", I meant the move correlation test that is the subject of the thread, not your obviously objectionable example."

Ok, it was our misunderstanding.

The origin of statistics is probability theory which was developed for random issues. Using it for not random issues is giving big risk and mostly wrong results.

I can agree that coin flipping is giving random results but choosing the moves by the algorithm is not random issue and can not be treated by statistics. It would be if algorithm will choose move on the base of random function.

All times the problem with 'p', if p is false than any result is true.

Rgds
Chris

I said this before, what is random is the choice of positions not the choice of the moves.

It is like opinion polls. The opinion of a person is not random, but you choose randomly a (relatively small) sample from the general population. You characterize that sample and use statistics to predict the behavior of the general population on election day.

The positions chosen in the tests are a small sample from the universe of different positions that an engine can face.

Miguel

[Hood]

I said this before, what is random is the choice of positions not the choice of the moves.

It is like opinion polls. The opinion of a person is not random, but you choose randomly a (relatively small) sample from the general population. You characterize that sample and use statistics to predict the behavior of the general population on election day.

The positions chosen in the tests are a small sample from the universe of different positions that an engine can face.

Miguel

I understand that but it is always a problem if you are getting the representative probe and if you selected proper distribution.
Stong assumptions weak points.

The voters know practicly nth about candidates, engine knows everything about position.

returning to the main thread, correlation could verify only similarity not clonity , as the cloner is able to fool corelation in an easy way.
rgds

[michiguel]

I said this before, what is random is the choice of positions not the choice of the moves.

It is like opinion polls. The opinion of a person is not random, but you choose randomly a (relatively small) sample from the general population. You characterize that sample and use statistics to predict the behavior of the general population on election day.

The positions chosen in the tests are a small sample from the universe of different positions that an engine can face.

Miguel

I understand that but it is always a problem if you are getting the representative probe and if you selected proper distribution.
Stong assumptions weak points.

There is no problem because no distribution of any kind is assumed.

The voters know practicly nth about candidates, engine knows everything about position.

If the engines know everything, how come they choose different moves?

Miguel

returning to the main thread, correlation could verify only similarity not clonity , as the cloner is able to fool corelation in an easy way.
rgds