Life In 19x19

the number of games you've played is massively far too few to be reliable in determining that one version isn't stronger than another.

Don't do that. He won't get it. He has no idea of statistics, but tries to explain the world. To make it even worse he is not able to communicate in english.

Bye ...

"Engine usually uses" - doesn't mean engines tests equivalency...

Your understanding of equivalency without variety would mean you can only play two games.

Contradiction is that 1h game on 1 core and 2h game on 4 cores couldn't be the tests for determining one rating of engines,

This has ...

On Your once again contradiction (that You never wrote)... The game longitude must depend on number of moves: it must not be the same for 91 moves and 291 moves games. I never used visit limitation in my tests. May be I had quoted other test in some context... I limit only move time (except one ...

Some mean that you can measure the strength with a few games as long as the quality is good enough.
I haven't seen such claims

You can start reading at #227.


If you use little time/playouts, you can determine the strength with little time/playouts. If you want to know the strength with much ...

Your basic oversight is only worrying about the absolute margin of error.

Indeed, this was the subject of debate. Some mean that you can measure the strength with a few games as long as the quality is good enough.

As you can see the stronger engine is expected to win more games under high ...

Almost all engines with neuronets are using MC search too (and are using it results for resign), for example, in LZ: neuronets - visits (and nneval win values), MC - playouts (and win %)...

You still didn't understand. It was related to "Monte Carlo Tree search" and not to "engines that use Monte ...

Almost all engines with neuronets are using MC search too (and are using it results for resign), for example, in LZ: neuronets - visits (and nneval win values), MC - playouts (and win %)...

You still didn't understand. It was related to "Monte Carlo Tree search" and not to "engines that use Monte ...

The number of playouts must be high enough to get a statistical significant result.
I am glad, that You understood the main idea...

I am sorry to say that, but once again you didn't understand at all... This was related to Monte Carlo Tree search...

You can't participate in such discussions ...

The size of the net and the low parity factor is closely related (larger nets are stronger but slower). Same-size nets tend to be closer in strength, that's why the curve is less steep. And again, there were plenty of other tests done beyond the single linked graph.

And where is the contradiction ...

Even this graph shows the same curve for nets of the same sizes starting with the same number of playouts, where the strength difference is smaller thus the low-mid point is at playout parity (1:1).

Well, I see 2 graphs of nets with the same size, at one you can adumbrate some kind of a u shape ...

Even this graph shows the same curve for nets of the same sizes starting with the same number of playouts, where the strength difference is smaller thus the low-mid point is at playout parity (1:1).

Well, I see 2 graphs of nets with the same size, at one you can adumbrate some kind of a u shape ...

Anyway, in simple estimation, the more instances you have the greater the convergence.

For every engine you can say the more playouts, the stronger. But the benefit is not linear. For above mentioned reason there is less benefit with a very small and a very high number of playouts.

The graph ...

BTW, has anyone come up with an explanation for the U shape graphs? Thanks. :)

I'd say with a few playouts the factor for the smaller net is higher because in a Monte Carlo search the number of playouts below 100 don't help much. The number of playouts must be high enough to get a statistical ...

BTW, has anyone come up with an explanation for the U shape graphs? Thanks. :)

I'd say with a few playouts the factor for the smaller net is higher because in a Monte Carlo search the number of playouts below 100 don't help much. The number of playouts must be high enough to get a statistical ...

The data is about how strength DIFFERENCE (thus the expected win%) between the same two nets changes with increasing playouts (for both nets, proportionally). There are data points even for nets of the same sizes (both here and elsewhere).

Exactly. And the data say nothing about statistical ...