214 vs 157

[hydrogenpi7]

on a bad computer still its 157 for the wins

[splee99]

I don't see this with 5000 playouts #157 vs 500 playouts #214.

[Vargo]

on a bad computer still its 157 for the wins

If your bad computer can handle 200 visits or more for #214, then #214 is probably stronger than #157 at equal time setting.

Two 100 game matches between #157 and #214 at equal time (LZ0.16, twogtp 1.5.0)

First match parameters :
-v 201 for #214 and -v 701 for #157, -m 20 --randomtemp=0.3 for both (and --noponder -komi 7.5)
Result : #214 wins 50-33 (60.2%)
Low visits and low randomtemp ---> 17 duplicate games, only 83 games used.

Second match : same parameters, -m 20 and --randomtemp=1
Result : #214 wins 61-38 (61.6%)
only 1 duplicate game, 99 games used.

With more than 200 visits and equal time, I suppose the strength difference between #214 and #157 would be increasing.

**********************************************

The games of the second match (2 and 6 are duplicates, 214 is B in the even numbered games)

214_157.zip

(86.63 KiB) Downloaded 532 times

The second match parameters and stats :

214 157.gif (69.91 KiB) Viewed 4678 times

[hydrogenpi7]

Thanks for the data, I was only giving it around 50... I guess 200 is a reasonable number for benchmark. Good to see that it has finally surpassed 157.

I recall when the dontbiteme guy was posting weekly updates on reddit he had already gone down to around 35 when he stopped posting.