I don't know how to post a smart go format game, but I just crushed it at 5 stones. I sacrificed 3 stones early on and it removed them from the board when that was totally unnecessary. It let me play the 3x4 point under my handicap stone with an shimari to the 7x3 point, solidifying a huge territory when it had easy invasions at various points.

As i said earlier, when every move is a sure loss, it will just play one of them more or less at random, since they all lead to the same result in its opinion.

It simply does not handle high handicap games. If you want a handicap opponent to play against, Leela Zero is the incorrect choise (you might be able to get away with 1-2 stones, depending a bit)

Try 7 stones, but this time against KataGo. It will crush, most likely. KataGo tries to maximize score.

I think the best network to play handicap go with regular LZ is THIS ONE



If you want to post a game, you can for example
1) upload it
2) click "place inline"
3) click "preview"
4) right-click on the attachment, and choose "copy link location"
5) paste it
6) select what you just pasted, and click the "sgf" button


it should be ok, the preview should look like this :
It's probably not the best method, but it works.
_________________________________________________________

BTW, network 14a3a5f wins ~80% of its H8 games, with komi 7.5 against network #48, which is supposed to be 1d, according to THIS SITE (I'm not sure about the exact settings they used for #48 to be 1d)

settings :
Black #48 : --noponder -r 1 -p 1601
White 14a3a5f : --noponder -r 0 -p 1601
(here, W uses 15 times more time than B at equal playouts !)

I will look into some of the other alternatives mentioned.

leelaz-model-swa-24-192000_quantized Is this the right weight file? https://www.youtube.com/watch?v=FV6dSjnx7jY It was recommended by that video.

I lost to leelaz by 3 points at 3 stones, so I have a hard time believing it could beat a top pro at 3 stones.

Bear in mind LeelaZero is not a single Go player, but an infinite family of them. Your LeelaZero is a combination of [network file, number of playouts/visits/time settings, LZ engine version, other factors]. My and anyone else's LeelaZero is a different set of those things and can thus be stronger or weaker. Did you run without the time setting parameters and just the playouts, so you can measure how long your computer takes to get 2500 playouts per move?

Here is a version of LeelaZero (LeelaMaster 20 block weights, beefy hardware) beating Kim Youngsam 8d on 3 stones (with komi) over a year ago.

I tried it without the time settings and it took about 40 seconds per move, so at 5 seconds, it is getting like 300 playouts. How much weaker would that make it?

Well, 300 playouts is very very low, so dont expect pro-crushing strenght. Also you said 5s per move - you responed in 5 secs too?

Giving 3 stones (komi=0) to top pros might be a bit too much.FineArt which is not only damn strong in even games, but also extremely good at handi games has special account which plays 2 stones with 7,5 komi.Against one Taiwan 9p he is 45 wins - 8 loses, so i guess they might be around even with just 2h (no komi).This FA played 2 games with Ke Jie with result 1-1.It was more than year ago, so FA might be stronger now, but 3 stones to top pros might be too much for bots.

Is somebody really that strong at computer science?

Sorry, could not resist.

I was averaging significantly less than 5 seconds per move, but since I am not a computer I wasn't taking exactly 5 seconds for each move.

So I know the breakthrough was using AI techniques, and that chess programs were successful mostly without them. However, I guess computational power is still really important.Those programs beating top pros are presumably running on powerful computers and looking at a lot more than 2500 playouts.

Chess is hugely easier to calculate, and as such there are strong ab search engines that can run on CPUs instead of GPUs.

You dont need a hugely powerful GPU to beat pros on a even game with Leela. A GTX 1650 for 150e running in fp16 mode would probably be enough (even though you might want to get a card with more memory, i dont know if a 40b network will fit in 4gb)

Naturally, a intergrated intel gpu, or a old laptop gpu wont work very well, but you can always let it think for longer.

Edit: none of this will however "solve" handicap play

Edit2: what are you using to run leela btw? If its Lizzie you should note that its distributed with a very old, rather weak network that you need to update yourself

That's the funny part, you will always lose by like 3 points. You will never beat it.

Maybe, but I beat it repeatedly by like 50 points at 5 stones.

I tried ignoring its moves in the corner, sometimes ignoring everything and making a loose line of stones down the center. It isn't the best way to win, but it leads to a better games than playing against 3x3 point invasions and just trying to make moyo into territory.

How many times does it need to be repeated that it does not know how to play high handicap?

Whatever you try at 5-4-3 stones, even 2 stones handicap, is completely irrelevant.. its not what it was designed to do..
(EDIT: And do note.. when the handicap goes higher, the bot gets much worse. It is NOT a linear weakening according to the amount of handicap stones..)

Now.. us wanting to have a bot that CAN play high handicap, thats normal, and many people are working on that (check out KataGo for example),
but just because you want it be able to play high handicap, and you thinking it should be able to, does not make it so..

Normal Go thinking around handicap and handicap systems does not apply here.. anyone testing or trying to say anything about the strength of go bots by handicap.. well.. useless information (sorry)

Yes. Top level bots are trained on even games between evenly matched opponents. High handicaps trigger attempts to win the game that is almost certainly lost. With humans that leads to plays called looking for a place to resign. OC, when humans give high handicaps they assume a certain level and type of mistakes by Black. Those bots have no training under such conditions.

_________________
The Adkins Principle:

At some point, doesn't thinking have to go on?

— Winona Adkins

Everything with love.

Is somebody really that unsubtle and crude ?

Yes, and direct and effective too.

I play @Tygem with LZ, exclusively with my Split-Fusekis (which are inferior Fusekis). I had no problems reaching 9d. Within 9d it's 50/50, but I also did beat a 2p once. I play with #158 the old traditional one.

I just checked and name of bot i mentioned earlier is "geomancer2".If you still dont believe in LeelaZero strenght (and i think you dont) try to give it a shot.I checked games and it beat some 1d on 9 stones,2d on 6 stones, and 4d on 3 stones - though it also lost to some 1d on 4 stones (all KGS games).Imho worth try if you doubt in Leela power.

Try Katago, anything above 2 stones for LZ starts to make it play erratically.

You probably won't beat the new version with 5 stones, I've only seen it lose at 8 stones