Engine-Test Experiment on Lichess

Solista
Posts: 202
Joined: Thu Jul 06, 2023 11:47 am
Location: Germany
Has thanked: 7 times
Been thanked: 293 times
Contact:

Re: Engine-Test Experiment on Lichess

Post by Solista »

Some people don't understand what I'm talking about because of IbaiBuR's private Leaderboard list.

I'm not complaining about the list, everyone can do whatever they want. The list suggests to the unsuspecting reader that it is a list by Lichess:

https://lichess.org/team/banned-of-leaderboard-of-bots

Image :shock:

The list is on Lichess, and anyone who clicks on the link will see that there is a list (with the Lichess logo) where players are banned. That's not nice, sorry! The list has nothing to do with Lichess. That's my point of criticism.
User avatar
massimilianogoi
Site Admin
Posts: 396
Joined: Thu Aug 04, 2022 1:42 pm
Has thanked: 604 times
Been thanked: 648 times
Contact:

Re: Engine-Test Experiment on Lichess

Post by massimilianogoi »

Solista wrote: Mon Nov 27, 2023 9:02 pm admete_bot is a weak program with an Elo of around 2400. I have stopped playing against such opponents otherwise I would be ashamed. ProteusSF is of course not ashamed, he would also play 100.000 games against such opponents if he could gain another 100 Elo. Maybe he'll do that at some point. He will never want to play against opponents like me. Other people (including here in the forum) have also reported to me that ProteusSF doesn't accept any challenges against them because they are too strong for him. ProteusSF only requires sacrifices for its Elo rating.

Proof, all bullet games from ProteusSF in PGN sorted by date:
https://pixeldrain.com/u/VFVdxAUR
This is an old trick used by books too: they exploit weakness and even go in negative lines just to reach an error of a certain book and gather points, but with other books the result would be different and even bring to lost games if the opponent architecture is very strong.

Thus it's not a proof of how strong a book is.

I've seen it by buying the opponents books. I'm not making names...
People who have lost the hope.
Solista
Posts: 202
Joined: Thu Jul 06, 2023 11:47 am
Location: Germany
Has thanked: 7 times
Been thanked: 293 times
Contact:

Re: Engine-Test Experiment on Lichess

Post by Solista »

I will soon be closing my news page on my homepage.

Before I close the news page, here's a final word on the reasons mentioned. PlayChess.com maintains official rating lists for computers. Computers are only allowed to play in the separate Engine Room! Here is the current top 100 rating list in Blitz --> Top 10:

Image

I'm in 3rd place. The two leaders cheated and that's proven. I've posted some games as proof. I'm happy with my performance but I'm not bragging about it. There are no computer chess rankings on Lichess. Here is the current rating list of Lichess --> Rating Lichess:

Image

As everyone can see, computers are not included in this list, although there is only one Elo for computers and humans. Computers can also play against humans. Lichess doesn't want to have engines in rankings where humans are. For some computer freaks, this is a reason to create their own private computer list. Everyone is welcome to do this as they wish, but they must not suggest that this is an official Lichess list. But that's exactly what IbaiBuR did: He created his own list with his own rules, whereby the list (Lichess logo and link to Lichess) suggests an official list to readers. Regarding ProteusSF: He has created a total of more than a dozen accounts on Lichess. For what? To brag! Like IbaiBuR, he has created his own private Lichess list in which he takes the first 6 places (crazy haha) --> Rating List by ProteusSF:

Image

Is he that good? No that's not him. I checked all the games on his Top 1 account: He played thousands of games against weak opponents (that's not forbidden), all of which he won. There were only 60 winning games in total against strong opponents. Of these 60 games he won 7 and lost 53. Is such an engine really the best? Nevertheless, he brags about it and constantly posts new photos of his nonsense private rating list. IbaiBuR also does the same thing, almost as a contrast to ProteusSF, and he created his own rules there so that he can ban players like ProteusSF (and me haha). I don't want to be on either of these arrogant and private (!) lists.

Eduard
https://solistachess.jimdosite.com/books/
ZamChess
Posts: 107
Joined: Thu Jul 06, 2023 12:30 pm
Has thanked: 1069 times
Been thanked: 246 times

Re: Engine-Test Experiment on Lichess

Post by ZamChess »

Thank you Eduard for Incognito 4 & 5 ! :sm165:

I wish you all the best.
Solista
Posts: 202
Joined: Thu Jul 06, 2023 11:47 am
Location: Germany
Has thanked: 7 times
Been thanked: 293 times
Contact:

Re: Engine-Test Experiment on Lichess

Post by Solista »

INCOGNITO is CorChess. What has changed is only the network. I love the network with Dimension 3072. Incognito is made for analysis.

I have now also implemented the large network in my private engine Light (hence Big Light) and two of my friends are currently playing with it on PlayChess. I test offline, but also with Ponder ON (~5000 kns/engine).

GUI Fritz 18, openings with color swap, Hash 512 MB, all 3456men, time control 5 min + 0.1s, current standing:

Image

The smaller network with dimension 2560 played weaker.

I don't like posting results without games. That would be unbelievable. That's why I would also like to see games from other testers. I don't believe anyone who doesn't post a PGN (ProteusSF and others).

PGN all games:
https://pixeldrain.com/u/hkwjHwfP
Solista
Posts: 202
Joined: Thu Jul 06, 2023 11:47 am
Location: Germany
Has thanked: 7 times
Been thanked: 293 times
Contact:

Re: Engine-Test Experiment on Lichess

Post by Solista »

IbaiBuR would like to forbid me (and some other players) from playing Lichess against weak opponents. Otherwise he will put us on a banished list. The fact is, server gaming is dying. This is what it looks like on PlayChess today:

Live Picture on PlayChess, you see only draws:
(and there are only a few players playing. This is the future when only your IbaiBuR rules will apply)
Image

The only chance to play against engines other than Stockfish is Lichess. There are really a lot of interesting engines there that are of course much weaker than Stockfish. But I still enjoy playing against these engines more than on PlayChess, where I might only win one game out of 100. A friend of mine had to play 150 games today before he won another game with at least 18000 kns. He plays non-stop in auto mode. I got his latest database with exactly 10723 games. I counted only 164 winning games.

Image

Some of these were gained through disconnect or time. According to IbaiBuR, such balances must now be continued on Lichess, otherwise there is a risk of banishment and discrimination. :thinking:
Solista
Posts: 202
Joined: Thu Jul 06, 2023 11:47 am
Location: Germany
Has thanked: 7 times
Been thanked: 293 times
Contact:

Re: Engine-Test Experiment on Lichess

Post by Solista »

Live image PlayChess: Only 16 players worldwide play in the ChessBase engine room. The reason is the many draws. Why should you play? Almost all games end in a draw, you can play all day and only draw.

Image :dunno:

All engines are Stockfish derivatives!
Solista
Posts: 202
Joined: Thu Jul 06, 2023 11:47 am
Location: Germany
Has thanked: 7 times
Been thanked: 293 times
Contact:

Re: Engine-Test Experiment on Lichess

Post by Solista »

I ran another experiment on Lichess. I wanted to see how quickly I could get first place in the Bullet. To date, ProteusSF was leading the table. But ProteusSF needed 5000 games for its current Elo of 3223 in Bullet. My nephew's account only needed 162 games to be currently top (Elo 3225). ProteusSF boasting is nonsense, it's not hard with appropriate setting and book (It's not the hardware).

TOP on Lichess:
Image

By the way: Engine that played had dimension 3072. The larger network is better against weaker opponents because it make less repetition of moves, which is particularly important in games with black pieces in order to win with black. Only those who win almost all games with black can become top.
Solista
Posts: 202
Joined: Thu Jul 06, 2023 11:47 am
Location: Germany
Has thanked: 7 times
Been thanked: 293 times
Contact:

Re: Engine-Test Experiment on Lichess

Post by Solista »

Watch Lichess Bots live without registering. Just enter this address into the browser:
https://lichess.org/games/bot

Also works with smartphones or tablets (only Browser like Chrome, no Lichess App). My accounts are SolistaBot (tournament mode) and EdoBot (experimental mode). The account in picture KQRBNPkqrbnp belongs to Khalid Omar, for example, who plays with Eman (he has at least two other accounts). All games can be downloaded by anyone as a PGN file. This morning there were 150 bots online. Only 35 people on PlayChess. It looks like so that PlayChess dies.

Live Lichess Bots playing:
Image
User avatar
IbaiBuR
Posts: 29
Joined: Thu Jul 06, 2023 3:30 pm
Location: Spain
Has thanked: 4 times
Been thanked: 29 times
Contact:

Re: Engine-Test Experiment on Lichess

Post by IbaiBuR »

Eduard, why do you have to change my words in every single forum? I just do not understand it. I removed you from the list and the team, which was what you wanted, and you still do not leave me alone. Please, I do not want anything from you. You are starting to be extremely annoying. You posted the same thing in 3 forums plus your homepage.

What is your problem?

I want to make something clear. The point why Lichess (regarding BOTs) is great is because you can play both against strong and weak BOTs, there is a wide variety of self-made engines which are very funny to play with, due to their different style and opening books. I really encourage people to play against this BOTs but also to play against stronger ones!

The Lichess BOT Leaderboard me and some friends created some months ago is just for fun and to have a reference computer chess rating list / ranking at Lichess, that is all. At first all the work was done manually, but after that we created an script that automatically creates the rankings and updates them every 2 hours, source code is here: https://github.com/TheYoBots/libot-lb. It is very simple, it just uses the Lichess API to create the rankings and update the available BOTs list.

For the list, we adopted rules to avoid unfair players to be in it. One way of artificially increasing your rating for example is to only play against BOTs weaker / much weaker than your BOT, so you can win every single game and increase your elo. This is exactly what Eduard did and what he keeps doing. And you know why he started doing it? Because he got so mad (quite a childish behaviour) after he did not get elo points during a game against an "apparently strong" BOT (which by the way uses a RaspBerry PI4, very weak books, and Stockfish 16, not even the dev version).

So, after criticising AlexChess during months, saying that he would stop playing against weak opponents such as admete_bot because he would feel ashamed (lie, he recently played about 20 games with this BOT: https://lichess.org/@/SolistaBot/search ... order=desc), he does the same thing as him, but he cannot be criticised it seems, in his mind, he is the only one playing fair in online servers, it does not matter what he do. He plays against BOTs with 500-600 elo less than his BOTs only, and then claims he is top 1 at bullet, which is nothing but a joke!

Do you know why your nephew's account reached that rating after that short amount of games? It is very easy. At Lichess you start with 2000 elo and a high rating deviation. What does this mean? Well, when your rating deviation is high (usually, above 60), when you win (even if your opponent is weaker) you almost always get elo (unless you play with the same opponent non-stop I believe). If the opponent has a similar rating to you or is stronger, if you win you get a lot of elo, and if you lose you will also lose more elo than normal. Your nephew's account has only played 162 bullet games (the vast majority of them against very weak BOTs) and it currently has a rating deviation of 71 (very high).

If you consider this behaviour as completely honest and fair, fine, have fun with it, but then don't be mad if we ban you, simply ignore us, it is not that difficult!

The list is not official, I said that a trillion times, but you seem to be blind and ignore it. I really want you to leave me alone but you don't stop. I tried to ignore you but after seeing the same thing posted in three forums I think it is already enough, so please stop.
Post Reply