- Blind mode tutorial
lichess.org
Donate

Question: How to use 3rd party engine to modify repertoire

Hello, I am trying to tweak some of my openings with an engine called patriciabot (lichess name) due to its hyper-aggresive style and quick victories against humans. I have a few questions on how to do this best.

I have the engine downloaded to scid vs pc. When I check what it prefers against e4, it chooses e6. On lichess, it’s games indicate it responds with e5 most. Is there a skill or style gap between these two? Which engine should I use?

If the answer is lichess, I want to somehow check lines similar to how analysis board works. Is this possible? If not, I was going to try to use an engine/repertoire book to test my openings against the bot but I read that this is not allowed, even against a bot. Additionally, since the bot does not accept casual challenges, I can not input a FEN to check its response against a certain opening position it does not traditionally play into. Is there a fix for any of this?

If the answer is SCID vs PC, I currently have to transfer the moves from SCID to a lichess study due to my lack of familiarity with the GUI and saving repertoire files in this program. Additionally, it is extremely inefficient in terms of search times and finding moves compared to lichess analysis board. Is there any way to combat either of these problems?

Overall, is there an efficient way to use patriciabot to create/modify an opening repertoire?

I apologize if these questions are naive. I am only used to lichess analysis boards and studies at the moment. As a result, I am completely unfamiliar with downloading/using local engines and chess GUI’s.

Hello, I am trying to tweak some of my openings with an engine called patriciabot (lichess name) due to its hyper-aggresive style and quick victories against humans. I have a few questions on how to do this best. I have the engine downloaded to scid vs pc. When I check what it prefers against e4, it chooses e6. On lichess, it’s games indicate it responds with e5 most. Is there a skill or style gap between these two? Which engine should I use? If the answer is lichess, I want to somehow check lines similar to how analysis board works. Is this possible? If not, I was going to try to use an engine/repertoire book to test my openings against the bot but I read that this is not allowed, even against a bot. Additionally, since the bot does not accept casual challenges, I can not input a FEN to check its response against a certain opening position it does not traditionally play into. Is there a fix for any of this? If the answer is SCID vs PC, I currently have to transfer the moves from SCID to a lichess study due to my lack of familiarity with the GUI and saving repertoire files in this program. Additionally, it is extremely inefficient in terms of search times and finding moves compared to lichess analysis board. Is there any way to combat either of these problems? Overall, is there an efficient way to use patriciabot to create/modify an opening repertoire? I apologize if these questions are naive. I am only used to lichess analysis boards and studies at the moment. As a result, I am completely unfamiliar with downloading/using local engines and chess GUI’s.
<Comment deleted by user>

Hi guy's, I'm fairly new to online chess and infact the game of chess. But I am struggling to understand this blog. Am I right in thinking that using a bot to help you with your game is cheating?
I say this because I am struggling to understand how most players I have played recently have been unbeatable. I understand that there will always be players that will be using bot's (cheating) but what I also can't understand is, it seems to come in waves. When I reach a score of over 1400 my opponents become invincible. My success rate falls to around 5%. Until I reach around 1250 and then it stops and things become more normal, my success rate becomes much more healthier.
It almost seems as though there is something in the system beyond people using bot's. Because this amount of cheating all at once doesn't add up.
Does this make sense?

Hi guy's, I'm fairly new to online chess and infact the game of chess. But I am struggling to understand this blog. Am I right in thinking that using a bot to help you with your game is cheating? I say this because I am struggling to understand how most players I have played recently have been unbeatable. I understand that there will always be players that will be using bot's (cheating) but what I also can't understand is, it seems to come in waves. When I reach a score of over 1400 my opponents become invincible. My success rate falls to around 5%. Until I reach around 1250 and then it stops and things become more normal, my success rate becomes much more healthier. It almost seems as though there is something in the system beyond people using bot's. Because this amount of cheating all at once doesn't add up. Does this make sense?

@Likestoplay63 said in #3:

Hi guy's, I'm fairly new to online chess and infact the game of chess. But I am struggling to understand this blog. Am I right in thinking that using a bot to help you with your game is cheating?

Hey it's better to open your own threat, just switching to your topic is kind of the online equivalent of joining a conversation and just ignoring the previous speakers maybe a bit worse because of the missing real life cues.

You can just copy your post and "create new topic" in the general Chess discussion section.

You'll get plenty of hopefully nice, answers that way.

@Likestoplay63 said in #3: > Hi guy's, I'm fairly new to online chess and infact the game of chess. But I am struggling to understand this blog. Am I right in thinking that using a bot to help you with your game is cheating? Hey it's better to open your own threat, just switching to your topic is kind of the online equivalent of joining a conversation and just ignoring the previous speakers maybe a bit worse because of the missing real life cues. You can just copy your post and "create new topic" in the general Chess discussion section. You'll get plenty of hopefully nice, answers that way.

Patricia bot sounds great. Please invite me to patricia-pate!

Patricia bot sounds great. Please invite me to patricia-pate!

Hi, Thank you for your reply.

I think you might be missing the point. My point being. Seeing that you are discussing the use of bot's on this chess site is telling me that everyone is using bot's and thinks it's quite normal to do so.

Is this so?

Because if it is, this would mean that everyone's score is purely artificial. And further more, what on earth is the point of playing chess with a computer making your every move. This is not chess, this is ridiculous!

Hi, Thank you for your reply. I think you might be missing the point. My point being. Seeing that you are discussing the use of bot's on this chess site is telling me that everyone is using bot's and thinks it's quite normal to do so. Is this so? Because if it is, this would mean that everyone's score is purely artificial. And further more, what on earth is the point of playing chess with a computer making your every move. This is not chess, this is ridiculous!

@rajma420 I think using Patricia or any other engine to construct or tweak your repertoire (the way you plan to do by looking what moves Patricia prefers) is most likely not possible. Patricia has been trained to be aggressive in the sense that it is trained so that it gets a higher 'Engine Aggressiveness Score' (EAS), as measured by a certain EAS tool that collects statistics of number of sacrifices, quick wins, and so on. But I don't think Patricia would prefer particularly aggressive moves in the opening. More likely you would see its aggressive behaviour in the middlegame. So it is not useful to base your repertoire on what choices Patricia makes in the opening. If you want aggressive repertoire, it may be easier to look at some repertoire books that propose more aggressive lines. Another possibility is to look at some players who play aggressively and use the lines in their games to make your repertoire. (This may be automated.) But again I am not sure if this is very useful. For example, if Carlsen says Arjun plays like madman, maybe only at a very high level Arjun's repertoire looks like madman's but at amateur level it may all look normal. I don't use Patricia but I follow the discussions about its development on the Talk Chess forum, where all matters of computer chess are discussed. You can ask the Patricia author there.

Regarding Scid vs PC - you may want to explore it more. For example, check if your Scid vs PC is configured to use an opening book. The program comes with 4 pre-installed books, and you can construct or download and install more books. The pre-installed books are normal books, not particularly aggressive or anything. If an engine is configured to use a book, it is actually the GUI that makes a book move as long as a position is in the book. Only when the game leaves the book, the GUI starts using the engine for moves. It means, as long as you are in the opening phase, it does not matter what engine you are using. You can install a more aggressive book, e.g., you can find a gambit book somewhere on the internet, and the engine (actually the GUI) will play the gambits.

Edit: in Scid vs PC it is possible to do automated bulk analysis of games. Also, for opening phase this site https://chessdb.cn/queryc_en/ has deep analysis and evaluations of opening phase with powerful computers and engines.

@Likestoplay63 OP is not talking about using 'bots' during a game, but talking about using 'bots' for their home preparation.

@rajma420 I think using Patricia or any other engine to construct or tweak your repertoire (the way you plan to do by looking what moves Patricia prefers) is most likely not possible. Patricia has been trained to be aggressive in the sense that it is trained so that it gets a higher 'Engine Aggressiveness Score' (EAS), as measured by a certain EAS tool that collects statistics of number of sacrifices, quick wins, and so on. But I don't think Patricia would prefer particularly aggressive moves in the opening. More likely you would see its aggressive behaviour in the middlegame. So it is not useful to base your repertoire on what choices Patricia makes in the opening. If you want aggressive repertoire, it may be easier to look at some repertoire books that propose more aggressive lines. Another possibility is to look at some players who play aggressively and use the lines in their games to make your repertoire. (This may be automated.) But again I am not sure if this is very useful. For example, if Carlsen says Arjun plays like madman, maybe only at a very high level Arjun's repertoire looks like madman's but at amateur level it may all look normal. I don't use Patricia but I follow the discussions about its development on the Talk Chess forum, where all matters of computer chess are discussed. You can ask the Patricia author there. Regarding Scid vs PC - you may want to explore it more. For example, check if your Scid vs PC is configured to use an opening book. The program comes with 4 pre-installed books, and you can construct or download and install more books. The pre-installed books are normal books, not particularly aggressive or anything. If an engine is configured to use a book, it is actually the GUI that makes a book move as long as a position is in the book. Only when the game leaves the book, the GUI starts using the engine for moves. It means, as long as you are in the opening phase, it does not matter what engine you are using. You can install a more aggressive book, e.g., you can find a gambit book somewhere on the internet, and the engine (actually the GUI) will play the gambits. Edit: in Scid vs PC it is possible to do automated bulk analysis of games. Also, for opening phase this site https://chessdb.cn/queryc_en/ has deep analysis and evaluations of opening phase with powerful computers and engines. @Likestoplay63 OP is not talking about using 'bots' during a game, but talking about using 'bots' for their home preparation.

Thank you for your explanation Kajalmaya, I do hope you're right! I must admit to thinking it was an easy access online bot that was being discussed. My bad!

Thank you for your explanation Kajalmaya, I do hope you're right! I must admit to thinking it was an easy access online bot that was being discussed. My bad!

I doubt you are using the best free available GUI with SCID.

Currently Nimzo_3d, the upgraded Tarrasch GUI and Lucas Chess are the best free GUIs available. Lucas Chess even comes with Patricia preinstalled.

I doubt you are using the best free available GUI with SCID. Currently Nimzo_3d, the upgraded Tarrasch GUI and Lucas Chess are the best free GUIs available. Lucas Chess even comes with Patricia preinstalled.

@kajalmaya Thank you for the in-depth response! I definitely learned a lot from this message. I also noticed patriciabot prefers less crazy aggressive openings when playing, it has a preference for the ruy lopez, which in my (limited) experience is less aggressive than other options. I was thinking to put the bot in an already aggressive position like a kings gambit, classical dutch, etc. 5-6 moves deep and see the ideas it uses. I noticed in early middlegames and late openings patriciabot sometimes loses 25-100 centipawns (according to stockfish) which allows it to gain a strong initiative, which is what I’m going for. I’ll definitely learn more about these features on SCID, thank you for the ideas.

@kajalmaya Thank you for the in-depth response! I definitely learned a lot from this message. I also noticed patriciabot prefers less crazy aggressive openings when playing, it has a preference for the ruy lopez, which in my (limited) experience is less aggressive than other options. I was thinking to put the bot in an already aggressive position like a kings gambit, classical dutch, etc. 5-6 moves deep and see the ideas it uses. I noticed in early middlegames and late openings patriciabot sometimes loses 25-100 centipawns (according to stockfish) which allows it to gain a strong initiative, which is what I’m going for. I’ll definitely learn more about these features on SCID, thank you for the ideas.

This topic has been archived and can no longer be replied to.