Patricia 3

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Rebel, chrisw

jefk
Posts: 773
Joined: Sun Jul 25, 2010 10:07 pm
Location: the Netherlands
Full name: Jef Kaan

Re: Patricia 3

Post by jefk »

ok about move 21, you (sylwy) give Nd5 for White with P3 at a shallow depth 17 in a sharp position (expecting
...Be4) and then you compare it with the move -thereafter- 21...Bxd5 for Black (with Sf at depth 21).
On my comp P3 is noticing a worse position at move 21 and is considering eg. 21.a3 at depth 20 (*).
If i feed in manually 21.Nd5 (all moves are losing now) it indeed takes some time to see 21...Bxd5
namely about two minutes on my comp(*) at depth 22/23 or so but who cares.
The White position in the game already got worse fast after at move 19.e5?? btw, and although P3 would not
play this (at ply 22 it play Bd3 or Qe3. If i feed in 19.e5? it also sees quick advantage for Black with dxe5 so
it's not that bad actually (although Nf2 as played was better); but then like someone else wrote,
we shouldn't compare it at this tage with something as SF which as we know is one of the -very few-
strongest engines, developed by a complete software team for many years

(*) although i get beyond dpt 17 pl. on my comp in reasonable time , i don't quickly get beyond depth 21 ply
with P3 in this position; wondering if it's using all (my 6) threads.. (my Cpu load in task manager
is not indicating this.. (also based on some/quite a few test games i played it wasnt playing at 3300,
more like 2900 or so but then against some 2600-2700 engines -and own gambit book- i already got many
interesting games; at 3300 it might be more fun but it may be due to my computer/gui config dunno)
clyjspdxk
Posts: 3
Joined: Fri Jul 19, 2024 2:53 pm
Full name: Shujingling

Re: Patricia 3

Post by clyjspdxk »

Archimedes wrote: Mon Jul 15, 2024 4:27 pm
pohl4711 wrote: Mon Jul 15, 2024 4:22 pm Can anybody do and share an Android Binary (arm) of Patricia 3?
Patricia 3 for Android:
https://sourceforge.net/projects/chess- ... p/download

Patricia 3 for Android (OEX):
https://sourceforge.net/projects/chess- ... k/download
Hello, Can you teach me How to set the NNUE in your new Android Patricia 3.1 in Droidfish,or can you embedded it into the engine, please. Thanks
Archimedes
Posts: 157
Joined: Tue Mar 05, 2019 3:43 pm
Full name: Archimedes

Re: Patricia 3

Post by Archimedes »

clyjspdxk wrote: Mon Aug 12, 2024 1:01 pm
Archimedes wrote: Mon Jul 15, 2024 4:27 pm
pohl4711 wrote: Mon Jul 15, 2024 4:22 pm Can anybody do and share an Android Binary (arm) of Patricia 3?
Patricia 3 for Android:
https://sourceforge.net/projects/chess- ... p/download

Patricia 3 for Android (OEX):
https://sourceforge.net/projects/chess- ... k/download
Hello, Can you teach me How to set the NNUE in your new Android Patricia 3.1 in Droidfish,or can you embedded it into the engine, please. Thanks
It is already embedded.
tapio
Posts: 54
Joined: Thu May 30, 2024 10:33 am
Full name: Michael Scheer

Re: Patricia 3

Post by tapio »

Will you use your aggression evaluation algo to create an inverse Patricia? Picture related.

Image
Whiskers
Posts: 219
Joined: Tue Jan 31, 2023 4:34 pm
Full name: Adam Kulju

Re: Patricia 3

Post by Whiskers »

tapio wrote: Sun Aug 18, 2024 1:30 pm Will you use your aggression evaluation algo to create an inverse Patricia? Picture related.

Image
If you're talking about one that minimizes EAS score, I've heard that joke from at least five people already 😭
tapio
Posts: 54
Joined: Thu May 30, 2024 10:33 am
Full name: Michael Scheer

Re: Patricia 3

Post by tapio »

Not joke, with your eas score, you're having an evaluation that could also make it play un-awesome, boring and defensive. Like in table tennis the defensive players, terrible to play.
User avatar
pohl4711
Posts: 2568
Joined: Sat Sep 03, 2011 7:25 am
Location: Berlin, Germany
Full name: Stefan Pohl

Re: Patricia 3

Post by pohl4711 »

tapio wrote: Wed Aug 21, 2024 12:28 pm Not joke, with your eas score, you're having an evaluation that could also make it play un-awesome, boring and defensive. Like in table tennis the defensive players, terrible to play.
Yes, thats true. My EAS-Tool gives just a number to measure aggressiveness of engines. Of course, you can use the EAS-Tool not only for making an engine playing more aggressive, but for making an engine playing more solid (=non aggressive): You just have to make the engine getting a small(er) EAS-Score.
tapio
Posts: 54
Joined: Thu May 30, 2024 10:33 am
Full name: Michael Scheer

Re: Patricia 3

Post by tapio »

Hi, Patricia exposes many variants to the UI, the lower the skill level. Skill level one: 5 variants. Can this be changed?

This is AcidApe Chess on Android.

Image
Whiskers
Posts: 219
Joined: Tue Jan 31, 2023 4:34 pm
Full name: Adam Kulju

Re: Patricia 3

Post by Whiskers »

tapio wrote: Fri Sep 13, 2024 7:25 pm Hi, Patricia exposes many variants to the UI, the lower the skill level. Skill level one: 5 variants. Can this be changed?

This is AcidApe Chess on Android.

Image
What arrow would you want to see? Only the move that is actually played? Currently Patricia decides which move to play only after completing a search and getting some set of moves.
tapio
Posts: 54
Joined: Thu May 30, 2024 10:33 am
Full name: Michael Scheer

Re: Patricia 3

Post by tapio »

Shouldn't the thought process be kept internal? IMO (not an engine expert) the # of variations should be shown, which are set in the GUI. In my case it was one.

Think about it the other way around... you set # of variations to 1 in your chess GUI, but GUI then shows 5 variations. Then you instantly think something is broken.