Since you’re here...

We hope you will consider supporting us today. We need your support to continue to exist, because good entries are more and more work time. Every reader contribution, however big or small, is so valuable. Support "Chess Engines Diary" even a small amount– and it only takes a minute. Thank you.
============================== My email: jotes@go2.pl



Amira 1.71 JA - new version chess engine

on
 


Amira - UCI chess engine
Author: FauziAkram

Amira 1.71 what's new?

Amira 1.71 marks a major leap in chess understanding, enhancing the engine's core with a deeply sophisticated evaluation function and a more relentless, efficient search. This version moves beyond simple heuristics to a more granular appreciation of complex positional factors, particularly in pawn and rook endgames.

Key Features in This Release:
🧠 Refined Strategic Engine:
The evaluation model has undergone a significant upgrade, providing Amira with a richer understanding of key strategic elements:

Advanced Passed Pawn Evaluation: The engine's handling of passed pawns is now far more sophisticated. It dynamically assesses passers based on their safety, whether they are blocked, and their proximity to both friendly and enemy kings for a more accurate evaluation of their true potential.
Nuanced Pawn Structure Intelligence: Amira now recognizes and rewards "phalanx" pawn structures (pawns side-by-side) based on their rank and file, encouraging the creation of powerful, mobile pawn chains.
Smarter Rook Play: The engine better understands the power of rooks on the enemy's side of the board. It now specifically rewards rooks on the 7th rank that attack key enemy pawns or restrict the enemy king, with an additional bonus for connected rooks working in tandem.
⚡ More Aggressive & Efficient Search:
The search algorithm has been sharpened with additional pruning techniques, allowing it to explore relevant lines more deeply and discard irrelevant ones faster:

Aggressive SEE Pruning: Static Exchange Evaluation (SEE) is now used to prune moves that are likely to lose material, even at deeper search depths. This allows the engine to quickly discard tactical blunders and focus on promising lines.
Futility Pruning: At shallower depths, Amira now employs futility pruning to cut moves that are very unlikely to raise the score, making the search more efficient in quiet positions.
Optimized Quiescence Search: The quiescence search is now more intelligent, using SEE-based checks to avoid exploring pointless or losing captures, leading to a faster and more accurate tactical search.
Adaptive Time Management: The engine's time management is now more dynamic. It can extend its search time if the evaluation of a position changes drastically ("panic" mode) and can decide to stop sooner if the best move has remained stable for several iterations, saving time for more critical moments.
Refactored Move Generation: The move generator has been streamlined for improved performance and code maintainability.

Jim Ablett compiles:

Comments

Post a Comment