History_second_order

[AliceRoselia]

Passed STC:
LLR: 2.97 (-2.94,2.94) <0.00,2.00>
Total: 72864 W: 19060 L: 18690 D: 35114
Ptnml(0-2): 208, 8420, 18832, 8738, 234
https://tests.stockfishchess.org/tests/view/6897e839f8a258623dda6e73

Passed LTC:
LLR: 2.95 (-2.94,2.94) <0.50,2.50>
Total: 52890 W: 13807 L: 13456 D: 25627
Ptnml(0-2): 29, 5643, 14752, 5990, 31
https://tests.stockfishchess.org/tests/live_elo/6897c5eaf8a258623dda6dfb

However, this patch involves quite a large tune.
The untuned version passed STC:
LLR: 2.93 (-2.94,2.94) <0.00,2.00>
Total: 38176 W: 10103 L: 9785 D: 18288
Ptnml(0-2): 121, 4371, 9803, 4655, 138
https://tests.stockfishchess.org/tests/view/6896da31f8a258623dda69d8
The untuned version failed LTC:
LLR: -2.94 (-2.94,2.94) <0.50,2.50>
Total: 97284 W: 24948 L: 24991 D: 47345
Ptnml(0-2): 54, 10576, 27418, 10547, 47
https://tests.stockfishchess.org/tests/live_elo/6896dfc7f8a258623dda69e1

Currently, removing the second-order history while maintaining the tuned value seems to be about neutral at LTC. The VLTC could either pass or fail. This suggests a very complex relationship.

bench: 3068005

What do you think? If it were me, I would merge the patch and wait for the VVLTC tune before attempting to simplify away this second-order history.

[github-actions]

clang-format 20 needs to be run on this PR.
If you do not have clang-format installed, the maintainer will run it when merging.
For the exact version please see https://packages.ubuntu.com/plucky/clang-format-20.

(execution 16859129105 / attempt 1)

[vondele]

CI:

Error: 

Test Suite: TestSyzygy
    Running test_syzygy_setup... 
    ✓ test_syzygy_setup (1092.34ms)
    Running test_syzygy_bench... 
    ✗ test_syzygy_bench (1883.26ms)
    File "/home/runner/work/Stockfish/Stockfish/tests/testing.py", line 221, in __run_test_method
      test_instance.afterEach()
    File "/home/runner/work/Stockfish/Stockfish/src/../tests/instrumented.py", line 425, in afterEach
      assert postfix_check(self.stockfish.get_output()) == True
    ⎯⎯⎯⎯⎯OUTPUT⎯⎯⎯⎯⎯
    history.h:89:112: runtime error: signed integer overflow: 1430416 * 2603 cannot be represented in type 'int'
    history.h:89:112: runtime error: signed integer overflow: 784996 * -2873 cannot be represented in type 'int'
    history.h:89:112: runtime error: signed integer overflow: 1106704 * -2114 cannot be represented in type 'int'
    info depth 3 seldepth 4 multipv 1 score cp 40 nodes 159 nps 53000 hashfull 0 tbhits 0 time 3 pv e2e4
    history.h:89:112: runtime error: signed integer overflow: 1085764 * -2195 cannot be represented in type 'int'
    info depth 4 seldepth 6 multipv 1 score cp 37 nodes 577 nps 57700 hashfull 0 tbhits 0 time 10 pv e2e4 e7e5 g1f3 b8c6
    info depth 5 seldepth 6 multipv 1 score cp 37 nodes 690 nps 57500 hashfull 0 tbhits 0 time 12 pv e2e4 c7c5 b1c3 b8c6
    info depth 6 seldepth 9 multipv 1 score cp 35 nodes 1369 nps 65190 hashfull 0 tbhits 0 time 21 pv e2e4 c7c5 b1c3 b8c6
    ```

[AliceRoselia]

CI:

Error: 

Test Suite: TestSyzygy
    Running test_syzygy_setup... 
    ✓ test_syzygy_setup (1092.34ms)
    Running test_syzygy_bench... 
    ✗ test_syzygy_bench (1883.26ms)
    File "/home/runner/work/Stockfish/Stockfish/tests/testing.py", line 221, in __run_test_method
      test_instance.afterEach()
    File "/home/runner/work/Stockfish/Stockfish/src/../tests/instrumented.py", line 425, in afterEach
      assert postfix_check(self.stockfish.get_output()) == True
    ⎯⎯⎯⎯⎯OUTPUT⎯⎯⎯⎯⎯
    history.h:89:112: runtime error: signed integer overflow: 1430416 * 2603 cannot be represented in type 'int'
    history.h:89:112: runtime error: signed integer overflow: 784996 * -2873 cannot be represented in type 'int'
    history.h:89:112: runtime error: signed integer overflow: 1106704 * -2114 cannot be represented in type 'int'
    info depth 3 seldepth 4 multipv 1 score cp 40 nodes 159 nps 53000 hashfull 0 tbhits 0 time 3 pv e2e4
    history.h:89:112: runtime error: signed integer overflow: 1085764 * -2195 cannot be represented in type 'int'
    info depth 4 seldepth 6 multipv 1 score cp 37 nodes 577 nps 57700 hashfull 0 tbhits 0 time 10 pv e2e4 e7e5 g1f3 b8c6
    info depth 5 seldepth 6 multipv 1 score cp 37 nodes 690 nps 57500 hashfull 0 tbhits 0 time 12 pv e2e4 c7c5 b1c3 b8c6
    info depth 6 seldepth 9 multipv 1 score cp 35 nodes 1369 nps 65190 hashfull 0 tbhits 0 time 21 pv e2e4 c7c5 b1c3 b8c6
    ```

How on earth did it even pass fishtest? It had the undefined behavior all along?

My bad. Let's close this PR. I'm going to test the fixed version.