SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2310.11518 · NeurIPS · 2023

Guarantees for Self-Play in Multiplayer Games via Polymatrix Decomposability

James Wright, Revan Macqueen

arXiv · PDF · Open in the Atlas

Code that ran

2 of 9 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 2 failures of our instrument; 7 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2310.11518.json · badge: /api/ran/2310.11518.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
revanmacqueen/self-play-polymatrix canonical (classifier label)
classifier basis: census
2 ran · 7 no recorded run
FunctionStatusWhere it lives
sample_strats Ran
ran; our contract draft was wrong, not the code
revanmacqueen/self-play-polymatrix/sgdecompose/sgdecompose_utils.py
code served (permissive licence) · get_code("6a1b94b9673edebf")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ArrayImpl, ndim 3, dtype family float, shape (3, 2, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import jax.numpy as np

# Define function arguments
num_players = 3
batch_size = 2
num_terminals = 4

# Create rand_idxs: shape (batch_size, num_players) with indices into strats
# Each entry should be a valid index into the first dimension of strats (after the player dim)
rand_idxs = np.array([
    [0, 1, 0],  # batch 0: player 0 uses strat 0, player 1 uses strat 1, player 2 uses strat 0
    [1, 0, 1]   # batch 1: player 0 uses strat 1, player 1 uses strat 0, player 2 uses strat 1
])

# Create strats: shape (num_players, num_strats, num_terminals)
# We need at least 2 strategy indices per player based on rand_idxs
num_strats = 2
strats = np.array([
    [[0.1, 0.2, 0.3, 0.4], [0.5, 0.3, 0.1, 0.1]],  # player 0: 2 strategies over 4 terminals
    [[0.2, 0.3, 0.25, 0.25], [0.1, 0.4, 0.3, 0.2]],  # player 1: 2 strategies over 4 terminals
    [[0.25, 0.25, 0.25, 0.25], [0.4, 0.1, 0.2, 0.3]]  # player 2: 2 strategies over 4 terminals
])

ARGS = {
    'rand_idxs': rand_idxs,
    'num_players': num_players,
    'batch_size': batch_size,
    'num_terminals': num_terminals,
    'strats': strats
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
jax — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ArrayImpl, ndim 3, dtype family float, shape (3, 2, 4) — shape (3, 2, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
unique Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
revanmacqueen/self-play-polymatrix/sgdecompose/sequence.py
code served (permissive licence) · get_code("28af04455767e75f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

# Construct valid arguments for the unique function
# The function takes a list and returns a list of unique elements

# Create a simple list with some duplicate elements
test_list = [1, 2, 3, 2, 1, 4, 3, 5]

ARGS = {
    "list": test_list
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family int — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
chance_terminal_reach_probs Not yet run
unverified (harvested, no recorded run)
revanmacqueen/self-play-polymatrix/sgdecompose/terminals.py
code served (permissive licence) · get_code("a4f6acece04d00f9")
How it ran

no invocation recorded.

construct_vars Not yet run
unverified (harvested, no recorded run)
revanmacqueen/self-play-polymatrix/sgdecompose/sequence.py
code served (permissive licence) · get_code("841c6361af5f151b")
How it ran

no invocation recorded.

get_terminal_returns Not yet run
unverified (harvested, no recorded run)
revanmacqueen/self-play-polymatrix/sgdecompose/terminals.py
code served (permissive licence) · get_code("1d23d239a4db1528")
How it ran

no invocation recorded.

is_root Not yet run
unverified (harvested, no recorded run)
revanmacqueen/self-play-polymatrix/sgdecompose/sequence.py
code served (permissive licence) · get_code("73a391bea90ed36d")
How it ran

no invocation recorded.

player_terminal_reach Not yet run
unverified (harvested, no recorded run)
revanmacqueen/self-play-polymatrix/sgdecompose/terminals.py
code served (permissive licence) · get_code("47a480a6d65cdd35")
How it ran

no invocation recorded.

tree_norm Not yet run
unverified (harvested, no recorded run)
revanmacqueen/self-play-polymatrix/sgdecompose/sgdecompose_utils.py
code served (permissive licence) · get_code("32d3351455213744")
How it ran

no invocation recorded.

tree_normalize Not yet run
unverified (harvested, no recorded run)
revanmacqueen/self-play-polymatrix/sgdecompose/sgdecompose_utils.py
code served (permissive licence) · get_code("e1a9347c7cfb64de")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 9 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Self-play is a technique for machine learning in multi-agent systems where a learning algorithm learns by interacting with copies of itself. Self-play is useful for generating large quantities of data for learning, but has the drawback that the agents the learner will face post-training may have dramatically different behavior than the learner came to expect by interacting with itself. For the special case of two-player constant-sum games, self-play that reaches Nash equilibrium is guaranteed to produce strategies that perform well against any post-training opponent; however, no such guarantee exists for multiplayer games. We show that in games that approximately decompose into a set of two-player constant-sum games (called constant-sum polymatrix games) where global ϵ-Nash equilibria are boundedly far from Nash equilibria in each subgame (called subgame stability), any no-external-regret algorithm that learns by self-play will produce a strategy with bounded vulnerability. For the first time, our results identify a structural property of multiplayer games that enable performance guarantees for the strategies produced by a broad class of self-play algorithms. We demonstrate our findings through experiments on Leduc poker.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2310.11518")
get_code_for_paper("2310.11518")
have("2310.11518")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2310.11518.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2310.11518.svg)](https://syntology.ai/paper/2310.11518)

Connect an agent — have() is free.