SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2002.06673 · ICML · 2020

Performative Prediction

Juan Perdomo, Tijana Zrnic, Celestine Mendler-Dünner, Moritz Hardt

arXiv · PDF · Open in the Atlas

Code that ran

5 of 10 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 4 have no contract verdict; 1 failure of our instrument; 5 no recorded run), measured 2026-11-21; 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/2002.06673.json · badge: /api/ran/2002.06673.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
mrtzh/whynot community (classifier label)
classifier basis not recorded on the link
5 ran · 5 no recorded run
FunctionStatusWhere it lives
check_and_cast_args Ran
ran on a synthesized input
mrtzh/whynot/whynot/causal_graphs.py
code served (permissive licence) · get_code("ddf30472cf64bb94")
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. The image it ran in was not recorded. If your copy returns the same type 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
import numpy as np
from collections.abc import Iterable

torch.manual_seed(0)

# The function check_and_cast_args takes a single parameter 'args'
# It can be either:
# 1. An Iterable (list, tuple, etc.) of items
# 2. A single item
#
# The function checks and casts each argument:
# - Raises ValueError if arg is a string
# - Converts int to float
# - Returns the arg otherwise
#
# Valid test cases would include:
# - A list of integers, floats, and tensors
# - A single integer, float, or tensor
# - An iterable of mixed types

# Test case: list with integers, floats, and tensors
test_args = [
    1,  # int - should be cast to float
    2.5,  # float - should pass through
    torch.tensor([1.0, 2.0, 3.0]),  # tensor - should pass through
    0,  # another int - should be cast to float
]

ARGS = {
    "args": test_args
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list — a list output has no array shape of its own; none is stated
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
compute_ate Ran
ran on a synthesized input
mrtzh/whynot/whynot/algorithms/propensity_score_matching.py
code served (permissive licence) · get_code("5ad5c75f160ddf41")
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 float32, ndim 0, dtype family float, shape () 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 numpy as np

np.random.seed(0)

# Create small sample data
n_samples = 10

# outcome: array of outcome values
outcome = np.random.randn(n_samples).astype(np.float32)

# treatment: binary treatment assignment (0 or 1)
treatment = np.array([0.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0, 0.0, 1.0], dtype=np.float32)

# weights: positive weights for each sample
weights = np.ones(n_samples, dtype=np.float32)

ARGS = {
    "outcome": outcome,
    "treatment": treatment,
    "weights": weights,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float32, ndim 0, dtype family float, shape () — shape () 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
dataclass_to_box Ran
ran on a synthesized input
mrtzh/whynot/whynot/causal_graphs.py
code served (permissive licence) · get_code("27663bc69d6c1aea")
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 tuple, ndim 1, dtype family float; 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 dataclasses
from autograd.tracer import trace_stack, new_box
import numpy as np

# Define a minimal dataclass for testing
@dataclasses.dataclass
class SimpleState:
    x: float
    y: float
    z: float

# Define a dataclass with a get_boxable/set_boxable interface
@dataclasses.dataclass
class BoxableState:
    value: float
    
    def get_boxable(self):
        return self.value
    
    def set_boxable(self, box_value):
        self.value = box_value

# Create a test dataclass instance
test_dataclass = SimpleState(x=1.0, y=2.0, z=3.0)

# Create a test dataclass with boxable field
test_dataclass_boxable = dataclasses.replace(
    test_dataclass,
    x=BoxableState(value=1.5)
)

# Use the trace_stack context manager as required
trace_context = trace_stack.new_trace()

ARGS = {
    "dataclass": test_dataclass,
    "trace": trace_context,
    "name_suffix": "test",
    "skip_args": ["z"]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
autograd, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple 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
get_matched_dataset Ran
ran on a synthesized input
mrtzh/whynot/whynot/algorithms/propensity_score_matching.py
code served (permissive licence) · get_code("1a7197978bda289e")
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 tuple, ndim 2, dtype family float; 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 numpy as np

# Set random seed for reproducibility
np.random.seed(0)

# Create small sample dataset
num_samples = 10
num_treated = 4
num_control = 6

# Create binary treatment array
treatment = np.array([1.0, 1.0, 1.0, 1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0])

# Create propensity scores
scores = np.array([0.8, 0.75, 0.85, 0.9, 0.1, 0.15, 0.2, 0.05, 0.12, 0.18])

# Create outcome array
outcome = np.array([5.2, 4.8, 5.5, 6.1, 1.2, 1.5, 1.8, 0.9, 1.3, 1.6])

# Number of neighbors for matching
num_neighbors = 2

ARGS = {
    "treatment": treatment,
    "scores": scores,
    "outcome": outcome,
    "num_neighbors": num_neighbors,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, sklearn — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 2, dtype family float — a tuple 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
register Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
mrtzh/whynot/whynot/gym/envs/registration.py
code served (permissive licence) · get_code("2718f2328806addb")
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 EnvSpec. The image it ran in was not recorded. If your copy returns the same type 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 gymnasium
from gymnasium.envs.registration import EnvSpec

# Create a simple mock environment class to use as entry point
class MockEnv:
    def __init__(self, param1=None, param2=None):
        self.param1 = param1
        self.param2 = param2
        self.spec = None
    
    @property
    def unwrapped(self):
        return self

# Define arguments for the register function
ARGS = {
    'id': 'MockEnv-v0',
    'entry_point': MockEnv,
    'max_episode_steps': 100,
    'reward_threshold': 95.0,
    'kwargs': {'param1': 'value1', 'param2': 42}
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
gymnasium, whynot — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EnvSpec — a EnvSpec output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a EnvSpec output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
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
extract_params Not yet run
unverified (harvested, no recorded run)
mrtzh/whynot/whynot/framework.py
code served (permissive licence) · get_code("5207cda67d97bf0e")
How it ran

no invocation recorded.

make Not yet run
unverified (harvested, no recorded run)
mrtzh/whynot/whynot/gym/envs/registration.py
code served (permissive licence) · get_code("98c1773aa7ded37f")
How it ran

no invocation recorded.

parallelize Not yet run
unverified (harvested, no recorded run)
mrtzh/whynot/whynot/utils.py
code served (permissive licence) · get_code("8a4d20789854051c")
How it ran

no invocation recorded.

parameter Not yet run
unverified (harvested, no recorded run)
mrtzh/whynot/whynot/framework.py
code served (permissive licence) · get_code("caf87616e0180c33")
How it ran

no invocation recorded.

spec Not yet run
unverified (harvested, no recorded run)
mrtzh/whynot/whynot/gym/envs/registration.py
code served (permissive licence) · get_code("bfa15a965f084103")
How it ran

no invocation recorded.

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 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.

FunctionAlso present inEarliest-datedOthers named
register 1 other paper
1 dated earlier
1810.12282 (2018-10)
Assessing Generalization in Deep Reinforcement Learning

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 2026-11-21.

Repositories linked to this paper

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

Abstract

When predictions support decisions they may influence the outcome they aim to predict. We call such predictions performative; the prediction influences the target. Performativity is a well-studied phenomenon in policy-making that has so far been neglected in supervised learning. When ignored, performativity surfaces as undesirable distribution shift, routinely addressed with retraining. We develop a risk minimization framework for performative prediction bringing together concepts from statistics, game theory, and causality. A conceptual novelty is an equilibrium notion we call performative stability. Performative stability implies that the predictions are calibrated not against past outcomes, but against the future outcomes that manifest from acting on the prediction. Our main results are necessary and sufficient conditions for the convergence of retraining to a performatively stable point of nearly minimal loss. In full generality, performative prediction strictly subsumes the setting known as strategic classification. We thus also give the first sufficient conditions for retraining to overcome strategic feedback effects. * Equal contribution.

For agents

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

get_harvested_code_for_paper("2002.06673")
get_code_for_paper("2002.06673")
have("2002.06673")

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

curl https://syntology.ai/api/ran/2002.06673.json

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

[![Syntology run record](https://syntology.ai/api/ran/2002.06673.svg)](https://syntology.ai/paper/2002.06673)

Connect an agent — have() is free.