<!-- Credit Taylor (gtmcknight): solved policy, Gaussian filter, qm tables and harness. -->
# qt1 engine baseline — eb-b

Credit: Taylor (gtmcknight), author of the solved policy, filter, qm tables and pinned harness. Qt1 is unchanged. All new measurements below are local engine observations, not leaderboard scores.

## Setup

| Step | Result |
|---|---|
| Clone amm-challenge | PASS; commit `6b6921f8d463314aa7c96234e1be25a25245f704` |
| Clone Taylor amm-sim, side by side | PASS; commit `4812a6382311624e1506aca131ca35c378f930d9` |
| Run `sh ../amm-sim/setup-fast.sh` once from amm-challenge | PASS, exit 0; Linux x86_64 |
| Installed engine, Python environment and compiler | PASS; setup installed pyrevm 0.3.7, py-solc-x 2.1.0 and solc 0.8.24 |
| Instrumented `.so` hash | PASS; `ff88b6353c2aad7342f0a26194f7347e78b9c3b38d683a7449006d423978cd14` |
| Wheel SHA256 | PASS; downloaded release wheel independently hashed to `aedb68280aca34c110968f00148387c95885207913fb2cf19a1ab4c5cfd6887a`, matching the supplied expected hash |
| StarterStrategy validation | PASS |
| Extract qt1 from Q | PASS; 42,990 bytes; SHA256 `c68887cff61180e12ba2b1ceb333edf76a830f91f63b5d775a8806fb9ec08aca` |

Pinned sources: [setup](https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/setup-fast.sh), [harness](https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/harness.py), [Q](https://api.imdnotdead.fun/artifacts/8a6e303b662e1a5ea9bb47f31acedd3988ef65b3ddb66d074517aa59ad41ca40). Hash checks and measurements are this run's local observations.

Setup output:
```
amm_sim_rs.abi3.so sha256 ff88b6353c2aad7342f0a26194f7347e78b9c3b38d683a7449006d423978cd14
Validating strategy...
Compiling strategy...
Strategy 'StarterStrategy' validated successfully!
```

## Validate

`amm-match validate contracts/src/Qt1.sol`, exit 0:
```
Validating strategy...
Compiling strategy...
Strategy 'TaylorGaussianQMSpline' validated successfully!
```

Deployed runtime: **20,605 bytes**, measured from `SolidityCompiler.compile(...).deployed_bytecode` in every chunk. Compiler: solc 0.8.24, optimizer enabled with 200 runs, viaIR, Paris, per the [pinned compiler](https://raw.githubusercontent.com/benedictbrady/amm-challenge/6b6921f8d463314aa7c96234e1be25a25245f704/amm_competition/evm/compiler.py). Below the 24,576-byte cap.

## Engine

Qt1 alone versus the [30 bps VanillaStrategy](https://raw.githubusercontent.com/benedictbrady/amm-challenge/6b6921f8d463314aa7c96234e1be25a25245f704/contracts/src/VanillaStrategy.sol), four disjoint chunks in ascending seed order, eight workers. Each seed executed exactly once. Each chunk calls `gas_stats_reset()`, `run_batch(qt1, vanilla, configs, 8)`, then `gas_stats()`, through Taylor's harness `run` helper. Edges are the returned `r.edges["submission"]`, not subtraction of the normalizer's edge. SE is sample standard deviation / sqrt(128).

Configs use the exact [MatchRunner._build_configs](https://raw.githubusercontent.com/benedictbrady/amm-challenge/6b6921f8d463314aa7c96234e1be25a25245f704/amm_competition/competition/match.py) body with only its loop range replaced by the chunk's range. Each seed independently initializes `np.random.default_rng(seed=i)`; skipping the unused prefix preserves all requested configurations. The unmodified harness CLI repeats A by default; its `main` was therefore bypassed to respect the no-rerun requirement. The wrapper appears under Commands.

Gas includes only role `Submission`. Means are total `sum_ok_gas` / total successful calls, weighted across chunks. All calls succeeded; all submission OOG, revert, other_halt, tx_error and bad_return counters were zero. Chunk times measure `run_batch` wall time and exclude compilation and config construction.

```json
{
  "strategy": "qt1",
  "seeds": "9800128..9800255",
  "n": 128,
  "mean": 513.5049728806645,
  "se": 8.588810151732844,
  "edges": {
    "9800128": 575.0714848105139,
    "9800129": 475.6252019871907,
    "9800130": 719.1114206940086,
    "9800131": 384.7419453844864,
    "9800132": 467.0871941428443,
    "9800133": 353.00257182483097,
    "9800134": 715.1251631716032,
    "9800135": 520.415370372429,
    "9800136": 545.800705131955,
    "9800137": 611.7712441276748,
    "9800138": 428.9693149517274,
    "9800139": 440.31575844331854,
    "9800140": 435.4192025416303,
    "9800141": 406.61565077703887,
    "9800142": 562.1980337163214,
    "9800143": 473.6185355387734,
    "9800144": 464.40493254179046,
    "9800145": 529.9043438287505,
    "9800146": 581.5862406018566,
    "9800147": 562.3226667411282,
    "9800148": 644.0074698913195,
    "9800149": 619.7617685873588,
    "9800150": 364.7783112924784,
    "9800151": 369.5664820054741,
    "9800152": 553.0100479069719,
    "9800153": 781.3041907881907,
    "9800154": 457.2947998756057,
    "9800155": 499.2091442130844,
    "9800156": 604.2197827572509,
    "9800157": 437.64190310681965,
    "9800158": 403.4876617499161,
    "9800159": 367.3968925116981,
    "9800160": 684.1071262766168,
    "9800161": 550.2630834626159,
    "9800162": 504.12051937167337,
    "9800163": 435.10386712585307,
    "9800164": 608.6606485574624,
    "9800165": 500.10828599613177,
    "9800166": 450.8811356712151,
    "9800167": 460.7547314380382,
    "9800168": 613.5160145379806,
    "9800169": 627.151588010116,
    "9800170": 481.42549042748476,
    "9800171": 560.5163089072701,
    "9800172": 648.4826641982822,
    "9800173": 560.9907156986683,
    "9800174": 403.7459092854356,
    "9800175": 452.9330943647069,
    "9800176": 455.3403295175043,
    "9800177": 520.1346256073443,
    "9800178": 441.59021813505257,
    "9800179": 430.8419948980403,
    "9800180": 554.0530295098395,
    "9800181": 618.7110608285203,
    "9800182": 561.4411239531878,
    "9800183": 408.914505376551,
    "9800184": 542.4436205914367,
    "9800185": 672.5279709071236,
    "9800186": 449.8774118461121,
    "9800187": 597.0975276658238,
    "9800188": 522.7024831101515,
    "9800189": 495.82559090063614,
    "9800190": 448.11696161445565,
    "9800191": 562.736690479302,
    "9800192": 605.5488419214121,
    "9800193": 528.2131339185435,
    "9800194": 642.0842107545084,
    "9800195": 454.15263630609684,
    "9800196": 414.3309887074359,
    "9800197": 598.6413669983391,
    "9800198": 494.35622067231185,
    "9800199": 673.9990550489474,
    "9800200": 611.2165386208662,
    "9800201": 442.72699914946054,
    "9800202": 763.2511209594478,
    "9800203": 704.5716953724557,
    "9800204": 433.67109252252794,
    "9800205": 424.02722524419545,
    "9800206": 366.7709245477803,
    "9800207": 502.2541491720047,
    "9800208": 508.8008937576638,
    "9800209": 628.4881391885355,
    "9800210": 409.09390837469755,
    "9800211": 391.06110151486394,
    "9800212": 612.1762562631736,
    "9800213": 450.64933861409696,
    "9800214": 355.3722772380656,
    "9800215": 555.1633784577178,
    "9800216": 574.9332889100654,
    "9800217": 643.6282044857821,
    "9800218": 398.2043121412238,
    "9800219": 471.20892993988355,
    "9800220": 386.44425251687704,
    "9800221": 398.48372483311755,
    "9800222": 427.865248075649,
    "9800223": 472.9893901384924,
    "9800224": 505.6488652285993,
    "9800225": 542.3764764326198,
    "9800226": 598.5634230283115,
    "9800227": 632.2864819446598,
    "9800228": 526.4278865190308,
    "9800229": 456.51210815047847,
    "9800230": 421.29314959211376,
    "9800231": 480.09868107240345,
    "9800232": 443.74011792648594,
    "9800233": 561.1104954180284,
    "9800234": 500.8134454375696,
    "9800235": 514.2266115612928,
    "9800236": 475.3821034560523,
    "9800237": 376.0013191122208,
    "9800238": 420.1541985061714,
    "9800239": 558.5611875386364,
    "9800240": 607.6922047088852,
    "9800241": 497.9098430774859,
    "9800242": 424.6557555702577,
    "9800243": 355.7695415789743,
    "9800244": 644.2760493799045,
    "9800245": 393.52787582975395,
    "9800246": 628.8802401301547,
    "9800247": 492.24282154612155,
    "9800248": 436.29791281502463,
    "9800249": 726.0195078086158,
    "9800250": 515.1534085457338,
    "9800251": 589.9640685211443,
    "9800252": 554.9376305231932,
    "9800253": 473.24060667512407,
    "9800254": 378.32455109597794,
    "9800255": 472.26735134111334
  },
  "gas": {
    "afterSwap": {
      "calls": 1371375,
      "oog": 0,
      "peak": 209112,
      "mean": 138190.15968790447
    },
    "afterInitialize": {
      "calls": 128,
      "oog": 0,
      "peak": 90858,
      "mean": 90858.0
    }
  },
  "peak_per_chunk": [
    208981,
    204494,
    209112,
    208981
  ],
  "wall_s_per_chunk": [
    22.39,
    23.252,
    24.575,
    20.795
  ]
}
```

## Commands

All commands ran inside the workspace. Checkouts and temporary state were confined to the explicitly permitted disposable `test/scratch/` tree. `C` below denotes `test/scratch/amm-challenge`; commands using `scratch/env.sh` ran with C as the working directory. Shell redirections only save local logs. No engine seeds outside the specified range were run. No command took eight minutes.

The command inventory (including read-only inspections and output generation) follows. Long Python heredocs are described by their complete operational content or supplied verbatim below.

```sh
pwd
rg --files -g '!lib/**' -g '!node_modules/**' | head -40
mkdir -p test/scratch artifacts/goal/batches/e1-engine-base/items/eb-b
git clone https://github.com/benedictbrady/amm-challenge test/scratch/amm-challenge
git -C test/scratch/amm-challenge checkout 6b6921f8d463314aa7c96234e1be25a25245f704
git clone https://github.com/gtmcknight/amm-sim test/scratch/amm-sim
git -C test/scratch/amm-sim checkout 4812a6382311624e1506aca131ca35c378f930d9
cat test/scratch/amm-sim/setup-fast.sh
sed -n '1,260p' test/scratch/amm-sim/harness.py
# from C:
sh ../amm-sim/setup-fast.sh > ../setup.log 2>&1
# from repository root:
curl -Lf https://api.imdnotdead.fun/artifacts/8a6e303b662e1a5ea9bb47f31acedd3988ef65b3ddb66d074517aa59ad41ca40 -o test/scratch/Q.txt
sed -n '1,65p' test/scratch/Q.txt
cat test/scratch/setup.log
rg -n 'Strategy.sol|```solidity' test/scratch/Q.txt
cat test/scratch/amm-challenge/amm_competition/evm/compiler.py | head -200
rg -n '_build_configs' test/scratch/amm-challenge/amm_competition/competition/match.py
sed -n '100,135p' test/scratch/amm-challenge/amm_competition/competition/match.py
# python3 heredoc: extract block following ## Strategy.sol and ```solidity;
# assert its length and SHA256, write unchanged bytes to C/contracts/src/Qt1.sol;
# print length, hash and first 500 characters.
curl -LsSf https://github.com/gtmcknight/amm-sim/releases/download/sim-6b6921f8/amm_sim_rs-0.1.0.post1-cp310-abi3-manylinux2014_x86_64.manylinux_2_17_x86_64.whl -o test/scratch/amm-challenge/scratch/engine.whl
sha256sum test/scratch/amm-challenge/scratch/engine.whl
# from C:
. scratch/env.sh && amm-match validate contracts/src/Qt1.sol > scratch/validate.log 2>&1
# from root:
sed -n '130,150p' test/scratch/amm-challenge/amm_competition/competition/match.py
# cat heredoc creates C/scratch/baseline.py, reproduced below.
# from C, each once:
. scratch/env.sh && python scratch/baseline.py 9800128 9800160 > scratch/chunk-9800128.log 2>&1
. scratch/env.sh && python scratch/baseline.py 9800160 9800192 > scratch/chunk-9800160.log 2>&1
. scratch/env.sh && python scratch/baseline.py 9800192 9800224 > scratch/chunk-9800192.log 2>&1
. scratch/env.sh && python scratch/baseline.py 9800224 9800256 > scratch/chunk-9800224.log 2>&1
# read-only inspections from root:
cat test/scratch/amm-challenge/scratch/validate.log
sed -n '1,85p' test/scratch/amm-challenge/amm_competition/evm/baseline.py
head -20 test/scratch/amm-challenge/contracts/src/VanillaStrategy.sol
git status --short
head -c 500 test/scratch/amm-challenge/scratch/chunk-9800128.log
ls test/scratch/amm-challenge/scratch/chunk*
# from C:
cat scratch/chunk-9800128.json
```

Python wrapper (`cat > C/scratch/baseline.py`):
```python
# Credit Taylor (gtmcknight): pinned harness.py and solved qt1 policy.
import sys, inspect, textwrap, json, time, hashlib
from pathlib import Path
sys.path.insert(0,str(Path('../amm-sim').resolve()))
import harness as h
import amm_sim_rs
from amm_competition.evm.compiler import SolidityCompiler
from amm_competition.competition import match
s0,s1=map(int,sys.argv[1:3])
out=Path(f'scratch/chunk-{s0}.json')
assert not out.exists(), 'never rerun a seed'
source=Path('contracts/src/Qt1.sol').read_text()
c=SolidityCompiler().compile(source)
assert c.success,c.errors
print('runtime_bytes',len(c.deployed_bytecode),flush=True)
# Execute the pinned _build_configs body verbatim except iterate only the requested
# seed range; each config uses its own default_rng(seed=i), no prefix dependency.
body=textwrap.dedent(inspect.getsource(h.MatchRunner._build_configs))
assert body.count('range(self.n_simulations)')==1
body=body.replace('range(self.n_simulations)',f'range({s0}, {s1})')
ns=dict(vars(match)); exec(body,ns)
runner=h.MatchRunner(n_simulations=s1,config=h.build_base_config(seed=None),n_workers=8,variance=h.BASELINE_VARIANCE)
cfgs=ns['_build_configs'](runner)
assert [x.seed for x in cfgs]==list(range(s0,s1))
vanilla=list(h.get_vanilla_bytecode_and_abi()[0])
r,e=h.run('qt1',list(c.bytecode),vanilla,cfgs)
d={'runtime_bytes':len(c.deployed_bytecode),'source_sha256':hashlib.sha256(source.encode()).hexdigest(),'seeds':[s0,s1],'run':r,'edges':{str(k):v[0] for k,v in sorted(e.items())}}
out.write_text(json.dumps(d,indent=2)); print(json.dumps(d),flush=True)
assert r['valid'], r
```

Other Python commands: a `python3` heredoc fetched the five raw GitHub URLs listed in sources.json using `urllib.request.urlopen`, asserted each listed quote appeared verbatim, and wrote sources.json with those URLs plus Q and its exact Taylor credit line. A `cat` heredoc wrote `test/scratch/report.py`; `python3 test/scratch/report.py` aggregated the four saved JSON outputs using `statistics.mean`, `statistics.stdev / sqrt(n)`, summed gas counters, maxima and weighted means, asserted all 128 distinct seeds and successful calls, and wrote this report. Final `python3` verification parsed both named files, checked all source quotes, required sections, JSON seed coverage and gas limits, byte limits, and unchanged qt1 source SHA256; `git status --short --untracked-files=all` checked output paths. Tool session polls only collected output; they executed no additional shell commands or simulations.

Final local checks: **PASS** — both JSON documents parse; all 128 expected seed keys occur once; mean and sample SE recompute exactly; every callback succeeded; size and gas caps hold; qt1 source SHA256 remains unchanged; all six source quotes were fetched again and matched. These checks are local evidence only. `result.md` is below 64,000 bytes and `sources.json` is 1,343 bytes. Required outputs remain untracked (ignored by the repository rules); nothing was staged or committed.

Additional inspection commands were a Python heredoc that printed each completed chunk's seed range, timing, validity and peak; `cat > test/scratch/verify.py` containing the checks described above; `python3 test/scratch/report.py`; `python3 test/scratch/verify.py`; and `git status --short --untracked-files=all`. This final Python heredoc inserted the observed check results into the report and checked the final byte limit and untracked status with `git ls-files --error-unmatch` for each required output.

## Problems

No setup, validation or simulation step failed. The measured afterSwap peak may exceed the historical 208,775 figure because these are fresh seeds; the measured maximum remains below 245,000. No policy changes, candidate comparison, replica Gate P/Q or fee-parity Gate F were attempted: this item is the unchanged qt1 engine baseline only. These results alone do not establish an improvement over qt1 or a leaderboard gain. Temporary engines, dependencies and logs are disposable scaffolding; the two requested files contain the deliverable.