"""Exact P&L for Polymarket soccer makers, from Polygon logs. (read-only) WHY ON-CHAIN. The REST route cannot do this. /activity caps at offset ~5400 and 40 of 42 real makers exceed it, and a truncated ledger biases P&L UPWARD without limit -- a SELL or REDEEM whose purchase falls outside the window has no cost basis and reads as pure profit (it produced a fake $8.8M row). On-chain there is no such cap: OrderFilled indexes `maker`, so one wallet's entire fill history is directly queryable. MAKER-FIRST, NOT MARKET-FIRST. `assetId` is NOT indexed, and the exchange emits ~46k OrderFilled per 1000 blocks (~2M/day) across all of Polymarket, so scanning a market's block range and filtering locally is impossible on a public RPC. Queries must be keyed on the maker address, then filtered to soccer tokens. DECODING (verified against the API to the cent on 0xfb101b79...fca7e): topics = [sig, orderHash, maker, taker]; data words = [makerAssetId, takerAssetId, makerAmountFilled, takerAmountFilled, fee, ...]. assetId 0 is USDC, so makerAssetId==0 means the MAKER PAID USDC (bought the token) and takerAssetId==0 means the maker SOLD it. Amounts are 1e6-scaled. P&L = net USDC from fills + settlement value of the net token position, where settlement comes from gamma outcomePrices. A soccer market's whole lifecycle fits inside the window, so every position has its cost basis -- which is the property the REST route could not provide. THE RELIABILITY FLAG. SPLIT/MERGE mint or burn a Yes+No pair for $1 without any fill, so a wallet that uses them has cash flow this script cannot see. Those counts are read from /activity and reported as `split_merge`; a row with a large value is NOT trustworthy and is flagged rather than silently ranked. STILL NOT ANSWERABLE, whatever we do here: QUOTES. Polymarket's CLOB is off-chain, so resting orders are public nowhere. Everything below is fills, so "legs" means legs FILLED, never legs quoted. """ import collections import csv import json import os import sys import time import requests RPC = 'https://polygon-bor-rpc.publicnode.com' EXCHANGE = '0xe111180000d2663c0091e4f400237545b87b996b' ORDER_FILLED = '0xd543adfd945773f1a62f74f0ee55a5e3b9b1a28262980ba90b1a89f2ea84d8ee' GAMMA = 'https://gamma-api.polymarket.com' DATA = 'https://data-api.polymarket.com' HERE = os.path.dirname(os.path.abspath(__file__)) OUT = os.path.join(HERE, 'pm_onchain_pnl.csv') CHUNK = 10000 # RPC hard cap on block span BLOCKS_PER_DAY = 43200 # ~2s blocks def rpc(method, params, tries=4): for i in range(tries): try: r = requests.post(RPC, json={'jsonrpc': '2.0', 'id': 1, 'method': method, 'params': params}, timeout=45) j = r.json() if 'result' in j: return j['result'] except Exception: pass time.sleep(1.0 * (i + 1)) return None def http(url, params, tries=3): for i in range(tries): try: r = requests.get(url, params=params, timeout=30) if r.status_code == 200: return r.json() if r.status_code == 400: return None except Exception: pass time.sleep(1.0 * (i + 1)) return None def soccer_tokens(): """token_id -> (match_slug, leg_question, settlement_price).""" tok = {} for closed in ('true', 'false'): for off in (0, 100, 200, 300): evs = http(f'{GAMMA}/events', {'limit': 100, 'offset': off, 'tag_slug': 'soccer', 'closed': closed, 'order': 'volume', 'ascending': 'false'}) if not evs: break for e in evs: ms = e.get('markets') or [] if len(ms) != 3 or ' vs' not in (e.get('title') or '').lower(): continue for m in ms: try: ids = json.loads(m.get('clobTokenIds') or '[]') pr = [float(x) for x in json.loads(m.get('outcomePrices') or '[]')] except Exception: continue for i, tid in enumerate(ids): tok[str(int(tid))] = (e.get('slug'), m.get('question'), pr[i] if i < len(pr) else None) return tok def maker_fills(wallet, from_block, to_block, log=None): """Every OrderFilled where this wallet was the MAKER, decoded.""" topic = '0x' + '0' * 24 + wallet[2:].lower() out = [] b = from_block while b < to_block: hi = min(b + CHUNK, to_block) res = rpc('eth_getLogs', [{'fromBlock': hex(b), 'toBlock': hex(hi), 'address': EXCHANGE, 'topics': [ORDER_FILLED, None, topic]}]) for l in (res or []): d = l['data'][2:] w = [d[i:i + 64] for i in range(0, len(d), 64)] if len(w) < 4: continue mk_asset, tk_asset = str(int(w[0], 16)), str(int(w[1], 16)) mk_amt, tk_amt = int(w[2], 16) / 1e6, int(w[3], 16) / 1e6 if mk_asset == '0': # maker paid USDC -> BOUGHT tk_asset out.append((tk_asset, +tk_amt, -mk_amt)) elif tk_asset == '0': # maker delivered token -> SOLD it out.append((mk_asset, -mk_amt, +tk_amt)) b = hi time.sleep(0.05) return out def main(): topn = int(sys.argv[1]) if len(sys.argv) > 1 else 20 pad_d = float(sys.argv[2]) if len(sys.argv) > 2 else 2.0 head = int(rpc('eth_blockNumber', []), 16) tok = soccer_tokens() print(f'soccer tokens mapped: {len(tok)}') # THE WINDOW MUST COVER WHEN THE SOCCER MATCHES WERE LIVE, not "recently". # Scanning the last N days found 13,720 fills for one candidate and ZERO in # soccer: these are generalist makers whose current book is elsewhere, and # the matches in the sample settled months ago. Derive the range from actual # soccer trade timestamps instead, padded either side so opening and # settlement fills are inside it. ts = [] seen = 0 # sample trade timestamps across the mapped soccer markets for tid_slug in list({v[0] for v in tok.values()})[:14]: d = http(f'{GAMMA}/events', {'slug': tid_slug}) if not d: continue for e in d: for m in (e.get('markets') or []): tr = http(f'{DATA}/trades', {'market': m.get('conditionId'), 'limit': 1, 'takerOnly': 'false'}) if tr: ts.append(int(tr[0]['timestamp'])) tr2 = http(f'{DATA}/trades', {'market': m.get('conditionId'), 'limit': 1, 'offset': 400, 'takerOnly': 'false'}) if tr2: ts.append(int(tr2[0]['timestamp'])) seen += 1 if seen > 20: break if not ts: print('could not date the soccer markets'); return 1 now = int(time.time()) hi_ts, lo_ts = max(ts), min(ts) frm = head - int((now - lo_ts + pad_d * 86400) / 2) to = head - int(max(0, (now - hi_ts - pad_d * 86400)) / 2) print('soccer activity spans %s .. %s' % (time.strftime('%Y-%m-%d', time.gmtime(lo_ts)), time.strftime('%Y-%m-%d', time.gmtime(hi_ts)))) print('blocks %d..%d (%d calls/wallet)' % (frm, to, (to - frm) // CHUNK)) head = to src = list(csv.DictReader(open(os.path.join(HERE, 'pm_maker_study.csv'), encoding='utf-8'))) cands = [r for r in src if float(r['maker_share']) >= 0.6] cands.sort(key=lambda r: -int(r['fills'])) cands = cands[:topn] print(f'candidate makers: {len(cands)}\n') rows = [] for i, c in enumerate(cands, 1): w = c['wallet'] fills = maker_fills(w, frm, head) cash = 0.0 pos = collections.defaultdict(float) n_soccer = 0 legs = collections.defaultdict(set) for tid, dq, dcash in fills: if tid not in tok: continue n_soccer += 1 cash += dcash pos[tid] += dq slug, q, _ = tok[tid] legs[slug].add(q) settle = 0.0 unresolved = 0 for tid, q in pos.items(): p = tok[tid][2] if p is None: unresolved += 1 else: settle += q * p act = http(f'{DATA}/activity', {'user': w, 'limit': 500}) or [] sm = sum(1 for a in act if a.get('type') in ('SPLIT', 'MERGE')) nlegs = [len(v) for v in legs.values()] rows.append({ 'wallet': w, 'pnl_soccer': round(cash + settle, 2), 'soccer_fills': n_soccer, 'matches': len(legs), 'mean_legs_filled': round(sum(nlegs) / len(nlegs), 2) if nlegs else 0, 'all3_share': round(sum(1 for x in nlegs if x == 3) / len(nlegs), 2) if nlegs else 0, 'total_fills_window': len(fills), 'unresolved_tokens': unresolved, 'split_merge_recent': sm, 'reliable': int(sm == 0 and unresolved == 0), }) print(' %2d/%d %s soccer_fills=%-6d matches=%-4d pnl=%12.2f %s' % (i, len(cands), w[:14], n_soccer, len(legs), cash + settle, '' if rows[-1]['reliable'] else '(flagged)')) rows.sort(key=lambda r: -r['pnl_soccer']) with open(OUT, 'w', newline='', encoding='utf-8') as fh: wtr = csv.DictWriter(fh, fieldnames=list(rows[0].keys())) wtr.writeheader(); wtr.writerows(rows) print(f'\nwrote {len(rows)} -> {OUT}') rel = [r for r in rows if r['reliable'] and r['soccer_fills'] >= 20] print('\nRELIABLE ROWS (no split/merge, all positions settled, >=20 fills): %d' % len(rel)) for r in rel[:12]: print(' %-44s pnl=%11.2f fills=%-6d matches=%-4d legs=%.2f all3=%.0f%%' % (r['wallet'], r['pnl_soccer'], r['soccer_fills'], r['matches'], r['mean_legs_filled'], 100 * r['all3_share'])) if rel: p = [r for r in rel if r['pnl_soccer'] > 0] print('\n profitable: %d of %d (%.0f%%)' % (len(p), len(rel), 100 * len(p) / len(rel))) if p: print(' profitable makers: mean legs filled %.2f, all3 %.0f%%' % (sum(r['mean_legs_filled'] for r in p) / len(p), 100 * sum(r['all3_share'] for r in p) / len(p))) l = [r for r in rel if r['pnl_soccer'] <= 0] if l: print(' losing makers : mean legs filled %.2f' % (sum(r['mean_legs_filled'] for r in l) / len(l))) return 0 if __name__ == '__main__': sys.exit(main())