"""Who are the PROFITABLE market makers in Polymarket soccer, and what do they do? (read-only; touches no account and places nothing) WHY. Our own maker is stranded 13.8% of the time against an 11.1% break-even, and the open question is whether assembling a 3-leg book is even the right construct. Polymarket settles on-chain, so unlike Kalshi -- where trades are anonymous and only AGGREGATE maker P&L is recoverable (research/mm_feasibility.py) -- individual makers can be named, ranked by realised P&L, and studied. Polymarket models a soccer match as THREE binary markets ("Will X win?", "...end in a draw?", "Will Y win?"), which is exactly our three legs, so the comparison is structural rather than analogical. HOW A MAKER IS IDENTIFIED. data-api /trades returns the TAKER side only by default; with takerOnly=false it returns BOTH sides of every fill. The set difference on (transactionHash, proxyWallet) is therefore the maker side. Verified 2026-09-14: 901 of 942 hashes resolve to clean 2-row pairs and 53% of rows classify as maker, which is the ~half you would expect. P&L. Per wallet per market: cash flow (BUY -size*px, SELL +size*px) plus the settlement value of the net position (outcomePrices is ["1","0"]-style and aligns with outcomeIndex). A negative net position is the complement leg and the same arithmetic holds. READ THE OUTPUT SCEPTICALLY: * SURVIVORSHIP. These are wallets that are currently large. Anyone who ran the same strategy into the ground is absent, which is exactly the inference error that produced the South America mirage. * NO QUEUE DATA. Fills are visible; resting orders are not. Nothing here can speak to queue position or refresh cadence. * DIFFERENT ECONOMICS. Polymarket's fee and tick structure is not Kalshi's, and our verdict rests on the 1c tick being the binding constraint. Findings transfer as HYPOTHESES. * SIZE. The top wallet moved 2.2M on one match. What works while warehousing that inventory need not work at $1/leg. * The offset cap on /trades is ~10k rows per market, so the very largest markets are TRUNCATED (reported per market as `truncated`). """ import collections import json import os import sys import time import requests GAMMA = 'https://gamma-api.polymarket.com' DATA = 'https://data-api.polymarket.com' OUT = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'pm_maker_study.csv') PAGE = 500 OFFSET_CAP = 10000 # the API refuses beyond roughly this def _get(url, params, tries=4): for i in range(tries): try: r = requests.get(url, params=params, timeout=30) except Exception: time.sleep(1.5 * (i + 1)); continue if r.status_code == 200: return r.json() if r.status_code == 400: # offset past the cap -- not an error return None time.sleep(1.5 * (i + 1)) return None def match_events(n_events, closed=True): """Highest-volume soccer matches that have the 3-leg structure.""" out = [] for off in (0, 100, 200, 300, 400): evs = _get(f'{GAMMA}/events', {'limit': 100, 'offset': off, 'tag_slug': 'soccer', 'closed': str(closed).lower(), 'order': 'volume', 'ascending': 'false'}) if not evs: break for e in evs: ms = e.get('markets') or [] if len(ms) == 3 and ' vs' in (e.get('title') or '').lower(): if all(m.get('outcomePrices') for m in ms): vol = sum(float(m.get('volume') or 0) for m in ms) out.append((vol, e)) if len(out) >= n_events * 3: break out.sort(key=lambda kv: -kv[0]) return [e for _, e in out[:n_events]] def trades(cid, taker_only): rows, off = [], 0 while off < OFFSET_CAP: p = {'market': cid, 'limit': PAGE, 'offset': off} if not taker_only: p['takerOnly'] = 'false' d = _get(f'{DATA}/trades', p) if not d: break rows += d off += len(d) if len(d) < PAGE: break time.sleep(0.12) return rows, off >= OFFSET_CAP def study(n_events=6, complete_only=True, pool=400): """Aggregate maker behaviour over matches with COMPLETE trade history. THE TRUNCATION PROBLEM. /trades refuses offset > ~10500, and before/after filters are ignored, so the biggest markets return only their most RECENT 10k rows per leg -- a slice pressed right up against settlement, where trades happen at 0.999/0.001 and are near-riskless cleanup rather than market making. Measuring "legs quoted per match" on that window biases hard toward 1, because a maker who worked all three legs during the match but only one at the end reads as a one-leg maker. So: walk DOWN the volume ranking and keep only matches whose every leg came back under the cap. Smaller markets, but the history is whole and the number means what it says. """ evs = match_events(pool) if complete_only else match_events(n_events) print(f'{len(evs)} candidate matches with the 3-leg structure' f'{" (selecting complete-history ones)" if complete_only else ""}\n') # wallet -> stats W = collections.defaultdict(lambda: { 'pnl': 0.0, 'maker_fills': 0, 'taker_fills': 0, 'size': 0.0, 'events': set(), 'legs': collections.defaultdict(set)}) used = 0 for e in evs: if used >= n_events: break title = (e.get('title') or '').encode('ascii', 'replace').decode() # Pre-flight: is this match's history complete? Probe each leg at the # cap; if anything is still returning rows there, the market is bigger # than we can page and the sample would be settlement-skewed. if complete_only: trunc_any = False for m in e['markets']: d = _get(f'{DATA}/trades', {'market': m.get('conditionId'), 'limit': 1, 'offset': OFFSET_CAP - 200, 'takerOnly': 'false'}) if d: trunc_any = True break if trunc_any: continue used += 1 print(f' {title[:52]:<52}', end='', flush=True) n_rows = 0 for m in e['markets']: cid = m.get('conditionId') try: prices = [float(x) for x in json.loads(m.get('outcomePrices'))] except Exception: continue allr, trunc = trades(cid, False) tk, _ = trades(cid, True) tkey = {(t['transactionHash'], t['proxyWallet']) for t in tk} n_rows += len(allr) # cash + position per wallet for THIS leg cash = collections.defaultdict(float) posn = collections.defaultdict(lambda: collections.defaultdict(float)) for t in allr: w = t['proxyWallet'] try: sz = float(t.get('size') or 0); px = float(t.get('price') or 0) oi = int(t.get('outcomeIndex') or 0) except (TypeError, ValueError): continue is_maker = (t['transactionHash'], w) not in tkey if is_maker: W[w]['maker_fills'] += 1 else: W[w]['taker_fills'] += 1 W[w]['size'] += sz W[w]['events'].add(e.get('slug')) W[w]['legs'][e.get('slug')].add(cid) if (t.get('side') or '').upper() == 'BUY': cash[w] -= sz * px; posn[w][oi] += sz else: cash[w] += sz * px; posn[w][oi] -= sz for w, c in cash.items(): settle = sum(q * (prices[i] if i < len(prices) else 0.0) for i, q in posn[w].items()) W[w]['pnl'] += c + settle print(f' {n_rows:>6} rows') return W def main(): n = int(sys.argv[1]) if len(sys.argv) > 1 else 6 W = study(n) rows = [] for w, s in W.items(): tot = s['maker_fills'] + s['taker_fills'] if tot < 20: continue mshare = s['maker_fills'] / tot legs = [len(v) for v in s['legs'].values()] rows.append({ 'wallet': w, 'pnl': round(s['pnl'], 2), 'fills': tot, 'maker_share': round(mshare, 3), 'size': round(s['size']), 'events': len(s['events']), 'avg_legs_per_event': round(sum(legs) / len(legs), 2) if legs else 0, 'all3_event_share': round(sum(1 for L in legs if L == 3) / len(legs), 2) if legs else 0, }) rows.sort(key=lambda r: -r['pnl']) import csv with open(OUT, 'w', newline='', encoding='utf-8') as fh: w_ = csv.DictWriter(fh, fieldnames=list(rows[0].keys()) if rows else ['wallet', 'pnl', 'fills', 'maker_share', 'size', 'events', 'avg_legs_per_event', 'all3_event_share']) w_.writeheader(); w_.writerows(rows) print(f'\nwrote {len(rows)} wallets -> {OUT}\n') mk = [r for r in rows if r['maker_share'] >= 0.6] print('TOP MAKER-DOMINANT WALLETS BY REALISED P&L') print(' %-44s %10s %7s %6s %6s %s' % ('wallet', 'pnl', 'fills', 'mk%', 'evts', 'legs/evt')) for r in mk[:12]: print(' %-44s %10.0f %7d %6.2f %6d %.2f' % (r['wallet'], r['pnl'], r['fills'], r['maker_share'], r['events'], r['avg_legs_per_event'])) if mk: prof = [r for r in mk if r['pnl'] > 0] print('\n maker-dominant wallets: %d profitable: %d (%.0f%%)' % (len(mk), len(prof), 100 * len(prof) / len(mk))) if prof: print(' THE QUESTION: legs quoted per match, profitable makers') a = sum(r['avg_legs_per_event'] for r in prof) / len(prof) s3 = sum(r['all3_event_share'] for r in prof) / len(prof) print(' mean legs/match : %.2f (3.0 = full book like ours)' % a) print(' share of matches all3: %.2f' % s3) return 0 if __name__ == '__main__': sys.exit(main())