LZLZL/Prediction markets/Engineering
FREEJUDGE IT G · EngineeringBasis

Subsidies do not appear
in trading P&L

2026-08-21 · a basis that makes you wrong in both directions

The leaderboard's "profit" is trading P&L, while the venue's subsidies are transfers, not trades — and never enter it. The consequence runs both ways: you systematically underrate subsidy-driven players and systematically overrate your own strategy.

1Two streams that never overlap

Trading P&LSubsidies
Formspread and settlement payoutstransfers
In the "profit" field?yesno
Where to find itthe leaderboard endpointon-chain receipts
Timingwhenever things settlebatched at a fixed daily time
real net income = trading P&L + total received two queries, then add. No endpoint does this for you

2Direction one: underrating the subsidy players

A pure market-making wallet can have persistently negative trading P&L — it quotes tight to attract volume, and each individual trade loses slightly. Its income comes from rebates.

I took one such wallet apart in full (thirty hours, over twenty thousand shares):

ComponentPer share
Pure trading−0.149¢
Maker rebate+0.248¢
Netpositive

On the leaderboard, that wallet looks like it is losing money. It is steadily profitable, and its entire approach is built on this basis gap.

⚠ And a more counter-intuitive finding

Across 62 boards, 25 had paired average prices summing to more than $1.00 (the highest reaching 1.250). It was deliberately assembling sets at more than they are worth — a set only ever pays $1, so those trades were certain losses.

Why? Because it is not buying spread, it is buying volume — volume produces rebates, and the rebates exceed the spread it gives up. Until you see that, every one of its actions looks like a mistake.

3Direction two: overrating yourself

What you computeWhat it may mean
"I made X this month"possibly "the strategy lost and subsidies more than covered it"
"the strategy works"possibly "the subsidy works"
"ten times the size, ten times the profit"subsidy rates change and can go to zero; the strategy component stays negative

This error is expensive because a subsidy is a product decision, not your skill. Rates, splits and eligible categories can all change — and when they do, your "working strategy" turns negative that day, and you may not even know what you had been relying on.

4How to do it correctly

DoNote
Keep two ledgersstrategy P&L and subsidy income, never merged
Judge the strategy on the first onlya subsidy is income, not evidence of skill
Add them only for total returnfine for "how much did I make"; not for "does this work"
Query both when reading othersthe leaderboard is half the picture; the other half is on-chain
Stress-test at zero subsidywhat remains is your strategy

5My own instance

I built a line whose case for profitability leaned heavily on subsidies. Result: rebates received $0, net −$496.51.

The lesson is not only "the subsidy did not come". It is that I never validated the subsidy column on its own. The strategy ran forty-one hours, while validating the subsidy would have meant a few small orders and one day of waiting.

Basisthe ledger recorded the $0 and not the reason. So the kill file says "no cause recorded" rather than supplying a plausible one. Full account in the autopsy.

6One line to keep

The leaderboard's "profit" is half the story; the other half is a set of on-chain transfers. Judging other people needs both. Judging your own strategy needs only the first. Counting subsidies as strategy skill means mistaking a product decision for your own ability.

EvidenceCheck it yourself

Two streams subsidies are paid as transfers and never enter the trading-profit field; real net income = trading P&L + total received, queried separately.
Market-making wallet thirty hours, over twenty thousand shares: pure trading −0.149¢/share, maker rebate +0.248¢/share, net positive; 25 of 62 boards had paired averages summing above $1.00 (highest 1.250). No identity attribution is made.
My own case a line leaning on subsidies: rebates received $0, net −$496.51. The ledger did not record why the rebates were zero.
Related traps see eight API traps.
Checked 2026-07 to 2026-08.

NextWhere to go

G · ENGINEERING
Eight API traps
B · REBATES
Four different things all called "rebate"
G · ENGINEERING
589 shares evaporated
B · REBATES
Rebates are a multiplier
This is an educational and research record. It is not investment advice, promises no returns, and offers no personalised trading recommendations. Rules and API behaviour are per the official documentation; this page states when it was checked and both can change without notice. Prediction markets are restricted or unavailable in some jurisdictions — confirm your own before taking part.