LZLZL/Prediction markets/Strategy map
FREEJUDGE IT F · MethodOverview

There are only four ways
to make money here

2026-08-21 · everything else is one of these four, repackaged

There are four sources of profit on this venue. Every approach that sounds novel decomposes into one or a combination of them. Knowing which one you are on matters more than any technique, because the four differ completely in cost structure, capacity ceiling, and how they fail.

1The four

PathYou earnYou payYou need
① Directionalthe amount by which your estimate beats the marketfee plus spreadinformation or judgement
② Market makingspread plus the maker rebateadverse selectionscreen time, fast cancels, inventory control
③ Structural arbitragethe gap between two prices for the same riskfees on every leg, plus unpaired riskspeed
④ Platform subsidythe subsidythe trading costs incurred to earn itvolume

They can be stacked — making markets while collecting rebates, or accumulating a tier while trading directionally. But stacking does not change the nature of the primary path: if ① is negative, ④ does not fix it (see §5).

2① Directional — the only one without a ceiling

You think something has a 70% chance and the market quotes 0.62; you buy, and you are earning those 8 percentage points.

the entire premise: your estimate is better than the market's and better by enough to cover fee and spread

It is the only one of the four with no structural ceiling: if you genuinely are more accurate, capital can keep scaling until your orders start moving the price. It is also the only one nobody can verify for you — hence the easiest to be wrong about.

How to know whether you really are accurate? Build a calibration table: write down your probability before each bet, then after a few dozen, bucket by your estimate and compare against outcomes. That measure cannot be faked — method in price is probability.

3② Market making — fixed income, variable cost

Quote both sides, earn the spread, collect a rebate on fills, and pay no fee to post. It sounds free.

But posters do not pay a fee; they pay adverse selection — when you are hit, the taker usually knows something you do not. And that cost varies violently with tenor:

IncomeCostNet
Hourly board+0.248¢/share−0.475¢/share+0.099¢, barely positive
Five-minute board+0.25¢/share≈ −7.2¢/sharean order of magnitude short

Same action, same rebate, and changing the tenor takes it from barely positive to deeply negative. See adverse selection.

4③ Structural arbitrage — certain, and already taken

The cleanest of the four: a full set of outcome tokens is always worth $1, so assembling one for less than $1 yields a certain difference. No judgement of any kind required.

Which is exactly the problem — opportunities requiring no judgement are the first to be automated. I censused the lane: live executable inventory was zero.

⚠ Paper opportunity and collectable opportunity are different things

You see the two sides summing to 0.96 with four cents sitting there — but those four cents usually live in the band you cannot buy: quoted, with a few dozen shares behind it, or unfillable entirely.

And every leg pays a fee, which on multi-outcome boards means N fees. Once you account for those, the paper opportunity is generally gone. See complementary sets and walking the book.

5④ Platform subsidy — a multiplier, not an addend

The four rebate schemes have entirely different rules, pools and payout times, and conflating them is the commonest error (the distinction is here). This section is only about subsidy as a path.

both rebate programmes maxed ≈ 0.17–0.18pp against a fee line of 1.75pp

In other words: fully maxed, subsidies only move the required gross edge from about 1.75pp to about 1.5pp.

⚠ I verified this with real money

I built a line whose case for profitability leaned heavily on subsidies. Result: about 41 hours, roughly 496 windows, net −$496.51, and rebates received of $0.

And even if they had arrived it would not have saved it — that line's shortfall was over a percentage point and subsidies cover at most 0.3pp: an order of magnitude apart. Full ledger in the autopsy.

Subsidies make a positive strategy better and cannot rescue a negative one.

6Choosing

If you haveTakeFirst confirm
genuine domain knowledge① directionalbuild a calibration table and verify you really are sharper
screen time and engineering② market makingcheck whether the category's population median is positive
low-latency infrastructure③ arbitragecheck whether any inventory remains — usually none
only capital and volume④ subsidyit is not a path on its own, only an addition to another

The last row is this entry's main point: ④ cannot stand alone. Trading in order to collect a subsidy always pays out more in fees than it returns (even the top tier returns only 50%). It only makes sense attached to a path that is already positive.

7One line to keep

Ask yourself which one you are on: I am sharper than the market (①), I provide liquidity (②), I capture a price inconsistency (③), or I collect a subsidy (④). An approach that cannot answer is usually ④ dressed up as ①.

EvidenceCheck it yourself

Market-making arithmetic hourly net +0.099¢/share, five-minute deeply negative; 47 same-shape wallets with a median of −0.74¢/share.
Arbitrage inventory full census of the lane; no live executable opportunity.
Subsidy magnitude both rebate programmes maxed ≈ 0.17–0.18pp against a 1.75pp fee line.
Verified with money a line leaning on subsidies: roughly 496 windows, net −$496.51, rebates received $0. See the autopsy.
Checked 2026-07 to 2026-08.
Not included my strategy parameters, signals and entry timing.

NextWhere to go

B · REBATES
Rebates are a multiplier, not an addend
F · METHOD
The real ceiling on each path
B · REBATES
The real cost of market making
E · MARKETS
The line that bet on subsidies
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.