Prediction Market Arbitrage

Financial Dashboard — Key Numbers

Investment Required ?

$5,000

to get started

Break-even ?

Month 9

from launch

MRR Target ?

$3,000

per month

Margin ?

40%

of revenue retained

Monthly Revenue Growth Forecast

Unit Economics — Numbers per Customer

Profitability ?

0–15% ROI/мес

of revenue retained

Customer Lifetime Value ?

$900

lifetime total

Max Acquisition Cost ?

$270

per new customer

LTV / CAC ?

3.3×

✓ Above benchmark

Development Scenarios

Scenario	Revenue by Month 6	Revenue by Month 12	Key Assumption
Pessimistic	$128	$779	CAC above forecast, conversion below 5%
Realistic	$368	$2,227	On plan: CAC at or below target, churn at or below 5%/month
Optimistic	$920	$5,567	Virality worked, CAC is 2x below target

Why This Verdict

✓ Arguments FOR

A unique niche with little competition among algorithmic players
Growing Polymarket volume creates more opportunities
The AI advantage in processing speed is real

✗ Why not higher

Market illiquidity limits scale
Regulatory uncertainty: prediction markets are under pressure from US regulators
Opportunities disappear quickly as competition grows

🛑 When to Stop the Project

K1

No conversions after 100 clicks

leads == 0 AND clicks >= 100

→ pause traffic, diagnostics

K2

CAC exceeded LTV × 0.5

cac > ltv * 0.5 AND leads >= 5

→ stop traffic

K3

Two Gate failures

gate_failures >= 2

→ close project

What to Validate Before Scaling

1

Prediction markets update their prices slower than new information appears

2

AI that analyzes news in real time can trade faster than a human arbitrageur

3

Stable inefficiencies exist between multiple prediction markets

What the AI Models Said

↓

Claude Opus (Critic)

Prediction market arbitrage is a professional strategy with minimal real opportunity. Polymarket and Manifold markets are illiquid. Most 'arbitrage' opportunities are information gaps, not genuine risk-free arbitrage.

↓

GPT-4.1 (Market Strategist)

Prediction markets are a growing niche. AI that compares probabilities across markets (Polymarket vs Kalshi vs media forecasts) can find real inefficiencies. Automation is the key to speed.

~

Grok-3 (Technical Analyst)

Prediction market arb is a unique niche with no serious algorithmic players yet. Polymarket volume is $500M/month and growing. AI that reads news and places bets before the market repricing gives a real edge.

Milestones & Stages

M1

Research prediction markets

Done

M2

Develop a price-scanning algorithm

Done

M3

Test with real trades

In Progress

M4

Scale capital and volume

Pending

M5

Implement trading automation

Pending

Investment & Exit Scenarios

Total Investment Needed

$5,000

to reach profitability

Marketing $2,500

Development $1,250

Infrastructure $750

Operations $500

🚦 Strengths & Risks at a Glance

✓ Green Flags

Combined monthly volume on Polymarket and Kalshi is growing fast. The market is clearly expanding, not stagnating.

Volume already runs into billions of dollars a month with steady growth (The Block). Check the exact figure again at launch time, the market moves fast.
Structural price gaps exist between offshore Polymarket and regulated Kalshi because of different audiences and listing rules. This is a real source of edge, not an invented one.

spread of 1-5% according to industry guides (Claw Arbs, TradingVPS)
The niche is barely touched by large algorithmic players. Competition is lower than in classic crypto arbitrage.
Kalshi received formal CFTC approval in March 2026. This legitimizes the regulated side of the market and lowers the risk of a full ban on that side.
The MVP is cheap: $5,000 covers a scanner prototype and a first paid test without complex infrastructure.

✗ Red Flags

The vast majority of traders and bots on Polymarket lose money. Only 0.51% of wallets ended up more than $1,000 in profit (StrongMocha analysis, 2026). If the service's clients lose money, they cancel their subscription in the first month.
high
Upcoming CFTC regulation for prediction markets creates uncertainty. The regulator is clearly moving toward formal rules for this niche, possibly including a ban on some contracts, but the outcome and timeline are not yet known. The regulatory base could change before the product gains any traction.
high
The arbitrage window closes within seconds as more bots spot the same gap. The edge is structurally temporary, not a permanent business asset.
medium
Low liquidity on some markets (especially Manifold and niche Polymarket contracts) caps trade size before execution itself moves the price.
medium
The technical bar to build a DIY scanner is low. The target audience (traders with technical skills) is more likely to write its own script than pay $300/month.
medium
Polymarket officially blocks users from the US. A product aimed at a US audience falls into a legal gray zone.
low

⚠️ Risk Matrix

Risk	Probability	Impact	Mitigation
CFTC bans some contracts as a result of the ongoing regulatory process 🛑 Kill trigger	55%	High	Build only around international Polymarket arbitrage, avoid US-regulated event contracts until the CFTC's final decision
The arb window closes faster than the client can execute a trade (API delay/liquidity) 🛑 Kill trigger	60%	High	Not technically fixable in full. Reduce it through direct API connections and speed priority, and never present the product as risk-free.
Clients cancel their subscription after the first losing month (high churn due to low real profitability in the niche)	50%	High	Offer a free trial with an honest P&L report before payment, instead of promising a guaranteed edge
CAC comes in above the $270 target: the prediction market trader audience is narrow and small	45%	Medium	Targeted outreach in niche trader Discord/Telegram communities, no paid search
Polymarket or Kalshi blocks bot access via the API (rate limit / ban) 🛑 Kill trigger	35%	High	Diversify across 3+ platforms (add Manifold, PredictIt) from the start
Large market makers/HFT firms enter the niche and kill the spread	40%	Medium	Treat the window of opportunity as temporary (12-18 months) and don't build long-term growth into the model

💸 Monthly Cash Flow (Realistic Scenario)

Period	Revenue	Expenses	Net	Cumulative
Start (investment)	$0	$5 000	-$5 000	-$5 000
Month 1	$7	$700	-$693	-$5 693
Month 2	$32	$700	-$668	-$6 361
Month 3	$80	$700	-$620	-$6 981
Month 4	$151	$700	-$549	-$7 530
Month 5	$246	$700	-$454	-$7 984
Month 6	$368	$700	-$332	-$8 316
Month 9 (claimed breakeven)	$900	$700	+$200	-$8,305 (estimate)

🏁 Competitive Landscape

📡 Market catalyst: Steady growth in prediction market trading volume through 2026, plus formal CFTC approval of Kalshi in March 2026, which legitimized the regulated side of the niche

Competitor	Size	Take Rate	Weakness
Claw Arbs (clawarbs.com)	—	—	Publishes guides and tools for Kalshi-Polymarket arbitrage. It is a content and education product, not a managed service with automatic trade execution.
TradingVPS (tradingvps.io)	—	—	Sells VPS infrastructure and guides for arbitrage bots. It is an infrastructure player; the client writes the actual edge and logic on their own.
Open-source Polymarket/Kalshi price scanners on GitHub	—	—	Free for anyone with basic programming skills. A direct threat to the paid product, it lowers willingness to pay $300/month.
Adjacent News (adj.news)	—	—	Focused on analytics and data aggregation for prediction markets, not on automatic arbitrage detection and execution

🛠 MVP — Week-by-Week Plan

Week 1

Deploy a price scanner between the Polymarket and Kalshi APIs
Manually find and log the first price gaps

≥5 gaps/day with a spread >1%

Week 2

Automate Telegram/Discord alerts for gaps found
Execute the first trades manually with $500-1000 in capital (matches portfolio milestone M3)

ROI ≥2% across 10 test trades

Week 3

Package the scanner as a paid subscription (Creem), sign up 10 beta users

3-5 paying beta clients at $99-150

Week 4

Collect feedback from beta clients
Measure net edge after exchange fees and slippage

net ROI after costs ≥1.5%, otherwise kill per criterion K2

🏰 Competitive Moat

✗ Easy to Copy

The price scanner itself, built on two public APIs, can be rebuilt in days. Dozens of open-source equivalents already exist.
Telegram/Discord alerts for price gaps are a standard feature with no technical barrier

✓ Hard to Copy

Execution speed and minimal latency on both exchange connections (this needs direct integrations, not the public REST API)
A track record of signal accuracy and trust from paying traders built up over months

⏱ Moat forms by: Practically unreachable. The edge itself is structurally temporary (12-18 months) and collapses under growing competition and regulation faster than any moat can be built.

📊 Acquisition Cost by Channel

Channel	CAC	Notes	Profitable?
Niche Discord/Telegram communities of prediction market traders	$150-250 (estimate)	A targeted but small audience: no more than 5-10 thousand active traders across the entire niche	✓ Yes
Paid search (Google Ads) on queries about Polymarket/Kalshi arbitrage	$300-400 (estimate)	An expensive, competitive auction because of the broader crypto trading topic	✗ No
SEO/content guides on arbitrage (like competitor Claw Arbs)	$80-150 (estimate, long cycle)	Cheap per lead, but the effect stretches over months. Not suitable for a 4-week MVP test.	✓ Yes

🔬 Anti-Optimism Audit

1

The portfolio financial model builds in an LTV of $900 at a 40% margin and ARPU of $300, as if this were an ordinary SaaS with low churn

→ A 0.51% share of profitable wallets on Polymarket means almost none of the product's clients will actually make money, and most will cancel within 2-3 months. A realistic LTV sits closer to $600-900 at best, and the model should assume churn of 30-40%/month, not the standard 5%.

pulls the score down, confirms the existing kill

2

The portfolio's base verdict text allows a conditional GO 'given technical expertise and $10-50K in capital'

→ That condition fits personal trading by one specific person, not a scalable product with paying clients. A qualified client is more likely to build the scanner themselves (the entry bar is low), while an unqualified one will lose money and leave. As a SaaS business, the model doesn't add up even if the underlying signal is correct.

doesn't change the verdict, but removes the illusion of a GO from the portfolio text

3

Regulatory risk is presented in the portfolio as one manageable line item

→ In reality this is an open-ended process that could end in a full ban on part of the US prediction markets. The CFTC is already moving toward formal regulation of this niche, and the final rules and timeline are still undecided. This is not background risk, it is a potential kill trigger for the entire niche.

strengthens the kill case, this is not just a risk

4

A market-size estimate of billions of dollars a month creates the impression of huge room for the business

→ That is exchange trading volume, not available arbitrage edge. The margin actually extractable is a fraction of a percent of that volume, and competition closes it within seconds. The right number to look at is not exchange GMV but the size of profitable trades available to a single bot with no competition, which is orders of magnitude smaller.

corrects the GMV green flag, doesn't cancel it out entirely