ArbAlpha

Guide

How prediction market arbitrage works and how to use ArbAlpha.

What is arbitrage?

The same event is listed on both Polymarket and Kalshi. Each contract pays $1 if it resolves true, $0 if not. Instead of predicting the outcome, you buy both sides — YES on one exchange and NO on the other. Exactly one side always wins, so one leg pays $1 at resolution regardless of the result. If the two prices together cost less than $1, the theoretical difference minus fees is the potential edge — assuming both legs fill at the shown prices and the markets resolve on identical terms.

Step 1 — Buy both legs once
YES on Polymarket @ $0.58+NO on Kalshi @ $0.36=$0.94 paid up front
Step 2 — At resolution, one leg pays $1.00
If the Fed…YES leg paysNO leg paysYou collect
cuts rates$1.00$0.00$1.00
holds rates$0.00$1.00$1.00
Step 3 — Your profit is the same either way
$1.00 collected − $0.94 paid = +$0.06 estimated, assuming both legs filled at the shown prices (~6% gross, ~5.4% after fees)

Anatomy of an opportunity tile

Every card on the dashboard looks like this. The numbered badges map to the explanations below.

#1Economicshigh3

Will the Fed cut rates at the June meeting?

1
+5.4%
net · after fees
2Polymarket YES @ $0.58+Kalshi NO @ $0.36
4Price & profit zone
5
Max size
4,200 shares
6
Capital
$3,948
7
Est. profit
$228.6
8
Avg YES / NO
$0.585 / $0.355
9
Edge (ask)
6.00%
10Trade Polymarket ↗Trade Kalshi ↗
  1. 1
    Net margin
    Your return after all estimated fees, as a percentage. This is the headline figure and the default sort key.
  2. 2
    The two legs
    Which side to buy on each exchange and at what price. Together one leg pays $1 at resolution — assuming both markets resolve on identical criteria.
  3. 3
    Category & confidence
    The event category and how strongly the two listings were matched as the same event. Always read both titles yourself to confirm.
  4. 4
    Realtime spread chart
    Polymarket and Kalshi implied probabilities plotted over time. The vertical gap between the lines is the spread. The green shaded band appears when the gap clears fees. Switch between time ranges or toggle to Edge % view.
  5. 5
    Max size
    The maximum share-pairs executable at a profit, found by walking both order books level by level. Arbs are not infinitely scalable.
  6. 6
    Capital
    Total dollars to deploy across both legs to reach the max size.
  7. 7
    Est. profit
    Fee-adjusted dollar profit if you fill the full max size. Fees are estimates — verify before trading.
  8. 8
    Avg YES / Avg NO
    Volume-weighted average price you'd actually pay across the order-book levels you consume. Usually slightly worse than the best quote.
  9. 9
    Edge (ask)
    The executable edge using live ask prices: 1 − (best YES ask + best NO ask). This is the real-world entry point.
  10. 10
    Trade buttons
    Open each market directly to place both legs.

How to execute

  1. 1
    Confirm it's the same event
    Open both market pages and verify they resolve on identical criteria — same outcome, date, and threshold. A mismatched pair eliminates the modelled edge entirely. This is the single most important check.
  2. 2
    Buy both legs simultaneously
    Use the trade buttons to open each market. Buy the indicated side at approximately the shown price. Speed matters — prices can move before both legs fill.
  3. 3
    Stay within Max size
    Beyond Max size the edge is gone. For smaller positions, split capital between legs in proportion to their prices.
  4. 4
    Prices are a 30-second snapshot
    Re-check the live order books before committing. The scanner refreshes every 30 seconds but exchange prices move continuously.

Risks & caveats

  • Resolution mismatch — Two similarly-worded markets can resolve on different criteria. Always read both descriptions before trading. This is the primary risk.
  • Fee estimates — Fees are estimated and vary by exchange, category, and position size. The real cost may differ.
  • Slippage & depth — Large orders move the price. Max size is a snapshot, not a guarantee of fill.
  • Execution timing — You may not fill both legs at the quoted price simultaneously. Market moves between leg one and leg two create residual risk.
  • Capital lock-up — Your capital is tied up until the market resolves. Factor the days-to-resolution into your sizing decision.

ArbAlpha is an information tool, not financial advice. You are responsible for verifying every trade.

What is ChainPulse?

Every Polymarket trade settles on the Polygon blockchain, which means the entire history of who traded what, when, and for how much is a matter of public record. ChainPulse reads that record and surfaces four things on each Polymarket-linked tile:

  • 🐋 Whale alerts — when a single taker BUY tops $5,000.
  • 📈 Volume spikes — when 24-hour volume is 2.5x its 7-day baseline.
  • 🧠 Top holder active — when one of the top 10 holders has traded the market recently.
  • ↑ ↓ Order flow — net direction of the last 20 trades by USD volume.

Think of it as seeing the activity behind the price. ChainPulse doesn't tell you what will happen — it tells you what is happening right now. Kalshi has no equivalent because it's a regulated company with a private order database, not a public blockchain.

On-chain data is an intelligence signal, not a prediction. A whale can be a hedger; unusual volume can precede silence. Cross-check with the news collapsible and the spread chart.

← Open the Scanner