Last Updated: 2026-10-26
A bot can execute a mistaken idea faster. Before comparing infrastructure or advertised returns, write down what each position pays in every relevant state. A correlated price gap, a logical implication and a fully hedged payoff are different claims. This worksheet uses invented contracts and prices; it is not a measurement of bot activity or a live trading recommendation.
What kind of strategy is being proposed?
Scroll sideways to see every column. Keyboard: focus the table region and use the arrow keys.
| Strategy | What it is trying to capture | What the explanation must include |
|---|---|---|
| Market making | A difference between purchase and sale prices | Inventory that remains unsold, adverse selection, cancellations and all costs |
| News-based trading | A change in an estimated outcome probability | Source reliability, time available to act and the possibility the estimate is wrong |
| Statistical relative value | A relationship expected to hold across prices | Why it should hold, how it can break and whether the contracts pay in the same states |
| Payoff matching | A minimum combined settlement payment | An exhaustive state table, matching rules, executable legs and settlement risks |
| Wallet copying | Someone else's observable activity | Hidden hedges, incomplete inventory, execution delay and a reason to study that selection |
None of these labels supplies a return forecast. A fixed $0 or $1 settlement payment does not tell you the probability of either outcome. A copied account is not automatically lower risk, and a quiet market is not automatically an opportunity for a human trader.
Start with a genuine implication
Define a fictional event A as “the named team reaches the final” and B as “the same team wins that final.” Under an ordinary competition format, B implies A. To use that relationship between two contracts, their rules must refer to the same team, competition and edition, with compatible cancellation and resolution treatment. Do not assume the wording alone proves those conditions.
For probabilities under that definition, P(B) ≤ P(A). A 60% probability of reaching the final and a 40% probability of winning it are consistent. The larger probability of the earlier stage is not an arbitrage violation. For a conjunction, P(A and C) ≤ P(A) also holds; adding marginal probabilities is not how to calculate the conjunction.
One hypothetical basket, every state
Assume the implication above is exact and settlement is binary. Consider one YES-A share plus one NO-B share. The following is a payoff table, not a quote or an instruction to buy:
Scroll sideways to see every column. Keyboard: focus the table region and use the arrow keys.
| Settlement state | YES A | NO B | Combined payment |
|---|---|---|---|
| Team does not reach the final | $0 | $1 | $1 |
| Team reaches the final but does not win | $1 | $1 | $2 |
| Team wins the final | $1 | $0 | $1 |
The state B=true and A=false is excluded only by the assumed implication. If contract rules permit that combination, the basket can pay zero and the argument fails. Add cancellation, invalid-market and dispute states if the actual rules allow them; do not hide them in a footnote.
Suppose the two executable asks are an invented $0.55 and $0.40. Equal share counts cost $0.95, so the minimum gross difference is $0.05 per pair under the stated settlement assumptions. Cash allocation is proportional to those prices: $55 and $40 buy 100 shares of each. Splitting $95 equally does not produce equal share counts.
If total assumed costs are $0.02 per pair, the modeled minimum net difference is $0.03. If costs are $0.06, it is -$0.01. If only one leg fills, there is no complete basket. A price displayed earlier, a mid-price, or a bid you cannot sell into is not the pair of executable asks used in this calculation.
Why the same argument does not cover related headlines
A candidate's nomination and a general-election win may have exceptions to their apparent relationship. A party-win contract and a named-candidate contract are not identical. A spot-asset position and a price-threshold contract have different payoffs. Draw their tables before calling one a hedge for the other. Correlation in historical observations is not a settlement guarantee.
The required evidence is specific: contract identifiers, full rules and versions, state definitions, share quantities, available depth, costs and the treatment of an incomplete fill. “The prices should converge” is a statistical hypothesis, not a substitute for those fields.
A playground that can reveal failure
Keep research separate from funded execution. Store a timestamped snapshot of the rules and observed book, then replay only information available at the decision time. Record stale data, missing updates, partial fills, rejected orders and unresolved outcomes alongside successful cases. Inventory limits and a persistent stop state matter because restarting a process must not silently erase its reason for stopping.
A fork of an on-chain state alone does not reconstruct an off-chain order book, queue position, cancellations or a real dispute process. Label every simulated fill assumption. An acknowledgment, fill and settlement are separate events; an unknown request outcome requires reconciliation before another attempt.
Use the offline paper-scenario example to inspect assumptions without an account or order submission, and the latency measurement guide to distinguish receive-time gaps from execution advantage. Neither demonstrates profitability. This page offers no empirical bot-dominance estimate, private test history or expected annual return.
Educational material, not financial advice. Verify eligibility and current platform rules; do not use location masking to evade restrictions. Affiliate relationships elsewhere on this site are explained in the disclosure; a commercial link is not evidence for a payoff claim.