| Market | Side | Now | Value | P&L | ||
|---|---|---|---|---|---|---|
| No positions yet. Buy yes or no on any market. | ||||||
imd.bet is a market for questions with two answers. You buy the side you think is right. If you are right, each share you hold pays out one $BET. If you are wrong, it pays nothing.
There is no bookmaker setting a line and no house taking the other side. Every price on the site is the last price someone paid another trader. When you see 62c on yes, it means the last person to buy yes paid 62 cents worth of $BET for a share that pays 1 $BET if the answer turns out to be yes.
That is the whole product. The rest of these docs is detail about how prices move, how a bet is placed, how a market ends and how the payout is worked out.
Every market has exactly two shares: a yes share and a no share. One of them will be worth 1 $BET when the market resolves and the other will be worth 0.
Because one of the two always wins, a yes share and a no share held together are always worth exactly 1 $BET. That is why the two prices on a market always add up to 100c:
yes 62c + no 38c = 100c yes 77c + no 23c = 100c yes 9c + no 91c = 100c
If yes is trading at 62c then no is trading at 38c, always, on every market, at every moment. You never have to look up the other side. Subtract from 100.
A price is a probability with a currency sign on it. 62c means the market thinks there is roughly a 62% chance the answer is yes.
| Price | What the market believes | Payout per share | Return if right |
|---|---|---|---|
| 5c | Very unlikely | 1 $BET | 1,900% |
| 25c | Unlikely | 1 $BET | 300% |
| 50c | A coin flip | 1 $BET | 100% |
| 75c | Likely | 1 $BET | 33% |
| 95c | Nearly certain | 1 $BET | 5% |
The cheaper the side, the more it pays if it comes in, because you are buying something the rest of the market doubts. The return figure in the ticket is worked out as (1 / price) - 1.
Under the chart on every market there are two columns of bids. The left column is what people are willing to pay for yes shares. The right column is the same for no shares.
Each row is a price and the number of shares wanted at that price. The bar behind the row shows the size relative to the rest of the book, so a long bar means there is real money sitting at that level and a short bar means there is not.
Two things are worth reading off the book before you buy:
From the market page:
Bets are placed in $BET. Other tokens sent in are not counted as bets.
When you press buy, the site builds one transaction and hands it to your wallet. Nothing is signed before you press the button, and the position is only recorded after the wallet returns a transaction hash.
While the wallet is open the button reads confirm in wallet and is locked, so the same bet cannot be sent twice by accident. If you reject in the wallet, nothing is recorded and the ticket goes back to how it was.
The transaction hash is kept with the position, so every entry in your positions table can be traced back to the transaction that opened it.
It helps to think in shares rather than in the amount you put in. The amount you type buys a number of shares at the current price:
shares = amount / price 100 $BET at 40c = 250 shares 100 $BET at 50c = 200 shares 100 $BET at 80c = 125 shares
Those shares are what settles. 250 shares pay 250 $BET if the answer is yes, whatever the price does in between.
Buying yes takes yes shares off the book, which lifts the yes price and lowers the no price by the same amount. Selling does the reverse.
How far it moves depends on the size of your order against the depth of the book. A small order in a deep market barely moves it. A large order in a thin market walks up several levels and your average fill sits above the price you first saw.
This is the reason the ticket reports an average price rather than the top of the book. It is what you actually paid across every level your order touched.
Open positions show in the positions table with:
| Column | What it means |
|---|---|
| Side | Yes or no, the side you hold |
| Shares | How many shares you hold, each paying 1 $BET if right |
| Avg | The average price you paid across all your buys on that side |
| Now | The current market price for that side |
| Value | Shares multiplied by the current price |
| P&L | Value minus what you paid |
Buying the same side twice adds to one position and averages the price, rather than opening a second row. Clicking a row loads that market back into the market panel.
You do not have to wait for a market to resolve. Shares can be sold back into the book at any point while the market is open, at whatever the price is then.
Selling early locks in the move in the price rather than the full payout. An example: you buy yes at 30c, news lands, yes trades at 70c.
hold to resolution, right 100 shares pay 100 $BET (paid 30, made 70) hold to resolution, wrong 100 shares pay 0 (paid 30, lost 30) sell now at 70c 100 shares pay 70 $BET (paid 30, made 40)
Selling gives up the last 30c of upside and removes the risk of the last 30c of downside. Which is right depends on how sure you are, not on which number is bigger.
Every market has a closing time, shown on the card and in the market panel as a countdown. The sequence is always the same:
| State | What can happen |
|---|---|
| Open | Buying and selling, prices move freely |
| Closed | Trading stops, positions are frozen as they stand |
| Resolving | The outcome is determined from the source named in the market |
| Resolved | Winning shares become claimable at 1 $BET each |
A market closes at its stated time whether or not the underlying event has happened yet. The gap between close and resolution is the time it takes for the answer to be known and checked.
Every market names the source that decides it before it opens. The wording of the question is what gets resolved, not the spirit of it, so read it closely. "Above $4,000 on the last day of the month" means the price at that moment, not the highest price during the month.
Three kinds of source cover nearly every market:
Questions are written so that one of those three can answer them with a yes or a no. If a question cannot be answered that way, it does not become a market.
Once a market resolves, every share on the winning side is worth exactly 1 $BET and every share on the losing side is worth 0. The payout is:
payout = winning shares x 1 $BET
Worked through, on a market that resolves yes:
| Position | Paid | Shares | Payout | Result |
|---|---|---|---|---|
| Yes at 25c | 100 $BET | 400 | 400 $BET | +300 |
| Yes at 60c | 100 $BET | 166.7 | 166.7 $BET | +66.7 |
| Yes at 90c | 100 $BET | 111.1 | 111.1 $BET | +11.1 |
| No at 40c | 100 $BET | 250 | 0 | -100 |
The price you paid does not change the payout. It only changes how many shares your money bought.
$BET is the token every market is priced, traded and settled in. Prices are quoted in cents of $BET, so a 62c share costs 0.62 $BET and pays 1 $BET.
The $BET on Etherscan button at the top of the page opens the token contract, where the supply, holders and every transfer can be read directly.
Two costs apply to a bet, and both are visible before you confirm:
There is no separate charge for opening a position, and no charge for holding one to resolution.
The console at the top of the page prints bets as they happen: the wallet, the market, the side and the size. Green squares are yes, rust squares are no.
The stat line above it carries three running figures:
The chart plots the yes price over the last 90 ticks on a scale from 0c to 100c, so the height of the line is the probability. The dashed line marks the current price and the dot on the right edge is the latest tick.
Under it sit the two bid columns. To the right is the ticket. Everything in the panel is about one market at a time, and clicking any card in the grid or any row in your positions swaps which one.
One full example, with every number shown.
The market is "Will more than 25,000 wallets claim an airdrop this week?". It closes Sunday at midnight and resolves from an on chain count of unique claiming addresses. Yes is trading at 77c, so no is at 23c.
You think 25,000 is low and the count clears it early. You buy yes with 150 $BET.
side yes price 77c amount 150 $BET shares 150 / 0.77 = 194.8 payout 194.8 $BET if yes return (1 / 0.77) - 1 = 30%
The count passes 18,000 midweek. Yes trades up to 89c. Your position is now worth 194.8 x 0.89 = 173.4 $BET, so you are up 23.4 on paper. Nothing has settled. You can sell at 89c or hold.
The count slows near 24,000 and yes falls back to 61c. Your position is worth 194.8 x 0.61 = 118.8 $BET, down 31.2 from what you paid. Same shares, different price.
The final count is 25,904. The market resolves yes.
shares 194.8 payout 194.8 x 1 = 194.8 $BET paid 150 $BET result +44.8 $BET
Every number between Monday and Sunday was noise. The only two that decided the outcome were the price you paid and the answer.
A question is only worth trading if two people who disagree about the world still agree about what the question means. That takes three things:
Wording is read literally at resolution. "More than 25,000" does not include exactly 25,000. "Before Friday" ends when Friday begins. "Under 10 gwei all weekend" fails on a single block above 10. If a phrase can be read two ways, the market is rewritten before it opens rather than argued about after it closes.
Liquidity is how much you can buy without moving the price. It shows up in two places: the size in the book and the gap between the best bid on each side.
| Book | What you see | What it costs you |
|---|---|---|
| Deep, tight | Thousands of shares within a cent or two | Your fill is close to the price on the card |
| Thin, tight | Small size, narrow gap | Fine for small orders, slips on large ones |
| Thin, wide | Little size, several cents of gap | You pay the gap on the way in and again on the way out |
Volume on the card tells you whether the market has been traded at all. A market with 500k $BET of volume and a tight book has been argued over. A market with 5k has not, and its price is closer to a guess than a consensus.
Four approaches cover most of what happens on a prediction market.
The first pays for being right. The last pays for being present. They are different jobs.
Markets on the same subject move together, and when they disagree with each other, one of them is mispriced.
If "BTC above $100k in June" trades at 40c while "BTC above $120k in June" trades at 45c, the second cannot be more likely than the first, because every world where BTC is above 120 is also a world where it is above 100. Buying no on the second and yes on the first captures that, whatever BTC actually does.
The same logic applies to time. A market closing later should never be cheaper than the same question closing sooner, because it has all the same paths plus extra time.
The connect button asks your browser wallet for an account. Once connected, the button shows the first and last characters of the address.
You can browse every market, read every book and work through every ticket without connecting. A wallet is only needed at the moment a bet is placed, and pressing buy while disconnected asks to connect first, then continues into the same transaction.
If no wallet is installed, the button says so rather than failing quietly.
The chart keeps the last 90 ticks. A tick is written whenever the price changes, whether from a trade or from the book being repriced, so a busy market fills the chart faster than a quiet one.
The sparkline on each card shows a shorter window, and its colour comes from the direction over the last 12 ticks: green when the price is higher than it was, rust when it is lower. The figure beside the volume is that same move in cents.
| Term | Meaning |
|---|---|
| Share | A claim on 1 $BET if its side is right |
| Yes / no | The two sides of a market, always adding to 100c |
| Price | What a share costs now, in cents of $BET |
| Average price | What you actually paid per share across all levels filled |
| Book | The list of bids waiting at each price |
| Depth | How many shares are wanted near the current price |
| Close | The moment trading stops on a market |
| Resolution | Determining the answer from the named source |
| Settlement | Paying 1 $BET per winning share |
| P&L | What a position is worth now minus what it cost |