Calculating Predictions

In this example we will take a look at how you can parse the the state of a market and calculate the current prediction based on the outcome asset prices.

1. Set up

We boot up the sdk in full context mode.

import {
  calcSpotPrice,
  create,
  FullContext,
  getIndexOf,
  IOMarketOutcomeAssetId,
  mainnet,
  Market,
  Sdk,
} from "@zeitgeistpm/sdk";

const sdk: Sdk<FullContext> = await create(mainnet());

2. Fetch Market and Pool

Then we fetch the market and pool we want to query the prediction for.

const marketId = 90;

const market: Market<FullContext> = await sdk.model.markets
  .get({ marketId })
  .then((market) => market.unwrap()!);

const pool = await sdk.model.swaps
  .getPool({ marketId })
  .then((pool) => pool.unwrap()!);

3. Fetching Base asset weight and balance.

Next we get the assets of the pool and fetch the pools base asset balance and base asset weight.

const assets = pool
  .getAssetIds()
  .filter(IOMarketOutcomeAssetId.is.bind(IOMarketOutcomeAssetId));

const poolBaseAssetBalance = await pool.getAssetBalance({ Ztg: null });
const poolBaseAssetWeight = await pool.getAssetWeight({ Ztg: null }).unwrap()!;

4. Calculating Prices

Now we can iterate through the markets assets, fetch their balances, weights and correlated market metadata category. Then we can calculate the price for the given assets.

info

The zeitgeist node also supports rpc calls to fetch asset prices. Read more about them here.

const assetPrices = await Promise.all(
  assets.map(async (asset) => {
    const assetBalance = await pool.getAssetBalance(asset);
    const assetWeight = pool.getAssetWeight(asset).unwrap()!;
    const category = market.categories?.[getIndexOf(asset)];

    const price = calcSpotPrice(
      poolBaseAssetBalance,
      poolBaseAssetWeight,
      assetBalance,
      assetWeight,
      0
    );

    return {
      name: category?.name ?? asset.toString(),
      price,
    };
  })
);

note

Note that the category metadata can be missing on markets.

5. Printing the current prediction

Now we can get the highest asset price and print the current prediction.

info

Here we are assuming the market is a Categorical market. To see how you can calculate the value if the market is a Scalar market, jump to the next section.

const predictedPrice = assetPrices
  .sort((a, b) => (a.price.gt(b.price) ? -1 : 1))
  .at(0)!;

console.log(
  `${market.question}`,
  `The prediction is ${predictedPrice.name} at ${predictedValue.toFixed(2)}`
);

Based on the market data at time of writing the output will be:

Will a DEX be deployed on Statemint or Statemine before the end of Q2 2023?
The prediction is `no` at 0.80767457218061127617

Full Code

Go to the full code snippet for this example