#4 A Busy Month for the Pyth network

Welcome to the 4th issue of the Pyth substack: HiFi for DeFi. Today's blog piece focuses on 3 major events for Pyth network this month.


The Pyth network is a vision coming to life: one that imagines the world of decentralized finance (DeFi) gaining comprehensive access to high-fidelity (HiFi) financial markets data securely and reliably.

In other words, the Pyth network is a specialized oracle network that focuses on sourcing continuous real-world and crypto-related market data originating off-chain and streaming it on-chain at sub-second speeds for smart contract consumption regardless of their blockchains.

This is an ambitious goal because financial market data is sufficiently unique: there are very few available sources, and those sources have very tightly controlled distributions. In addition, the oracle network needs to be able to combine such latency-sensitive data in way that optimizes not only accuracy but which increases security as well given how dependent many blockchain applications are on the accuracy of this type of data.

Website, Medium, Twitter & Discord

What happened on September 7th, 2021?

NEW YORK, Sept 7 (Reuters) - The price of cryptocurrencies plunged and crypto trading was delayed on Tuesday, a day in which El Salvador ran into snags as the first country to adopt bitcoin as legal tender.

Bitcoin and ether plummet 17% as broad sell-off batters crypto on the day El Salvador adopts bitcoin as legal tender (source Markets Insider)

Analysts point to overleveraged traders after Bitcoin flash crashes to $43K. Traders were hit hard as $3.5 billion in liquidations took place as Bitcoin price crashed below $43,000. (source CoinTelegraph)

For seasoned crypto participants, a 17% selloff on its major coin (Bitcoin) is not such a rare event:

  • During the "COVID crash", on the 14th March 2020, BTC quickly lost about 40% of its value in a one-day timeframe.

  • On January 21st, 2021, Bitcoin also dropped by 20% — less than 24 hours breaking its previous ATH.

On the other hand, for "traditional" and established markets and assets, such events would not even occur most of the time due to the trading halt rule. For example, NYSE imposes three trading curb levels – 7%, 13%, and 20%. If a 7% or 13% drop in the S&P 500 occurs during a single trading day, then all trading on the exchange is stopped for a period of 15 minutes. If the 20% drop level is hit, then all trading on the exchange is stopped for the rest of the trading day.

But more than the “appropriate valuation” of an asset, the more interesting part for the Pyth network (and the protocols that rely on us) was to see how both our individual data publishers and the overall Pyth network was able (and also not able) to reflect a sharp price evolution in such a short time frame.

For comparison purposes, looking back at the spot price of BTC(/USD) on the top 10 worldwide crypto exchanges (which represents on average about 75% of the total volume traded), anyone had a few minutes to purchase one Bitcoin for approximately $43,000 (varying across exchange) on September 7th, 2021.

Compare the BTC/USD spot price graphs for FTX, Binance, and Bybit (respectively):

Below is a snapshot of the Pyth network aggregated price feed for BTCUSD taken during the September 7th selloff. The lowest Pyth BTCUSD price aggregated and distributed to the Solana blockchain (and apps leveraging our feeds) was almost exactly similar to the drop experienced by all stakeholders that day: $43,000. Reviewing other oracles and data sources proves that not all price feeds were able to keep up with this volatility.

This demonstrates well that both our individual 1st party-data providers (more than a dozen were active that day on this specific feed, with exchanges and trading firms among them), as well as the Pyth aggregated price feed, managed to match this highly volatile market environment with a great level of accuracy.

This event strongly affirms our decision to onboard only 1st party data providers — which often, if not always, have the best and fastest access to market data, and are able to serve it to the Pyth network and (by extension) our feed recipients with a delay almost indistinguishable from real life. Likewise, this experience also adds conviction in our choice to not rely on 3rd party data providers, who often have access to delayed data or are unable to match such a highly volatile environment. While third-party data may be perfectly sufficient for certain protocols, the aim of Pyth is to provide high fidelity reliable data from institutional grade sources in order to serve a more complex and sophisticated market.  

What happened on September 20th, 2021?

If 9/7 was a testament to our core infrastructure, then 9/20 was a reminder of the work still to be done. Between 12:21 and 12:23 UTC on Monday, Sep. 20, 2021, the Pyth BTC/USD aggregate price had several sharp dips below $40,000, reaching as low as $5,402, and the confidence intervals became extremely wide. Here is a graph of the Pyth BTC/USD price series for the time period in question, plus two minutes on either side.

Several Solana programs relying on Pyth prices were impacted by this incident. The impact was exacerbated due to some programs relying on the aggregate price feed without using the confidence, which allowed liquidations to occur even though the published price was highly uncertain.

The issue was caused by the combination of (1) two different Pyth publishers publishing a near-zero price for BTC/USD and (2) the aggregation logic overweighting these publishers’ contributions.

Pyth core developers are taking several steps to prevent these issues from happening again:

  • Software changes to reduce the probability that publishers produce incorrect prices due to software errors, including developing a suggested integration testing protocol that publishers can run to validate their software changes.

  • Adjustment of the aggregation logic to properly weight prices that span a large range of values. The new aggregation logic will derive its weights from relative price differences, which will prevent the naturally smaller confidence intervals of lower prices from overly affecting the aggregate.

  • Enhanced documentation, best practices, and example code for protocols integrating with Pyth focusing on the importance of utilizing both the price and confidence interval as the best way to accurately reflect the market.

You may review the full root cause analysis here.

This event only fueled our motivation to continue onboarding the strongest first-party data providers and building the most robust oracle available. What does not kill you makes you stronger, and the lessons from this incident will bolster Pyth into a stronger state.

What happened on September 22nd, 2021?

Supported by our great partners at Jump, Cumberland, LMAX, Hudson River Trading, and Jane Street, the Pyth community gathered for its inaugural event in New York City on the back of the Messari Mainnet conference. It was truly amazing to see all of our amazing partners come together (in person) alongside others who were interested to learn more about Pyth. The sizeable attendance was testimony to the energy and enthusiasm behind this growing project. The gathering further fueled our commitment to Pyth and our belief in the long-term impact it can have on the industry.

New Providers

Another month in, and the stream of newly onboarded data providers is still flowing. In the last newsletter, we presented probably three of the worldwide biggest financial providers (Jane Street, Susquehanna, and Hudson River Trading). This time around, we welcome five significant players who are more oriented towards cryptocurrency markets.

These new additions (to our already incredible roster of market participants) will provide 24/7 high-fidelity cryptocurrency market data, further strengthening the quality and reliability of the Pyth aggregated crypto feeds.


  • Bitstamp is the world’s longest-running cryptocurrency exchange, supporting investors, traders, and leading financial institutions since 2011.

  • Bitstamp has consistently ranked in the top 10 spot markets in terms of volume for almost a decade and is now trusted by over 4 million customers.

DV Chain

  • DV Chain is an over the counter (OTC) trading platform offering deep liquidity and cutting-edge technology to institutional clients, professional traders, and exchanges

  • DV Chain is an affiliate of DV Trading, a Chicago-based proprietary trading firm, and also an affiliate of Independent Trading Group, a Canadian broker-dealer.

  • DV Chain provides 24/7 liquidity and market-making services to institutional clients and exchanges within the cryptocurrency ecosystem.

Amber Group

  • Amber Group is one of the world's leading crypto finance service providers, operating 24/7 with a presence in Hong Kong, Taipei, Seoul, and Vancouver.

  • Founded in 2017, the Amber Group has over $1.5 billion in AUM.

  • Since Amber Group has cumulatively traded over $500 billion and is now integrated with 100+ exchanges and trading venues across the globe.

Galaxy Digital

  • Galaxy Digital Holdings Ltd. is a diversified financial services and investment management company in the digital asset, cryptocurrency, and blockchain technology sectors. Galaxy Digital (TSX: GLXY) is a publicly-traded company on the Toronto Stock Exchange

  • In partnership with Alerian and S-Network Global Indices, Galaxy Digital has recently launched 8 blockchain indexes for passive investors to access the emerging digital-assets ecosystem. Constituents include Facebook, Alphabet, Microsoft, Mastercard, Square, and Coinbase.

  • Galaxy Digital also partners with Invesco to offer a Bitcoin Exchange-Traded Fund (ETF) in the United States. The filing is still currently pending the SEC decision.

GMO Coin

  • GMO Coin was established in October 2016 as a crypto assets exchange business company and registered as a crypto assets exchange in July 2018.

  • GMO Coin is also registered as a Type I Japanese Financial Instruments Business since May 2020

  • GMO Coin has a total volume traded of about $500 million traded on a daily basis across its various assets — with all markets being traded are versus the Japanese Yen

We still look forward to welcoming many more reputable and high-quality data publishers to our network and co-build the next digital marketplace of institutional grade, high fidelity financial data.

Find all our existing data publishers on our website.

That's all for our 4th newsletter — Thank you for reading, and don’t forget to subscribe and share!


We can’t wait to hear what you think! Feel free to join the Pyth Discord server, follow Pyth on Twitter, join the Telegram and the Substack or the Medium to learn more, and ask any questions you may have.