Curious facts from Michael@Curve
11.02.202516:02
Few thoughts about DeFi and the future of Finance
https://cointelegraph.com/news/institutional-players-shape-the-future-of-de-fi
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
https://cointelegraph.com/news/institutional-players-shape-the-future-of-de-fi
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
09.01.202501:35
Recently, I've heard people referring to LVR (Loss Versus Rebalance) framework a lot [https://anthonyleezhang.github.io/pdfs/lvr.pdf]. This is a famous paper done in Columbia University, University of Chicago and a16z research. A good summary is presented by Delphi [https://members.delphidigital.io/learn/loss-versus-rebalancing-lvr].
I believe that main interpretation of the results found in this paper is not correct (although the framework might be useful):
- Paper assumes that most of "losses" in CFMMs (Constant-Function Market Makers) occur towards arbitrageurs who try to earn profits from arbitraging between a CEX and a DEX. In reality, if prices on CEXes changed not in steps but as a continuous function, immediate arbitrage would having zero profit, but if CFMM had zero fee, one would realize the same zero profit or loss if price returned back to the same value after some time. So arb trader would have made zero profit, as well as the pool would have made the same PnL as the same pool brought from one price to another along any trajectory. Only true for a zero-fee pool, but this just disproves the assumption that "everything becomes magically good when arb traders are not making any money".
- Paper assumes that the best AMM would be realized if one always tweaked the reserves to ALWAYS have the same price as price on a CEX. In reality, "informed flow" on such a DEX would be always zero, so the AMM would likely make near-zero volume.
- There seem to be a silent assumption by some researchers that AMMs only make money on "uninformed flow" or "noise trades" - something which is NOT arbitrage. I think that one should ONLY assume that all trades are arbitrage and that "uninformed flow" is zero.
- I propose (and do it since 2019!) to calculate "profit on top of zero-fee CFMM bonding curve", and only simulate arbitrage between the "real" (non-zero-fee) bonding curve and external exchanges. The "ground zero" (zero-fee bonding curve) will be universal and not dependent on how exactly we got from price A to price B, but the "real" CFMM will have trajectory-dependent, volatility-dependent profits. Moreover, this model ALWAYS shows profits and never shows losses.
- "Impermament loss" can be dealt with separately, but within my model it would not be considered a loss.
If there are any DeFi theory researchers here - please tell if I should find some time to maybe make a writeup on my approach to simulating CFMMs.
I believe that main interpretation of the results found in this paper is not correct (although the framework might be useful):
- Paper assumes that most of "losses" in CFMMs (Constant-Function Market Makers) occur towards arbitrageurs who try to earn profits from arbitraging between a CEX and a DEX. In reality, if prices on CEXes changed not in steps but as a continuous function, immediate arbitrage would having zero profit, but if CFMM had zero fee, one would realize the same zero profit or loss if price returned back to the same value after some time. So arb trader would have made zero profit, as well as the pool would have made the same PnL as the same pool brought from one price to another along any trajectory. Only true for a zero-fee pool, but this just disproves the assumption that "everything becomes magically good when arb traders are not making any money".
- Paper assumes that the best AMM would be realized if one always tweaked the reserves to ALWAYS have the same price as price on a CEX. In reality, "informed flow" on such a DEX would be always zero, so the AMM would likely make near-zero volume.
- There seem to be a silent assumption by some researchers that AMMs only make money on "uninformed flow" or "noise trades" - something which is NOT arbitrage. I think that one should ONLY assume that all trades are arbitrage and that "uninformed flow" is zero.
- I propose (and do it since 2019!) to calculate "profit on top of zero-fee CFMM bonding curve", and only simulate arbitrage between the "real" (non-zero-fee) bonding curve and external exchanges. The "ground zero" (zero-fee bonding curve) will be universal and not dependent on how exactly we got from price A to price B, but the "real" CFMM will have trajectory-dependent, volatility-dependent profits. Moreover, this model ALWAYS shows profits and never shows losses.
- "Impermament loss" can be dealt with separately, but within my model it would not be considered a loss.
If there are any DeFi theory researchers here - please tell if I should find some time to maybe make a writeup on my approach to simulating CFMMs.
15.12.202414:58
― No, I'm just resting. This is all because someone is too narrow of a door.
― No, all because someone eats too much.
// Winnie the Pooh
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
Кайра бөлүшүлгөн:
TON Society
06.10.202417:13
Curve Finance x TON Livestream: Don’t Miss It!
We’re excited to share that Curve Finance has recently announced its collaboration with the TON Foundation to integrate Curve’s constant-function market maker (CFMM) technology into the TON Blockchain. This partnership aims to streamline the swap process and enhance the overall user experience on TON.
As a next step, join us for a special livestream featuring Curve Finance Founder, Michael Egorov, alongside representatives from the TON Foundation.
▪️ Date: October 7th
▪️ Time: 18:00 CEST
▪️ Live on: YouTube & Telegram
During the stream, we’ll cover:
▪️ Curve’s journey and evolution in the market.
▪️ TON’s narrative, achievements, and upcoming plans.
▪️ The significance of the Curve-TON collaboration and why stable swaps are essential for the TON ecosystem.
Don’t miss out — set a reminder and join us to learn about this pivotal development!
We’re excited to share that Curve Finance has recently announced its collaboration with the TON Foundation to integrate Curve’s constant-function market maker (CFMM) technology into the TON Blockchain. This partnership aims to streamline the swap process and enhance the overall user experience on TON.
As a next step, join us for a special livestream featuring Curve Finance Founder, Michael Egorov, alongside representatives from the TON Foundation.
▪️ Date: October 7th
▪️ Time: 18:00 CEST
▪️ Live on: YouTube & Telegram
During the stream, we’ll cover:
▪️ Curve’s journey and evolution in the market.
▪️ TON’s narrative, achievements, and upcoming plans.
▪️ The significance of the Curve-TON collaboration and why stable swaps are essential for the TON ecosystem.
Don’t miss out — set a reminder and join us to learn about this pivotal development!
06.02.202521:12
Gas limit on Ethereum was raised very slightly (by 3%), and tx price on Ethereum went much more notably down [https://www.coindesk.com/markets/2025/02/04/ethereum-raises-gas-limits-for-first-time-since-2021-boosting-eth-appeal]. What's happening?
Apart from the fact that necessary improvements were made to do this safely, it's a reflection of a curious property of "markets for a limited resource". Imagine the opposite: demand for blockspace going up slightly. This happens sometimes, and gas fee can spike up x100-x1000 at those times! And with gas limit slightly up, we see the opposite: tx fee significantly down.
We see the same when there's a fight for electricity costs in real markets [https://fortune.com/2024/05/18/texas-power-prices-1600-percent-heat-wave-record-energy-demand-electric-grid/], so that should not surprize. That would not have happened if there was a good storage for electricity over there, or if it was possible to buy enough somewhere else, for example.
Chains are software. Why wouldn't it be possible to "pre-buy" block space on Ethereum natively, to be able to spend later? Maybe protocols for who it is important (e.g. lending protocols) could scoop blockspace reserves cheap for users, to be used when gas is too high, so that users can close or adjust positions cheaper? Of course, real blockspace is limited, so it'd increase prices further for those who doesn't have such reserves. But still, sounds like a way to get this scarce resource for who need it most when they need it most.
Apart from the fact that necessary improvements were made to do this safely, it's a reflection of a curious property of "markets for a limited resource". Imagine the opposite: demand for blockspace going up slightly. This happens sometimes, and gas fee can spike up x100-x1000 at those times! And with gas limit slightly up, we see the opposite: tx fee significantly down.
We see the same when there's a fight for electricity costs in real markets [https://fortune.com/2024/05/18/texas-power-prices-1600-percent-heat-wave-record-energy-demand-electric-grid/], so that should not surprize. That would not have happened if there was a good storage for electricity over there, or if it was possible to buy enough somewhere else, for example.
Chains are software. Why wouldn't it be possible to "pre-buy" block space on Ethereum natively, to be able to spend later? Maybe protocols for who it is important (e.g. lending protocols) could scoop blockspace reserves cheap for users, to be used when gas is too high, so that users can close or adjust positions cheaper? Of course, real blockspace is limited, so it'd increase prices further for those who doesn't have such reserves. But still, sounds like a way to get this scarce resource for who need it most when they need it most.
08.01.202512:24
Seeing a lot being written about Terra and Do Kwon out there. So here's my take.
Design of Terra was based on Basis Cash where the stablecoin is backed by its own protocol token. Moreover, the system would sell the protocol token (Luna) to prevent the stablecoin from becoming under-collateralized if collateral price is too much down. This design is inherently unstable.
Backing by protocol token was tried before - it was called NuBits [https://nubits.com/]. Eventually, it's depegged forever, examplifying that this is not a good idea. It could have been ok if Luna's use case was much beyond just Terra, but near the end its only purpose was just backing UST. Terra introduced the fund called LFG to address this risk (and to diversify backing to BTC and ETH), but it was too little too late (also totally not decentralized).
Many Terra investors (reputable VCs!) were advertising it, so I started questioning - am I wrong? Are they right? Anyhow, it appeared that I should have trusted my own assessment and not listening to KOLs of that time.
When Terra collapsed - some people lost money, but money cannot just disappear. They just changed pockets. Who made money on Terra collapse? Apparently it is everyone who bought Luna lower and sold higher! Assets can have "paper value" much exceeding their real value, but after the sale of those assets, the true distribution becomes apparent. So victims of Terra collapse lost their money to numerous Luna investors who were fast enough to take profits before the collapse - likely including Luna VCs (although perhaps not deliberately from their side).
Design of Terra was based on Basis Cash where the stablecoin is backed by its own protocol token. Moreover, the system would sell the protocol token (Luna) to prevent the stablecoin from becoming under-collateralized if collateral price is too much down. This design is inherently unstable.
Backing by protocol token was tried before - it was called NuBits [https://nubits.com/]. Eventually, it's depegged forever, examplifying that this is not a good idea. It could have been ok if Luna's use case was much beyond just Terra, but near the end its only purpose was just backing UST. Terra introduced the fund called LFG to address this risk (and to diversify backing to BTC and ETH), but it was too little too late (also totally not decentralized).
Many Terra investors (reputable VCs!) were advertising it, so I started questioning - am I wrong? Are they right? Anyhow, it appeared that I should have trusted my own assessment and not listening to KOLs of that time.
When Terra collapsed - some people lost money, but money cannot just disappear. They just changed pockets. Who made money on Terra collapse? Apparently it is everyone who bought Luna lower and sold higher! Assets can have "paper value" much exceeding their real value, but after the sale of those assets, the true distribution becomes apparent. So victims of Terra collapse lost their money to numerous Luna investors who were fast enough to take profits before the collapse - likely including Luna VCs (although perhaps not deliberately from their side).
04.12.202422:43
Just routinely relocking my veCRV positions till Nov 30, 2028
17.09.202417:22
Here's the schedule of my talks on Token 2049:
- BTC ECO Summit: "Is Bitcoin the best prime asset to use in DeFi?". 18 Sep, 9:20, MARINA BAY SANDS，4 / F
- Token 2049: "Stablecoin swaps: a pathway from jurisdictional arbitrage to foreign exchange markets". 19 Sep, 17:20, Sky Stage
- Stablecoin Summit: fireside chat. 17:45, Andaz Singapore
Livestreams will appear on https://www.youtube.com/channel/UCFOIkOy5HWzhPmSHs7ZBaQQ
- BTC ECO Summit: "Is Bitcoin the best prime asset to use in DeFi?". 18 Sep, 9:20, MARINA BAY SANDS，4 / F
- Token 2049: "Stablecoin swaps: a pathway from jurisdictional arbitrage to foreign exchange markets". 19 Sep, 17:20, Sky Stage
- Stablecoin Summit: fireside chat. 17:45, Andaz Singapore
Livestreams will appear on https://www.youtube.com/channel/UCFOIkOy5HWzhPmSHs7ZBaQQ
29.01.202521:33
Interesting about Thorchain. A reminder to all builders: the protocol should be resilient to protocol token going to zero. If that can cause users going bankrupt - it will happen!
We saw that with Terra and USDN, we are seeing it with Rune
https://rekt.news/thorchains-iceage/
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
We saw that with Terra and USDN, we are seeing it with Rune
https://rekt.news/thorchains-iceage/
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
31.12.202413:31
Happy New Year everyone!
According to DeFi calendar, it is gonna be a year ofVyper Snek, so hopefully great yields ahead 😃
According to DeFi calendar, it is gonna be a year of
30.10.202423:21
Using CrossCurve bridge. Always nice to make some ETH on the way when bridging 😅
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
18.01.202513:15
Seeing many people going crazy about Trump launching a memecoin on Solana and not Ethereum.
Taking aside that the fact of Donald launching a memecoin is gross by itself, Solana is an obvious choice for memecoins. Memecoins are a casino thing, purchased by typical casino visitors who live on Solana. Ethereum is for serious things like DeFi because that's where people keep their crypto safe.
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
Taking aside that the fact of Donald launching a memecoin is gross by itself, Solana is an obvious choice for memecoins. Memecoins are a casino thing, purchased by typical casino visitors who live on Solana. Ethereum is for serious things like DeFi because that's where people keep their crypto safe.
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
19.12.202407:24
Curious liquidation here!
- It is acutally Uwu, not Aave;
- Liquidated CRVs do not exist!
How? In June, Uwu platform where I had a position got hacked. Since then, these CRVs backing my position haven't been returned. Recently I've made a bet that veCRV wrappers will repeg faster than repayment of the stolen CRVs on Uwu will happen, so I borrowed maximum against nonexistent CRV and bought veCRV wrappers with a discount using those funds.
Of course, when was possible, operators of Uwu liquidated those nonexistent CRV. As a result, they don't have to repay me that liquidated part of stolen funds in future, but they also make sure that lenders don't have any bad debt.
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
- It is acutally Uwu, not Aave;
- Liquidated CRVs do not exist!
How? In June, Uwu platform where I had a position got hacked. Since then, these CRVs backing my position haven't been returned. Recently I've made a bet that veCRV wrappers will repeg faster than repayment of the stolen CRVs on Uwu will happen, so I borrowed maximum against nonexistent CRV and bought veCRV wrappers with a discount using those funds.
Of course, when was possible, operators of Uwu liquidated those nonexistent CRV. As a result, they don't have to repay me that liquidated part of stolen funds in future, but they also make sure that lenders don't have any bad debt.
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
28.10.202416:52
Just noticed this extremely solid research piece from Circle. Will read, but sounds like a breakthrough?
https://www.circle.com/blog/confidential-erc-20-framework-for-compliant-on-chain-privacy
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
https://www.circle.com/blog/confidential-erc-20-framework-for-compliant-on-chain-privacy
🥳 News Curve | ✈️ Telegram Chat | 🤟 Twitter/X
🛸 China Chat | 📱 YouTube | 📱 Discord
Көрсөтүлдү 1 - 14 ичинде 14
Көбүрөөк функцияларды ачуу үчүн кириңиз.