GM(x) frens 🫐❤️ After messing around with @GMX_IO V2 for almost 2 weeks I want to breakdown why I think it is (once again) defining a new gold standard for the DEX perp sector. I start with a breakdown of the liquidity and fee model. Welcome to the beautiful new world of GM.
The success of GMX is strongly connected with the liquidity model GLP. At the recent top GLP had a TVL of $576m and undoubtedly started a wave of iterations, protocols building on top of it and utilisations of GLP for defi dapps (@muxprotocol, @Dolomite_io, @rage_trade,...)
GLPs success is strongly connected to its simplicity. It offers muted exposure to BTC/ETH volatility and providing "real yield". It allows for no-price-impact trading, which is interesting for large trading volumes, yielding high fees, paid back to liquidity providers and stakers
Despite its success GLP has a series of limitations:
🫐 Low capital efficiency (volume caps to reduce risk)
🫐 No incentives to balance OI (directional market exposure)
🫐 Limitations of tradable assets (full backing in GLP necessary)
🫐 Comparably high fees for small trades
With V2, GM is introduced as the new liquidity token. GM splits liquidity into isolated pools limiting risk to the respective pools.
Every pool consists of 3 token
🫐 Long backing
🫐 Short backing
🫐 Index
Both native token pools (e.g. ARB-USDC) and synthetic pools are available
Similar to GLP the price of GM is determined by the composition and the price of its backing assets, but I would like to point out some differences and common misunderstandings: (1) With GLP, fees are collected within one week and distributed in the following week.
This opens up to mercenary capital buying into GLP after a strong week, diluting the yield for GLP holders. GM on the other hand is auto-compounding and adds accrued fees directly to the respective GM pool, rewarding liquidity providers immediately.
(2) A common misunderstanding of both GLP and GM is that those LPs can experience impermanent loss (IL). This is neither the case for GLP or GM, since IL is a defined mechanism that only occurs for auto-balancing pools (e.g. UNI V3 pools).
GM pools are not auto-balancing, the asset ratio is independent from the price of its backing assets. Simply put: There is no IL in GM pools. BUT balancing of the pools IS crucial and a key difference to GLP. There are several mechanisms available for that in V2:
(1) Swap impact: When you buy into a lopsided pool with the overrepresented asset you pay a price impact fee, which is collected in a swap impact pool.
Example: There are $3.9m wBTC and $1.9m USDC in the BTC-USDC pool. The swap impact BTC pool contains $535.
Those funds are proportional distributed when swapping USDC against wBTC via this pool. If you would swap $100k USDC vs wBTC right now, you essentially get paid $91 from the swap price impact pool for doing so (wen integration of V2 pools @1inch, @paraswap,...)
The stronger the lop-side, the higher the positive price impact. Keep in mind that this is a dynamic setup. For example there are now $1.2k in the ARB-USDC swap impact pool, but the pool itself is almost perfectly balanced, the swap impact pool is currently not utilised.
(2) Price impact: Besides the pool balancing there is also balancing of open interest in V2. Balancing of OI is encouraged by price impact on positions. Every time a position on the overrepresented OI side is opened a price impact is deducted from the position.
This fee is collected into a position impact pool.When opening a position on the underrepresented side of OI there is proportional positive price impact on the position.
Example: There are $3.6k in the position impact pool for ARB-USDC. OI is lopsided $427k : $228k long : short.
Opening a $150k short position ($30k collateral, 5x leverage), gives a positive price impact of $149. After paying the opening fee the position is still $73in profits before any price movement.
(3) Funding fees: V2 introduces funding fees to additionally support OI balancing. Holding positions on the overrepresented OI side incentivises positions on the underrepresented side. APRs can be quite interesting for market makers to take advantage of.
Example: Check current funding fees for V2 pairs in @SniperMonke01 V2 bot. XRP funding for shorting was high indicating a strong OI misbalance in the first pic. This was incentivising people to open more short positions (second screenshot)
Negative Funding Fees are settled against the collateral automatically and will influence the liquidation price. Positive Funding Fees can be claimed under Claimable Funding after realizing any action on the position.
(4) Borrow fees: The borrow fee applies to all positions leverage trading and is redistributed to GM providers and $GMX stakers. It is currently set to zero if you are holding a position on the side with less OI, additionally favouring OI balancing.
Short breakdown of V2 liquidity and fees
🫐 Isolated GM pools
🫐 No IL
🫐 Higher capital efficiency
🫐 Strong incentives to balancing pools and OI (risk reduction)
🫐 Utilisation of pools for leverage trading but also for swaps
Why is this going to be a game changer now ?
Some thoughts:
🫐 Risk management: Isolated pools allow for superior highly scalable LP risk management
🫐 GM allows to build “custom made GLP”, which again can be automatised by third parties as it was done for GLP
🫐 Fee structure: The fee structure in V2 allows traders to open big positions with essentially positive price impact and zero fees under certain market conditions. This strategy is usually utilised by market makers on centralised exchanges and eventually transferred to Defi
🫐 GMX V2 allows unique combinations of providing liquidity and trade- and swap-arbing, enabling advanced trading strategies, that will eventually be utilised by third parties.
➡️ GMX V2 evolves the simplicity of GLP into a modular toolkit for liquidity providers, traders and third party protocols, putting a major focus on risk management
There are more features in-house for V2, but if you made it until here I think it is fair to call it a day. GN(x)
1/7 Today @chaos_labs published an article in the GMX gov forum presenting their @GMX_IO risk dashboard. This one really lifts trading analytics for GMX to a new standard. Let´s take a look.
2/7 The dashboard offers a variety of tabs beginning with an overview tab visualising high-level stats of the platform like pool value and composition, open interest and fees. You can toggle between stats for @arbitrum and @avax. Very smooth and professional UI in my opinion.
3/7 Next, the "Markets" tab shows you more detailed information about the utilisation of different markets. When clicking on a market it takes you to another page showing more details. Here it gets interesting as it also shows the OI imbalance for a selected market.
1/4 Let me see if I got the line of events correct: 1) We got Arbitrum, it was improved multiple times (NITRO) and has proven to be a reliable base layer for DEFI 2) The ARB foundations launches token as the biggest stimulus check to the community ever
2/4 3) After some chopping up and down $ARB token is nuking now and all tokens on arbitrum are nuking as well (bad tec) 4) We start trading meme coins on mainnet again and pay $50 for a single transactions pretty much like back in the days
3/4 5) The experience on ETH mainnet is too good, we start deploying stuff on BITCOIN and in addition to high transaction costs also have to wait ages for confirmation 6) "EVMs on bitcoin" are popping up, which are asking approvals from your EVM wallets (no, don´t approve those)
1/13 My recent tweet about DEX perps made me think more about how to reliably compare the user base/trading volume. What came out is the following:
-Number of unique addresses trading in 30 days
-Average trading size
-% of volume contributed from top traders
-Token incentives
3/13 With $10.5b @GMX_IO had the largest margin trading volume of those 4 in the last 30 days. The volume was contributed from circa 12k individual addresses, trading a position size of $63k on average The top 5 traders contributed 43% to the total monthly trading volume.
1/9 Perp DEXes are popping up everywhere, let´s take a breakdown of some high level stats of the Perp DEX sector. I gathered most of the data from the protocol stats pages / Dune dashboards. Big shoutout to @JamesCliffyz for his fantastic overview dune.com/shogun/perpetu…
2/9 The total volume of selected protocols more than doubled on average from Nov/Dec 22 to Feb/March 23. While @GMX_IO was dominating the market in 22, volume got fragmented in 2023. Anyhow, volume on GMX doubled on average as well, showing heavy growth of the overall sector.
3/9 After their launch on @arbitrum on new years eve, @GainsNetwork_io clearly won some market share in early 2023 (on their strongest day 57% of the selected protocols). It took them only 2 weeks after launching on ARB to replace MATIC as their preferred network for trading.
1/9 A lot of people have seen the open trades from @Rewkang on @GMX_IO and are asking: "What will happen to $GLP when he closes these trades". Let´s check.
2/9 Currently, @Rewkang has circa $16m of open profits in his trades including positions both on @arbitrum and on @avalancheavax. He withdrew most of his collateral already or the collateral was reduced by fees since most of the trades started at very high leverage close to 50x.
3/9 This results in some very high leverage numbers which are circulating, but keep in mind that these numbers are coming from reduced collateral and that PnL is not taken into account into the effective leverage.
2/10 The address they are talking about is this one: gmx.house/arbitrum/accou…
On GMX house you can check the trading history of the respective address, including PnL, position size and applied leverage
3/10 The address started trading on GMX October 6th 2022. Since that time it has traded more than $300m in volume, position sizes usually $10-30m mostly with high leverage (>20x). Both shorts and longs are traded. In total circa 20 closed trades.