Trading Strategy

Check the latest self-custodial open strategies here: tradingstrategy.ai The Open Strategies #trade automatically on #DEXes and #DeFi protocols. Trading capital is in on-chain vaults so you can withdraw your crypto whenever you want.

K Pool vault on Lighter keeps grinding up Vault description: Discretionary MTF low lev set ups on "tokens in play" through a strict selection criteria. Disciple of orderflow, price action, market structure and volume. tradingstrategy.ai/trading-view/v…

Long HYPE and BTC, short everything else. This simple strategy keeps winning on Hyperliquid tradingstrategy.ai/trading-view/v…

How do you get DeFi insurance for under 20bps? It all relates to how much insurance protocols should actually pay for per TVL. From shopping around, I can confidently say that the price in the market being offered is between 1.5-4% on lending and borrowing crime insurance. "but.. but.. you said the expected value is 3bps" yes but the insurers are: 1) skittish about cover, and 2) they have to take into account Expected Shortfall, and given no-one is buying cover they have pretty high concentrations on what they write. How the hell do you get to <20bps when you're paying 2% ? Start with defn of protocol crime insurance for hacks: - exclude bridge-related events - Solana and EVM - Protocol exploits or oracle manipulation involving a theft, a few others things too but it's PROTOCOL stuff Lets also give an explicit counter-example... START EXAMPLE SIDE QUEST >>> As a concrete example, most people in crypto talk about the "Aave hack". @aave didn't get hacked by any stretch of the imagination, you would definitely know if they did. @aave had a credit event. A collateral lost its peg because of an error in technical configuration. That error was propagated from a criminal cyber attack / hack on a bridge, who had also made an error in a recommendation to their client on their DVN setup. This is not a protocol hack of @aave and no crime insurance policy covering Aave would pay out for this. This is a credit default risk and protected by a Credit Default Swap. Where it gets confusing is that Aave's DAO appoints risk managers that are supposed to price the margins for collateral. However, you spin it though, the risk manager missed a credit event and the protocol remained resilient. What should ideally have happened is @aave have umbrella, collateral have Crime & E&O coverage and Bridge has Crime & E&O coverage. Insurers then get triggered by both collateral and bridge and then sue each other to decide who pays what %. There could be some argument that the risk manager might have some liability but it's going to be hard ground to justify because they have a framework, it's very public, they published their homework and we all deposited in a non-custodial fashion. << END EXAMPLE SIDE QUEST Getting back to the quant stuff. When you look at all the data of the 600 lending/borrowing protocols and 100+ hacks, especially on the larger protocols (of which there are nearly 20 hacks now) you notice something kind of interesting. The data suggests that the majority of protocols get pinged for a low proportion of their overall TVL. In fact the largest protocol hack was @eulerfinance and even that zero-day took about 63% There is a myth that is misunderstood, by even some brokers I discussed this with this morning, that hacks are total loss scenario. It might be for random off-the-book hedge funds but in the public and open-source Ethereum / Solana landscape we see all and it says the losses are varied but on average a small ratio. In fact you would cover 73% of all losses if you had just 5% of the TVL insured. Now hold up. If quotes in the market are around 1.5-4% and we can insure a large chunk with just 5% of cover then the passthrough to lenders is 1.5-4% / 20 per unit of TVL However, Increased Limits Factors kick in (honestly I don't even know if that's the correct acronym expansion, I just know it as "ILF") Insurers will generally insure blocks and create a "tower" which means that those at the bottom of the tower have higher risk. If the bottom of the tower is priced at 1% then the next guy might be happy to do next-loss cover at 90bps which implies an ILF of 90/100 = 90% On average ILF can vary but I think it's conservative to estimate around 95-99%. One of my broker buddies has also said there is a commercially minimum rate around 60bps so no matter how high the tower, you'll pay at least 60bps to get someone out of bed on each tower chunk. When you look at how this scales, on $1b of coverage you would buy actually just $50m of cover which would cover 73% of all losses. You would pay between 120.4bps and 325bps on that $50m You would pass through ONLY 5-19bps though to lenders. From a protocol perspective this is a total no brainer and insane that people don't do it more. The caveat? And why I care. "Unpermissioned DeFi is broadly uninsurable", a well known DeFi broker. Plug @KeyringNetwork... Lucky for you fuckers we charge no money on this and Keyring can onboard users with ZKPs in 3-30 seconds without selfies or passports. And most importantly, there is no backend review process, so you remain fully decentralised without triggering "points of control" for regulators. If you want to get insured, it's likely you'll want to look at putting a door lock on the think you insure to get a non ridiculous quote. Or get a quote at all.

@reisnertobias @GrowiHF Currently, $20B is earning better than the US treasury note rate in DeFi vaults 82% of vault TVL is earning better than US bank savings rate average tradingstrategy.ai/trading-view/v…

People still sleep on Vaults gg @GrowiHF

Building an AI trading agent is still way harder than it should be: Fragmented docs. Poorly structured APIs. We’ve been investing heavily in the integration stack to fix that. If you’re operating agents, GMX is officially your execution layer. ⬇️ 1/2

The @alphagrowth1 AUSD is the number one vault on Monad at the moment. This Euler vault has a 16% historical monthly return. Unlike more traditional Euler vaults, this vault does the leverage on the Balancer AMM. BPT is the liquidity provider (LP) token issued by Balancer, a decentralised automated market maker (AMM) on Monad and other blockchains. tradingstrategy.ai/trading-view/v…

A sign that the crypto industry is maturing: Onchain revenue is now much more distributed.

The vaults on @monad is the only chain where the yield has been going up for the last few months tradingstrategy.ai/trading-view/v… Both TVL and TVL-weighted average returns are increasing

Prop AMMs enable literally 0 latency order updates. Your contract can change its entire orderbook based on trades throughout the block. Thats very useful if you are the top venue for a token, and I think prop amms for long tail tokens with use this well.

Paul Mulvaney has done it again. Up 24.40% in April. Up 105.52% YTD 2026. #TrendFollowing #CTA #ManagedFutures iasg.com/groups/mulvane…

BredoStrategy is currently the #1 DeFi vault across 3000+ vaults. 66.9% net returns for the last 30 days. This is a discretionary vault. The vault leader, Bredo, is preparing for Hantavirus: "It's complicated, but it seems like a "logical decision" from all sides. I've cut my positions (closed them) by 30-50% each in Vault. There are two reasons. The first and main one is the hype around the Hantavirus. I don't want to argue about how dangerous it is or that there's no chance of a new pandemic, or whatever GPT told you, which OFC knows everything in advance))) What's more important here is that the messages are becoming increasingly critical and hype-driven, and the market doesn't need much to find a reason to sell, especially after the Corona dump. I'm not saying or thinking that there will be a new pandemic, I just think there's a chance that in the next 2-3-4 days we'll either see exponentially more messages about the Virus or not." This is his second vault. tradingstrategy.ai/trading-view/v…

Insertive Capital vault by @onchainquant on @Lighter_xyz is currently the #1 DeFi vault across 3000+ benchmarked vaults. The monthly net return is 68.9%. The vault strategy is a high-risk, high-return discretionary trading fund that actively pursues asymmetric opportunities across volatile markets. tradingstrategy.ai/trading-view/v…

Aave e-mode borrowers were running 30-50x loops on Eth LSTs. Gets you nearly 30% in ETH yield, but clearly comes with risks. People might think this is a purely DeFi phenomenon, but the same broad trade happens in TradFi with the Yen-USD carry trade.

A note on DeFi. DeFi will be FIXED INCOME. Below are views purely on the state of fixed income in DeFi. No payments/stables etc. The below are all best in class early stage fixed income plays that give exposure to a variety of market dynamics. Cheap liquid tokens - markets: @eulerfinance - super cheap beta @aave - kelp / dao dislocation @GearboxProtocol - exceptional tech Cheap equity - markets: @infiniFi - defi banking book @Zharta - best bond market (non fungible tokens) Cheap equity - strategy vaults (QIS): @ipor_io - best strategy vault infra (also leaders in swap space) @TradingProtocol - best quant /backtest infra @symbioticfi - best vaults for usd efficiency (reuse each $) Cheap equity - infrastructure: @KeyringNetwork - leader in full stack vault and vault infrastructure for fixed income Cheap equity - issuance: @AssetoFinance - largest in Asia @paretocredit - largest levered RWA TVL on morpho Cheap token - issuance: @plumenetwork - underpriced beta Cheap token - distribution: @turtledotxyz - category owners A note on agents: I haven’t put agents on here because it is my feel that Agentic AI for Fixed Income in DeFi will be cannibalised by OpenAI / Anthropic. I feel Agents in FI DeFi will largely be a UX and data analysis play that will be levered by quants and PMs for portfolio analysis and research tasks. It is my view that Agents will not do convex portfolio optimisation / allocation across an efficient frontier as they are not designed for this task. I am not sure whether quants / PMs will indeed pay for this from a startup. My own circle of quants and PMs in HFs would suggest they will (are) building in house. Attached article below about why we are where we are in DeFi and why DeFi is fixed income right now.

The LayerZero / KelpDAO hack caused a drop of $4B (12%) in stablecoin vault TVL $32B (all-time high) To $27.5B (today) tradingstrategy.ai/trading-view/v… The drop seems to be equal across all vaults and protocols, no clear winners and losers.

Time for some weekend bragging: This is my equity curve since I started my systematic trading using RealTest in live trading in Q4 2023. Before I was grinding with python for a year. It took off when I applied the right methods with the right tools (RealTest) and data (Norgate).

We didn’t expect this many people to show up for our brunch. Great conversations around vaults and DeFi infra. Huge thanks to our partners for making this happen: @ipor_io @TradingProtocol @1deltaDAO

I'll be on the @Vault__Summit NYC main stage presenting how institutions can allocate to vaults without triggering the regulator. Pinpointing the architectural choices that make institutional allocation actually work, based on real setups. Who'll be there?

@definikola Peckshild made a mistake attribution on this tweet and refuses to delete the tweet despite admitting it is incorrect, so even their tweet is AI slop x.com/TradingProtoco…

expect way more faster (obvious) exploits like this in the future, thanks to AI "free money" discovery setups

@safetyth1rd @peckshield It would be okay if they were sometimes incorrect. But @peckshield refused to remove the misleading tweet despite requests and their own admission that they were wrong.

I was also kind of surprised to see trading strategy “hacked” Turns out it wasn’t. @PeckShield ‘s alerts are overly sensitive Arguably better than under coverage but don’t always take them at face value