Original Author:CBB (@Cbb0fe)

Compiled by｜Odaily Planet Daily (@OdailyChina); Translator｜Azuma (@azuma_eth)

Background

• Odaily Note: Detailed story of this part can be found in HyperEVM "Top One" Arbitrage Team Public Strategy: How to Move Out $5 Million in Six Months.

Let's turn the clock back to October 2025.

In the eight months prior, we had been operating the top-ranking arbitrage bot on HyperEVM.

But this grueling task was about to end. We had been competing with Wintermute for the past few months, and now, new players had joined, and profits had been significantly compressed.

It was okay. My buddy and I were already used to this situation. We never tried to compete face-to-face with institutions and their massive armies of workers for the long term. We couldn't do it; there were just two of us.

Our advantage had always been — to deploy a strategy as quickly as possible and to extract all the profits before the big players came in.

They cannot deploy a strategy within 48 hours. They have regulatory restrictions, internal processes, approval procedures, and so on, while we don't have those; we just need to be as fast as possible.

So, it was time to look for the next grueling task.

We started to think, where is the next opportunity?

On October 10, that event just happened. The crypto market looked completely hopeless; everyone had suffered significantly, and there were no exciting new opportunities to excavate.

So, we began searching for a new battlefield. On October 13, HIP-3 launched on Hyperliquid. Three days later, TradeXYZ officially launched its first perpetual contract market for stocks — XYZ100.

Considering that Hyperliquid still had over 40% of its token supply yet to be distributed to the community, we thought it might be a good idea to generate some trading volume on HIP-3.

In fact, this line of thought was exactly the same as when we discovered the arbitrage opportunity on HyperEVM eight months ago: at that time, we simply wanted to generate some spot trading volume on Hyperliquid.

We didn't know if it would be useful, but we decided to give it a shot — to develop and operate a perpetual contract bot that arbitrages between HIP-3 and traditional financial markets.

Step One: Entering the Traditional Financial World

One thing needs to be clarified in advance: prior to this, we had never traded stocks in our lives.

We weren't even clear on what futures really were. Basically, we knew nothing about traditional finance.

All we knew was that for what we wanted to do, IBKR was an extremely competitive platform, so we decided to study it first.

In the first few days, I was mostly trying to figure out how to use the IBKR platform. I basically took screenshots of everything and sent them to Claude: "What is this?" "What does this mean?" "How should I operate here?" "How do we hedge XYZ100?"

This was basically how we started learning about TradFi. Meanwhile, my buddy began studying the IBKR API, trying to understand what could be done and what couldn't be done.

Coming from the cryptocurrency world, he was used to connecting to an exchange API and quickly getting the program running.

But IBKR was a completely different world. Market data subscription, contract specifications, order types, permissions, API restrictions, TWS, IB Gateway... We had a lot to figure out, and initially, we weren't even sure if this could be pulled off.

However, after a week of working with IBKR, we started to find some paths.

Building the Arbitrage Bot

This strategy was actually quite simple.

We treated IBKR's price data as "true price" and continuously checked for arbitrage opportunities on HIP-3.

If the trading price of a market on HIP-3 was at a discount compared to IBKR, we would go long on HIP-3 — only after our order on Hyperliquid was executed would we establish a corresponding short position on IBKR.

If the trading price on HIP-3 was at a premium compared to IBKR, we would do the opposite — going short on HIP-3, and after the order was executed, going long on IBKR.

In theory, it was very simple, but in practice, we needed to set a lot of parameters for each HIP-3 market.

Let's take the NVDA arbitrage strategy on the IBKR side as an example:

["NVDA", 55, 400, { maxDelta: 800, slippage: 0.1 }]

• “55 shares” is our minimum hedging size. Since IBKR's minimum commission is $1, we do not want to execute a lot of micro trades. Thus, we let the Delta position accumulate until it reaches 55 shares of NVDA before hedging on IBKR.
• “400 shares” is our maximum hedging scale per IBKR order to avoid excessive slippage.
• “maxDelta: 800” is our safety valve. If for some reason, our trades on IBKR keep failing, causing the position difference between the two markets to reach 800 shares of NVDA, the bot will stop trading that market.
• “slippage: 0.1” is the maximum slippage we allow when hedging on IBKR.

Next is the HIP-3 side:

NVDA: pair("NVDA", "xyz:NVDA", {makerSize: 400, makerOffsetBuy: 0.12, makerOffsetSell: 0.12, cancelDelta: 0.02, takerRatioBuy: 0.05, takerRatioSell: 0.1, takerMin: 1, takerMax: 2000, limit: 110000, makerEnabled: true, preMarketOffset: 0.04 })

It looks quite complex, but the logic is actually quite simple.

• “makerSize” defines our order size, while “makerOffsetBuy / makerOffsetSell” defines how much price difference we wish to achieve relative to the fair price.
• “cancelDelta” tells the bot how much price change would require it to cancel existing orders and place new ones.
• For taker trades, “takerRatioBuy / takerRatioSell” defines how large of a price difference we need before we actively take orders; “takerMin / takerMax” controls the trading scale we are willing to execute.
• “limit” is the maximum total position size we allow in that market, while “makerEnabled” just allows us to choose whether to enable or disable placing orders (maker).
• Finally, “preMarketOffset” adds some price difference during the pre-market trading period, as liquidity on the TradFi side is significantly lower during this time.

The First Trade

By the end of October, we finally felt ready to start trying.

The first few days were a bit chaotic. We were constantly struggling with the IBKR API, sometimes losing connection, and my buddy had to find a way to keep everything constantly connected and functioning normally.

But we quickly realized that there were tremendous opportunities. Basically, it felt like we were picking up money.

Throughout November, we executed about $850 million in trading volume on HIP-3, with profits over $500,000. That was good.

December was a bit calmer. We completed about $550 million in trading volume, and profits were still decent, but we actually began to think whether we should focus on other things.

It definitely made money, but it didn't seem like a gold mine.

However, we still decided to continue. As always, as long as there was still profit to be extracted, we often found it hard to stop.

Precious Metals Frenzy

The real explosion occurred in January this year.

Gold and silver began to rise crazily, and demand on Hyperliquid was absolutely insane. It began to feel almost too easy to make money, and our preparations over the previous two months were precisely for such a market.

One issue we faced was liquidity. Basically, everyone wanted to go long on commodities on Hyperliquid, which meant we needed to continually inject more funds into the IBKR side for hedging.

We kept adding funds to IBKR, but transferring such large amounts of money brought about banking issues. EtherFi was basically an MVP in this regard, allowing us to complete large fund withdrawals very quickly.

In the entire month of January, we completed $1.7 billion in trading volume on Hyperliquid, with over $600,000 in funding fee income alone.

But, as I mentioned before, there were just two of us. We didn't have internal processes, and we operated very quickly. Basically, everything was tested directly in the production environment, and sometimes this came at a cost.

On January 27, I had just arrived in Dubai and was preparing to grab coffee with my buddy and talk about the bot. Suddenly, I received a margin call alert from IBKR on my phone.

I completely didn't understand what was happening. It was still early, and the market hadn't seen any significant fluctuations.

I logged into IBKR. It turned out we had net shorted $120 million worth of gold futures while gold was experiencing a sharp rise.

We immediately shut down the bot. At that moment, I was shaking all over. I was genuinely terrified of being liquidated because essentially all the trades on IBKR were executed by the bot; I wasn't that familiar with IBKR myself.

In the next 15 to 30 minutes, I manually covered the $120 million short position in gold.

Later that afternoon, when the market opened, we finally calculated the loss — $1.1 million.

That was really painful. But we had no time to cry over it; we had to figure out what had happened and fix it as soon as possible.

Eventually, it turned out that the reason was actually ridiculously stupid; there was a problem with the data refresh from the IBKR API.

The bot believed there was a Delta mismatch between our positions on Hyperliquid and IBKR, so it kept shorting gold on IBKR in an attempt to correct a Delta exposure that effectively didn't exist.

Again and again; again and again; again and again...

Until it had cumulatively shorted $120 million worth of gold, we began receiving liquidation alerts.

Clearly, we needed more control measures.

We needed to ensure that the data we obtained from IBKR was indeed up-to-date. Before allowing the bot to continue expanding its position, we needed to add extra checks. More importantly, the bot was originally not designed for the massive trading volume and dense opportunities of today.

We spent an entire day fixing all the problems. The next day, the new version of the bot went back online, but our confidence had been shaken. Perhaps we really didn't know what we were doing. Maybe the risk-reward ratio wasn't worth it at all.

We had just lost $1.1 million due to an exceedingly stupid issue, and we now began to feel that the bot might have issues again at any moment. This was the first time since the bot went live that we seriously thought — should we just stop here?

But you should now understand us... We are just two greedy bastards. After losing seven figures, we wouldn't give up. We would just go harder.

And I feel this might indeed be something we are particularly good at. In the past, we had developed many bots together, and almost every time we went through a phase of insane losses, but for some reason, we always managed to climb back up in the end.

We wouldn't sit there for hours crying. We would try to figure out where things went wrong, fix it, and then keep moving forward.

Before recovering the losses, this would basically become a taboo topic between my brother and me.

The next day, after silver made a historic high, it underwent a mad pullback, and at one point, there was even about a 3% price difference between Hyperliquid and IBKR.

We made about $600,000 in profit from that.

We TMD fought our way back!

Liquidity Management and Speed Upgrade

By this stage, we had already made quite a bit of money. Moreover, we understood how this game worked.

If there's this much money to be made, then there will definitely be more people and their massive armies of workers coming to compete for this pie. So, we must become stronger as quickly as possible.

The first issue is funding. This is completely different from pure cryptocurrency arbitrage — in the crypto market, capital rebalancing between different trading venues can be completed in less than five minutes; but here, we have to go through banks to transfer funds in and out of IBKR.

Therefore, we designed a dynamic system that adjusts strategies based on the available liquidity on the IBKR side. When liquidity on IBKR is low, we are willing to pay a certain cost to close existing trades to free up funds.

Meanwhile, we would require a greater price difference to establish new positions. When liquidity on IBKR is high, we would do the opposite, willing to accept smaller price differences to open new positions and deploy funds more aggressively.

The second area that needed improvement was speed. Until now, we had been using IBKR's price data as our source of true price. This worked, but the data source was relatively slow.

As more and more participants entered this game, we knew that sooner or later this would turn into a speed competition. Continuing to rely on IBKR's data was no longer sufficient. So we started looking for alternatives and found Databento.

With Databento and a Nasdaq data authorization, we could access a much faster direct market data source.

We applied for access in January and finally got approved by the end of the month.

From Precious Metals to Oil

By February, the metal market was still hot, and we completed about $1.5 billion in trading volume. As if that wasn't enough, at the end of February, Trump decided to bomb Iran, causing crazy market fluctuations and pushing oil prices above $100.

At this point, we could make about $60,000 to $120,000 a day from arbitrage spreads and funding fees. Of course, excluding Saturdays and Sundays, as traditional financial markets would be closed, and we were bored to death.

If we "only" made $40,000 in the past 24 hours, we would even feel that something had to be wrong.

So, we would check the bot, adjust parameters, try to figure out what was happening, and how we could further optimize.

Claude gave us tremendous help in this process. We could feed all the trading data from Hyperliquid and IBKR to it, letting it analyze: where did we lose the most money? What exactly went wrong? What else could be improved?

In fact, this was our first time using AI to analyze trades, and it brought considerable changes. Even when everything was going smoothly, we still maintained this almost obsessive state.

Basically, this was the only method we knew to stay ahead.

My buddy and I discussed the bot all day. He would push code updates almost every day, while I continuously adjusted parameters based on what was happening in the market.

Semiconductor Frenzy

By the end of April, as the Iran conflict began to cool down, we thought that this crazy profitability was finally coming to an end.

In the past few months, we had been making about $500,000 a week, and we couldn't imagine anything that could continue to create such large-scale arbitrage opportunities. Then, semiconductors and all "bottleneck trades" suddenly began to explode.

Stocks like SNDK and MU started trading like pure Meme coins. It was absolutely insane.

We started building this bot in October when basically nothing was happening in the market. After that, we had experienced precious metals trading, oil trading, and now, semiconductors were being hyped up like scam coins on BSC.

What on earth was happening in this world?

Clearly, there were many elements of luck involved. We happened to be in the right place at the right time, and we just happened to have a product ready that was perfectly suited for this market environment.

But I believe that believing in HIP-3, stock perpetual contracts, and more specifically, believing in TradeXYZ, has indeed been a good judgment from early on.

In May, June, and July, we maintained trading volumes between $1.5 billion and $2.5 billion each month, consistently earning about $400,000 to $500,000 a week.

Conclusion

Now it's early September. Since we started, many institutions have entered this market.

Ethena also announced plans to enter the stock basis trading field in the coming weeks. For us, this opportunity may be coming to an end.

But it has definitely been a wild journey. In these 10 months, during a period that basically felt like a "crypto winter," we have achieved:

• $32 billion trading volume combined on HIP-3 and IBKR;
• 1.5% of TradeXYZ's total trading volume;
• $10 million in profits.

Of course, all of this was possible because we had a large amount of deployable liquidity, but in terms of actual deployed capital, the annualized return rate was still around 35% to 45%, depending on different stages.

More importantly, this was an excellent opportunity for us to truly enter the traditional financial world for the first time and understand how it actually works.

Ten months ago, we had never traded stocks, and we barely knew what futures were. Now, we have traded $32 billion worth of stocks and related contracts.

The only thing left to do now is to pray for Hyperliquid Season 3 to come soon.

Thank you for reading to the end. We will be back with the next story.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。