杰夫·严 - 高频加密货币做市与Hyperliquid交易所（第六季第二集）

当然，对我来说这并不清楚，我认为这些法律都是具体的。

Certainly, it's not clear to me, and I think a lot of these laws are concrete.

也许这就是为什么会发生这种情况的原因之一。

Maybe that's a bit of why this happened.

运营交易所本身就很难，所以他们还有其他事情要操心。

And it's just hard to run an exchange, so they've got other things to worry about.

举一些具体的例子，我认为幌骗是一个非常重要的问题。

For concrete examples, I think spoofing is a really big one.

至于幌骗，我其实不清楚它的技术定义。

And by spoofing, I don't really know the technical definition.

我认为美国证券和期货法律中有很多术语，但我只是泛泛而谈。

I think there are many terms in US securities and futures laws, but I just mean broadly.

当我提到幌骗时，指的是在订单簿和由此产生的价格图表中可以非常清楚地看到这种行为。

When I say spoofing, it's just you see it very clearly on the order books in the resulting price graphs.

就像人们下了大量明显不打算成交的订单一样。

It's like people place these massive orders that they clearly don't want to get executed in some sense.

比如，如果这些订单被执行了，他们会很不开心，或者是为了证明意图，但这非常明显。

Like, if they were executed, they would be unhappy or to prove intent, but it's very clear.

这些订单根本不是为了成交。

These orders are not to get filled.

它们只是为了制造出这一侧有需求的假象。

They're to give the impression that there is demand on that side of the book.

因此，如果某个算法将订单簿上的流动性作为价格走向的信号，那么就可能被误导而在他们希望的一侧下单。

And as a result, if there's some algo that is looking at the liquidity on the order book as a signal for where the price will go, then hopefully tricking these maybe to place orders on the side that they want.

然后，根据欺骗行为的成功与否，这个幌骗算法可以随后下出被主动成交的挂单，甚至主动击穿那些被误下的被动订单。

And then depending on what trickery is accomplished, then the spoofing algorithm can then either place making orders that get aggressed into or even aggress against passive orders that are placed mistakenly, I guess.

我的意思是，这是一个非常常见的例子。

As I mean, that's a super common example.

另一个例子是，我不确定这是否算真正的市场操纵，但确实存在一些拉高出货的圈子。

Another one is, like I don't if this is, like, really market manipulation, but there are certainly these pump and dump circles.

为了好玩，我偶尔加入过几个。

For fun, I sort of, like, joined a few.

我从未参与过。

Never participated.

我只是个旁观者，天啊，这些东西真够厉害的。

Just it was a lurker, and, man, these things are quite something.

我觉得最近这些行为已经被清理了很多，这挺好的。

And I think they've been cleaned up a lot recently, which is cool.

但以前，它们会制造出惊人的交易量。

But back in the day, they would generate crazy volume.

基本上就是一些内部人士宣布某个币种，然后所有散户都跟着行动。

It would just basically be, like, some insider announcing some coin, and then all this done retail.

我不知道他们是怎么找到这么多人的，但他们成功说服了很多人同时买入，然后内部人士趁机抛售。

I have no idea where they find these people, but they manage to convince a lot of people just, like, buy at once, and then the insiders, like, sell into that.

作为高频交易员，你可能会觉得这没什么，但实际上这非常棘手，因为这种扭曲效应太强了，你很容易被误导。

And as an HFT, like, you might think that's okay, but it's actually surprisingly tricky to navigate because there's such a strong perversion effect that you can kinda be tricked.

这是两个具体的例子。

Those are two concrete examples.

我想在处理这些问题时，这又回到了你之前关于基础设施、模型与策略的提问。

I guess in terms of, like, dealing with them, I think it goes back to the earlier question you had about infrastructure versus model versus strategy.

你到底在做什么？

Like, what do you work on?

我认为这属于另一类：各种突发的、你不得不处理的杂事。

And I view this as another category, miscellaneous random stuff that comes up that you just need to do.

也许是风险管理。

Risk management, perhaps.

风险管理。

Risk management.

是的。

Yes.

特殊场景。

Special scenarios.

我不知道。

I don't know.

如果你不做这件事，而其他所有事情都做得完美，那么根据你所处的市场环境和交易品种，这可能会决定你的平均盈亏成败。

It's really like if you don't do this and you do everything else perfectly, depending on the regime we're in and what you're trading, this could make or break your average p and l.

我想我们最初看到这种情况时，感觉非常可怕，因为我们在刚开始时运气不错。

I think when we first saw this, it was pretty scary because, I guess, we were lucky when we first started.

也许我们最初交易的那些标的资产本身就很难被操纵，或者人们还没来得及下手。

Maybe the symbols we were trading on initially just were pretty hard to manipulate or people hadn't gotten around to it yet.

总之，我们完全没预见到这个问题，无知地构建了系统，直到盈亏表现还不错。

Anyway, like, we just completely did not foresee this problem and naively built in ignorance of it and got to a point where things were going well at this P and L.

但一旦我们中了这些套路，情况就变得非常剧烈。

And then once we fell for these tricks, it was very dramatic.

你可能在一分钟内就损失掉一整天的盈亏。

You could lose a day's worth of P and L in a minute.

如果你不调整你的策略，它们就会做出愚蠢的事情。

If you don't tune your strategies, they will do dumb things.

有时候，自动化交易是最愚蠢的交易方式，因为它只是一个没有人为判断的简单状态机。

And sometimes automated trading is the dumbest trading because it's some simple state machine that has no human discretion.

它只会执行被编程好的行为。

It will just do what it's programmed to do.

我们的做法只是对此采取非常务实的态度。

Our approach was just to be pretty practical about it.

你可以坐下来分析，尝试建立模型来预测是否存在操纵行为。

You could sit back and analyze it, try to come up with models to predict whether there's manipulation going on.

但我们的一个主要优势，至少后来发现是，我们行动极其迅速，根本不在乎所谓的正确做法。

But in one of our big edges, at least turning out, was that we just moved super fast, and didn't didn't really care for, like, the proper way to do things.

这非常注重基于数据的实际操作。

It was very much, like, grounding the data.

所以对我们来说，就是这样：

So for us, it was like, okay.

这种情况正在发生。

This is happening.

但发生得并不频繁。

It's not happening that much.

当我们以某种特定模式亏损时，就直接关掉它。

Let's just shut it off when we lose money in a specific pattern.

这东西你可以在生产环境中用一个小时就写出来。

And this is something you can code up in, like, an hour without in production.

那时候，我就是那个80/20法则的践行者。

And, I mean, I was, the eighty twenty back then.

没错，你可能会错过一些机会，但它能腾出你的时间去扩展规模，专注于那些真正能带来10倍回报、影响你损益表的事情，而不必太在意这些。

And, yeah, like, you're missing out on some opportunities, but it frees up your time to start scaling out and working on the things that actually are, like, 10x multipliers in your P and L and not really worry about this.

也许只有5%的时间你会被关掉。

Oh, maybe this 5% of the time you're, like, shut off.

你可能会错过一些本可以赚到的钱。

You're losing something that you could have been making or something.

所以这里需要一点判断。

So there's bit of a judgment call there.

让我们专注于到底什么才是最该做的事情。

Let's concentrate off of, like, what's the best thing to be working on?

从那以后，我们有了更多时间来研究这些问题，因此我们现在有了更复杂的模型来预测这些市场状态，弄清楚发生了什么，而不是采取这些非常离散的行动，而是对策略进行连续调整。

Since then, we've had a lot more time to work on things, and so we do have more complicated models now of predicting these regimes and, like, figuring out what's going on and instead of doing these very discrete actions, rather, like, having a continuous adjustment to the strategies.

所以到目前为止，是的，我认为我们对这些操纵者的行为方式以及如何检测他们已经有了相当好的理解。

And so at this point, yeah, mean, I'd say we have a pretty good understanding of how these manipulators operate and detecting them.

但同样，我认为对于初学者来说，二八法则非常重要。

But, again, I think for people starting out, 20 principle is super important.

你是否发现这种市场操纵行为（如虚假挂单）在交易量更大、币种更丰富的交易所中更普遍？还是说即使在币安上交易比特币和以太坊时，你也会看到这种现象？

Do you find that that sort of market manipulation spoofing is more prevalent in sort of the fatter tale of exchanges with the fatter tale of coins, or is that something that you will still see even on Binance with Bitcoin and Ethereum?

比特币和以太坊在任何交易所上都很少出现这种情况，因为它们的流动性实在太强了。

It's pretty rare for Bitcoin and Ethereum on any exchange because there's just a

流动性要高得多。

lot more liquidity.

我认为这更多取决于资产本身，而不是交易所。

And I would say it's more about the asset and less about the exchange.

我在几乎所有交易所都见过这种现象。

So I've seen it on almost all exchanges.

人们在不同的交易所会做不同的事情。

People do different things on different exchanges.

你可以看出他们是不同的人，但他们都遵循相同的模式。

You can tell they're, like, different people, but they all follow the same patterns.

某种流动性适中的状态。

Some sweet spot of liquidity.

我想，如果某个代币的流动性极低，那可能就不值得做了。

I guess if there's, like, really no low on the token, then it may be not worth doing.

但一些有一定交易量的收益型资产，你可以借此误导算法。

But some sort of payout asset that has some volume going on, so you can kind of, like, trick the agos.

算法们通常预期有一定交易量和交易活动，但你可以误导它们，让它们做出错误的交易。

Like, the agos sort of expect some amount of volume and some amount of trading, but you can kinda trick them to make some bad trades.

我相信，我们看待市场的方式往往受到我们交易时间周期的影响。

I'm a believer that the way we see the market is often influenced by the horizon over which we trade it.

作为高频交易员，你对市场运作的理解，可能基于你对微观结构的直觉，而像我这样在更长周期操作的人，则更关注长期回报的基本面驱动因素。

You as a high frequency trader, I think, probably have a different perspective of the way markets work given your intuition around microstructure versus, say, someone like me who operates at a longer tail time horizon that might focus on more fundamental drivers of returns over the long run.

你发过一条推文，说市场的一种心理模型是粘性流体。

You had a tweet where you said one mental model for markets is a viscous fluid.

系统中的文档在价格发现过程中表现为阻尼振荡。

Docs to the system play out as damped oscillations in the price discovery process.

但这是一个非常有趣的想法。

But that was a really interesting idea.

我希望你能谈谈这一点，进一步解释一下你这句话的含义。

I was hoping you could talk about that a little bit, expand upon what you meant with that quote.

我也非常相信对事物的根本理解。

I'm a big believer as well in the fundamental understanding of things.

这有点像我接受的数学和物理教育背景。

It was kinda like the math and physics upbringing I had.

如果我不理解某件事，我就很难在这个黑箱上进行创新。

It's like if I don't understand it, then I find it hard to innovate on this sort of black box.

因此，我喜欢创造这些心理类比，用隐喻来解释事物是如何运作的。

And so I like to just come up with these mental analogies, sort of metaphors for how things work.

如果是粘性流体模型，那么正确的问题可能是：为什么高频交易还能赚钱？

If it is viscous fluid model, maybe the right question is, like, why does HFT even make money?

如果你问散户，他们通常认为这是一种掠夺性行为。

And if you ask retail often, they view it as this predatory thing.

他们是在抢先交易我们，或者猎杀我们的股票，诸如此类。

They're front running us or, like, I don't know, hunting our stocks or whatever.

但其实不是。

But no.

我不是说高频交易在做上帝的工作，但我认为它为这些市场提供了必要的服务。

I'm not saying that HFT is doing god's work or anything, but I think that it's providing a needed service to these markets.

因此，就这些系统冲击而言，我们的模型认为，撇开市场结构和抽象因素，价格变动是由一些对我们而言本质上随机的外部因素驱动的。

And so in terms of these, like, shocks to the system, our model is, like, outside of market structure and abstract, price moves as these, like, external factors that are essentially random for our purposes.

也许有人只是需要大量买入，并立即要求流动性。

Maybe somebody just needs to buy a lot and, like, demands that liquidity now.

也许有新闻事件改变了这个代币的实际公允价值，因此一些人会据此交易。

Maybe there's a news event moving the actual fair value of this token, and so some people are gonna, like, trade that.

但这种需求往往突如其来，等它出现时，订单簿已经变了。

But this demand kinda just comes out of nowhere and often, by then leave, it's the book.

这就像一个非常激烈的对抗场景，因此人们急于完成交易。

It's, like, a pretty PVP scenario, so there's a lot of urgency for people to execute.

而且这常常会形成一个循环。

And it's often, it can be a cycle.

比如，有些人可能基于动量交易，或者交易会触发其他交易，存在许多不稳定的平衡状态。

Like, some people might be trading off momentum or, like, trades can trigger other trades, and there's a lot of unstable equilibria.

会出现一个大的冲击，然后人们涌入，几乎开始讨论真正的公允价值是多少。

There'll be, like, a big shock, and then people come in and almost have this, like, discussion about what the actual fare is.

因此，第一个动作通常最大，然后往往他们会说，哦，超调了。

And so the first move will be the biggest, and then often, maybe they'll say, oh, like, overshot.

有人会进来，交易这种均值回归。

Someone will come in and, like, trade that mean reversion.

可能是中频交易者。

Maybe it's, like, a medium frequency trader.

也许是高频交易员，他们就是知道，五秒后，公允价格平均会回归。

Maybe it's a high frequency trader who, like, just knows, like, oh, five seconds from now, the fair price is on average in revert.

然后另一个人会说，不，不是这样的。

And then someone else will say, oh, no.

不。

No.

不。

No.

这是一件大得多的事，我们会一路狂涨，直到价格上涨20%左右。

This is, a much bigger deal, and we're gonna start to whopping until the price hits 20% increase or something.

埃隆把DOGE添加到推特上，这确实是件大事，你们都错了。

Elon adding DOGE to Twitter is, like, a real thing, and, like, you guys are wrong.

于是他们会去捕捉那些做均值回归的交易员。

And so, like, they'll, like, go pick off the, like, mean reversion traders.

因此，不同参与者之间几乎爆发了一场大规模的讨论或较量，但关键特征是，价格波动越来越小。

And so there's almost, like, this big discussion slash battle going on between the different actors, but the key characteristic is, like, the moves get smaller and smaller.

对吧？

Right?

人们实际上是在用钱投票。

People are kinda, like, voting with their money.

或多或少，人们都会进入他们想要的位置，然后就会出现一种按美元加权的平均操作，价格最终趋于合理水平。

And, like, more or less, people get into the positions that they want to get into, and then there's sort of this, like, dollar weighted averaging going on and the price settles at the fair.

这其实就是市场运作的方式。

Like, that's kinda just how markets work.

在这片混乱中，高频交易的使命就是低买高卖。

Within all this chaos, HFT is, the mandate is to, like, buy low and sell high.

如果你想想那条上下波动的曲折线条。

If you think about that, just like the squiggly line that's, like, moving up and down all over the place.

如果高频交易平均在曲线低点买入、高点卖出，那么高频交易对市场的整体影响就是平滑这条曲线，这对所有人都是有益的。

If HFT on average buys when the squiggly line is low and sells when the squiggly line is high, then the market impact of HFT on average is to smooth out this squiggly line, and that is good for everyone.

它让价格更快地回归公允价值，并确保价格在其变动过程中尽可能接近公允值。

It makes the price snap to the fair price much faster, and it sort of ensures that it's as close to fair as possible along its trajectory.

所以，HFT越好，市场上的流动性就越多，这种‘流体’的黏性也就越强。

And so it's sort of like the better the HFT is, that's just like the more liquidity there is on your market, the more, like, viscous this fluid is.

我不知道这个心理模型有多完整，但推文就是这个意思。

I don't know how full this mental model is, but that was what the tweet was about.

你问我是否认为HFT的盈亏具有正自相关性。

You asked me whether I thought HFT p and l was positively autocorrelated.

我大概能提出一些理由来解释为什么它会是这样。

I could probably come up with some arguments as to why it would be.

我觉得这可能取决于市场环境。

I could see it being regime dependent.

我特别觉得在左尾部会这样。

I could see it particularly on the left tail.

一旦你开始亏损，我就能理解为什么这个算法因为某种原因不再盈利了。

Once you start incurring losses, I could see it simply just being a case where that algo for whatever reason was no longer printing.

所以一旦你开始亏钱，就会持续亏损。

And so once you start to lose money, you would continue to lose money.

你做了一项有趣的研究，分析了你的盈亏并不与自身自相关，而是作为预测你所交易品种中频价格的模型的输入。

You performed an interesting study where you looked at your p and l not being autocorrelated to itself, but as an input to a predictive model on mid frequency prices of the things you were trading.

如果你问我，是否认为你的高频交易盈亏能以任何方式预测你所交易品种的价格，我会说，除非你们全是同一方向的吃单策略，否则可能不会。

And if you ask me whether I thought your HFT p and l would be predictive in any way of the prices of the things you were trading, I would say maybe not unless they were taker strategies all in the same direction.

我不太期待在做市策略中，这种情况会具有预测性。

I wouldn't expect particularly on like a maker strategy that to be predictive.

你发现实际上确实存在某种信号，即你们在高频交易端的盈亏本身，是中频价格走势的有效预测因子。

You found there actually was some signal there that actually your own p and l on the HFT side was a meaningful predictor of mid frequency price movements.

给我解释一下。

Explain that to me.

这是我们一个疯狂的想法。

This was one of our crazy ideas.

所以我想我之前提到过，几乎总是更好基于已经有效的东西进行工作。

So I think I mentioned earlier that it's almost always better to work off of something that already works.

这样你的成功率会高得多，而且你有一个可以以此为基础扩展的起点。

It's just a lot your hit rate's gonna be a lot higher and you have this base to scale off of.

但我们确实为一些一次性、疯狂的探索留出了空间，有时它们会带来回报。

But we definitely leave room for the one off, crazy explorations, and sometimes they pay off.

所以这算是我们较为成功的一个业余项目之类的。

So this was one of our more successful hobby projects or something.

我们在开展这项研究时并没有很强的先验预期。

We didn't have strong priors going into the study.

主要的动机是：嘿。

The motivation mainly was, hey.

我们拥有的资本超过了我们能部署到高频策略上的量。

We have more capital than we can deploy the high frequency strategies.

我们已经接入了大量的交易所。

We've onboarded ton of exchanges.

这些是恒定因素的扩展。

Those are constant factor scaling.

由于交易所的规模越来越小，回报呈递减趋势。

It's like diminishing returns because the exchanges get smaller and smaller.

所以也许我们可以进入这种中频交易领域。

And so maybe we can get in this whole, like, mid frequency.

这就像Sharp3、Sharp4策略的金鹅，容量是高频交易的10倍、100倍。

That's like the golden goose of sharp three, sharp four strategies that have 10x, 100x, the capacity of HFT.

听起来很棒。

Like, sounds great.

所以这是最初的动机，但我们普遍非常相信有效市场。

So that was the initial motivation, but we're generally pretty strong believers in efficient markets.

基本上，是的，我们在高频交易中拥有所有优势，但如果我们给你一些市场数据，比如收益率之类的，让你预测日收益率，我们甚至都不知道从哪里下手。

Basically, yeah, we have all this edge in HFT, but give us some market data, I don't know, like, returns, whatever, and ask us to predict daily returns, and, like, we don't even know where to start.

带着这种谦逊，这个疯狂的想法成为我们进入中频交易的一种切入点。

So with that humility, this crazy idea was a way to kinda get a foothold into medium frequency trading.

通常，如果你能找到一个别人没有的、有用的数据源，这本身就可以成为一个交易策略。

Often, like, if you can just get some data source that is useful that people don't have, that itself can be a trading strategy.

我们也不会去发射卫星，盯着停车场看什么的。

And we're not about to, like, send satellites to go look at parking lots or whatever.

比如，经典例子是，但我们有什么数据呢？

Like, the classic examples are, but what data do we have?

嗯，我们有我们的HFTP和L。

It's like, well, we have our HFTP and L.

而且，显然这些数据对我们来说是私有的。

And, like, obviously, that's private to us.

而且，显然这些数据不是随机的。

And, obviously, that's not random.

你只要看看图表就行了。

You just look at graphs.

这非常有趣。

Like, it's very interesting.

如果你仔细想想，它和什么相关呢？

And if you think about it, like, what is it correlated with?

回到关于有毒流和散户流的讨论，它与散户流高度相关。

Going back to the discussion about toxic versus retail flow, it's pretty correlated with the retail flow.

我想，你通常的先验假设是，如果你能将市场中的某些参与者分组，并弄清楚他们在做什么，那通常就是一个很好的信号。

I guess your priors in general is, like, if you can segment some actors in the, like, market and, like, figure out what they're doing, then that's a very good signal in general.

这种先验想法某种程度上能预测某些东西。

Like, prior sort of that thing is predictive of something.

方向则没那么明显。

The direction is, like, less obvious.

所以我们一开始只是说，好吧。

So we kinda went in with just saying, like, okay.

我们有这个东西。

We have this thing.

它和另一个东西相关。

It's correlated with this other thing.

让零售流动，嗯。

Let the retail flow and, like, yeah.

这可能和价格相关，那我们为什么不直接深入分析一下呢？

Like, that's probably correlated with the price, and why don't we just work through it and analyze it?

那就是我们的动机。

That was, like, the motivation.

所以我们做了这个分析。

So we did this analysis.

我们基本上回归了各种基于利润和损失的特征，比如利润和损失的差值，以及利润和损失对一系列中频价格波动的导数。

We basically regressed, like, various p and l based features, the delta m p and l, like, the derivative of the p and l against kind of, like, a wide range of mid frequency price movements.

我们也不太确定中频是如何运作的，所以我们广泛地尝试了各种可能性。

We're also, like, just not sure how the mid frequency work goes, we kinda cast a wide net.

好吧。

Like, okay.

也许它是针对五分钟回报的。

Like, maybe it's for five minute returns.

我们将其指数级扩展到数小时。

We kinda exponentially scale it out to, like, hours.

这就是整个研究。

That was the whole study.

我们恰好拥有这些数据，因为我们有一个仪表板，它可以报告我们所有策略的盈亏情况。

We happen to have this data because we have a dashboard, and it, like, reports all the PNLs of all of our strategies.

因此，我们也可以按交易所、策略和币种进行切片分析。

So we can also slice it on exchange, on strategy, on symbol.

所以我们做了所有这些工作。

So we did all these things.

数据非常嘈杂，我认为有一些方法可以处理这个问题。

It's really noisy, so I think there are techniques to deal with this.

显然，我们不会用某个币种的盈亏去预测该币种的中频价格走势。

Obviously, we wouldn't regress one coin's p and l and try to, like, predict that coin's mid frequency movements.

我觉得这噪音实在太大了。

I think it's just, like, way too much noise.

我们基本上对这个问题做了二八分析，做了一些分组和分箱，遵循我们的先验假设，避免过度拟合。

Basically just did an eighty twenty on this and, like, yeah, we did some bucketing, some binning, like, following our priors to not overfit too much.

总体而言，我们发现了一个相当有趣的现象，我认为这与我所交流过的每个人直觉相反：我们的高频交易盈亏，无论是做市还是吃单，实际上都与加密货币的收益呈负相关。

And by and large, found this pretty interesting effect, which I think is counterintuitive to everyone I've talked to about this, which is that our HFT p and l, whether it's maker or taker, it doesn't actually matter, is negatively correlated with returns in crypto.

这个效应非常强烈，当我们看到这个结果时，我们非常兴奋。

And this effect is, like, pretty strong, but if you zoom in on actually trying to capture so we were super excited when we saw this, by the way.

我们当时想，天啊，干脆直接转型吧。

We were like, holy shit, let's just pivot.

我们就直接转向做空来实现亏损。

We'll just money shift to get a loss.

我们就去交易中频策略。

We'll just trade mid frequency.

一切都会变得很好。

Things are gonna be great.

这是一个非常显著的效应。

It was a very strong effect.

我不记得确切的数字了，但可能是几十个基点，在一到两小时的时间范围内，而且容量非常大。

I don't remember the exact numbers, but tens of basis points, in maybe an hour, two hour horizon, but with very high capacity.

所以问题在于，如果你仔细看，这个信号只会在告诉你做空时才会触发。

So the problem is if you actually look, the signal only triggers to tell you to short.

没有反向效应。

There's not a reverse effect.

也许会有，但我们不会运行我们的策略。

Guess maybe there would be, but we don't run our strategies.

我们调整策略以避免亏钱。

We tune the strategies to not lose money.

所以你赚钱了，好的，做空。

So it's like you make money, alright, short.

当你做空时，比如做空期货。

And when you short, right, like, short, like, the futures.

但如果你仔细观察实际操作，就会发现有一个效应，就是资金费率。

But if you actually look into doing it so there's this one effect, which is, like, funding rates.

当这种情况发生时，很多精明的交易者都在做空，我不认为所有人都使用相同的信号，但总的来说，阿尔法与其他阿尔法高度相关。

When this happens, a lot of sophisticated people are shorting, and I don't think everyone is using the same signal, but this is just in general, alpha is super correlated with other alpha.

人们可能在关注完全不同的东西，但归根结底，阿尔法高度相关，我认为人们很聪明，他们做出了正确的交易。

People can be looking at totally different things, but at the end of the day, alpha is super correlated, and so I think people are smart, and they're making the right trades.

所以存在资金费率。

So there's a funding rate.

然后还有另一件事，表现绝对最好的那些标的看起来像是异常值，极端成功。

And then there's this other thing where the symbols that perform absolutely best sound like outliers, extreme success.

我们显然会在任何研究中关注这些情况，而这些正是非常难以做空的标的。

We obviously look at those as in any study, and these were the things that are very hard to short.

净效应仍然很有趣，因为我们在交易时会积累库存，而库存是可以有偏差的。

The net effect is still interesting because, like, we accumulate inventory when we're trading, and you can bias inventory.

你可以在你的策略之间进行某种内部化调整。

You can sort of, like, internalize between your strategies.

不同的公司对此有不同的看法，但显然有一些你可以做的事情。

Different firms think about this differently, but there's obviously something you can do.

即使你不这么做，你也可以在这种情况最为强烈时，让你的策略偏向于不持有库存，这样会产生积极影响。

Even if you didn't do that, you could just bias your strategies when this is, like, absolutely strongest to, like, bias to not holding inventory, and, like, this will have a positive effect.

但这并不是一个孤立存在、明确无疑的交易机会。

But it's not, like, a surefire, like, obvious trade you can make in isolation.

我觉得期货方面确实有些东西，但我觉得这还不够有说服力，不足以单独围绕它建立一个策略，这就是为什么我认为这最接近于阿尔法收益，这种东西大概就是能在推特上分享的内容。

I think there's something there with the futures, but I think it was not compelling enough to really look into it and make a stand alone strategy around that, which is why I think it's, like, the closest thing to alpha that this is the kind of stuff that's, like, shareable on Twitter, I guess.

但我认为，这取决于你的策略组合和你正在运行的系统，这实际上可能是一个非常可操作的阿尔法收益。

But I think depending on your set of strategies and what you're running, this could actually be, like, super actionable alpha.

我本来想说，我喜欢这个观点：它可能并不是一种可操作的阿尔法收益，因为如果你真的想做空期货，这种效应可能已经被计入期货的融资费率了。

I was gonna say I love this idea that it might not be an actionable alpha in the sense that if you actually want to short the futures, it might actually be priced into the funding rate of the futures.

但通过调整你的持仓头寸，是另一种实现这种阿尔法收益的方式，这种方式能对你的盈亏产生实质性影响。

But biasing your inventory is another way to actually implement that alpha in a way that can have a meaningful impact on your p and l.

这让我想到，以DFA为例，他们交易的频率下，并不专门交易动量，但当他们买入价值股时，会筛选掉那些动量极低的股票。

It reminds me and the frequency I trade, DFA for example, they don't trade momentum specifically, but when they go to buy value stocks, they're gonna screen out the ones with really low momentum.

他们并没有明确将动量作为因子纳入策略，而是在买入价值股之前，等待那些发生在完全不同时段的负动量消退。

That they're not explicitly incorporating momentum as a factor, but they're waiting for that negative momentum which occurs at a totally different time horizon to abate before they buy their value stocks.

虽然这是完全不同的因子组合，但核心理念相似：即采用一个理论上正交的阿尔法信号，不直接交易它，而是将其融入你的交易方式中，为你的策略带来一些边际优势和微小改进。

Totally different set of factors, but similar idea of taking a theoretically orthogonal alpha signal, not trading it explicitly, but incorporating it into the way you're trading to add some marginal edge, marginal improvement to what you're doing.

我喜欢这个理念。

I love that concept.

我本来还想补充一点。

I was gonna add on to that.

我觉得这是一个非常有趣的例子，我以前没听说过。

I think that's a really interesting example that I hadn't heard of.

但我听过一些疯狂的故事，比如一些手动交易者会进行大额交易。

But I've heard of some crazy stories, like some manual traders who swing large size.

所以我假设他们知道自己在做什么。

So I assume they know what they're doing.

我们会说，哦，是的。

We'll say things like, oh, yeah.

在加密货币中，当50日均线突破某个水平时，我就认为这是一个信号，但这并不是技术分析。

In crypto, when the fifty day moving average crosses whatever, like, that's what I it's like, I have a signal that's not technical analysis.

但当这种情况发生时，我就会触发交易，我还没仔细研究过这个特定情况，但它让我想起了很多类似的东西。

But when that happens, that's when I trigger, and I haven't looked at that particular, but it reminded me a lot of that.

就像在等待你认为相当不错的一些其他信号。

It's like waiting for some other thing that you think is pretty good.

一些条件信号来改变。

Some conditional signal to change.

是的。

Yeah.

很有趣的东西。

Fascinating stuff.

你提到过一件事，我们一直在谈论中心化交易所。

One of the things you've mentioned, we've been talking a lot about centralized exchanges.

我们还没有怎么讨论过链上策略或去中心化交易所。

We haven't really talked about on chain strategies or decentralized exchanges all that much.

你提到你最喜欢的一个已停用的链上策略是交易RFQ。

You mentioned one of your favorite discontinued on chain strategies was trading RFQs.

我希望你能解释一下这是什么，为什么你曾经如此喜爱并认为它如此有效，以及为什么你后来停止了它。

I was hoping you could explain what that was, why it was a strategy that you loved and that worked so well, and then why you discontinued it.

这大概是半年前的事了，我想，那时我们正处于DeFi扩张的中期。

This was about half a year ago, I think, when we were in the middle of expanding DeFi.

我们听说最好的机会正逐渐转向链上，而中心化交易所的交易似乎正面临收益递减。

We had heard a lot of the best opportunities were starting to move on chain, and centralized exchange trading was kind of hitting, kinda like, you know, diminishing returns.

成交量很低，所以我们想，好吧。

Volumes were pretty low, and so we were like, okay.

让我们花更多时间研究DeFi。

Let's just spend more time looking at DeFi.

我认为那时RFQs还是一种潮流。

And I think back then, RFQs were a bit of a fad.

我觉得Crocswap的Doug最近写了一些关于这个的有趣帖子。

I think Doug from Crocswap has written some interesting threads about this lately.

我基本同意Doug的观点。

I tend to agree with Doug.

这并不是一个良好的设计，我认为它试图将TradFi中有效的东西套用到DeFi上，但并没有很好地适应。

It it's not a good design, and I think it's trying to take something that works in TradFi but not really applying it well to DeFi.

所以，为了给听众提供背景，RFQs指的是报价请求，我认为。

So I guess for context for listeners, RFQs stands for request for quotation, I believe.

这个想法不错。

The idea is good.

就像是，我们试着过滤掉做市商非常讨厌的那些不良交易流。

It's like, well, let's try to filter out this toxic flow that makers hate so much.

让我们尝试让散户直接与做市商互动。

Let's try to have retail interact directly with makers.

于是散户进来就说：嘿，我是散户。

So the retailer will come in and say, hey, I'm retail.

给我报个价。

Give me a quote.

然后做市商会给出一个报价，通常在VVO内，或者至少针对散户想要的规模，可能比散户直接下单到订单簿更好。

And then the maker will give them a quote, usually inside the VVO or certainly, like, for the size the retail wants, maybe better than if the retail were to hit the book directly.

然后散户获得报价，在DeFi中，这就像一个签名负载，你将其广播到某个智能合约，合约验证后就在散户和做市商之间完成资金转移。

And then the retail gets the quote, with DeFi, it's like a signed payload that you broadcast to some smart contract, which then verifies it and then does the fund transfer between them, retail and the market maker.

这就像场外交易，只不过围绕它构建了一个协议，我想。

It's just like OTC, but this protocol built around it, I guess.

是的。

Yeah.

这听起来不错，而且在传统金融中经常发生。

This might sound good, and it happens a lot in TradFi.

我认为Jane Street就做了很多这类事情，我觉得这非常好。

I think Jane Street, like, does a lot of this kind of stuff, and I think it's really good.

如果你想参与这种其他零售流，通过提供更大的交易量并避免被HNT抢先交易，你实际上为零售客户提供了很好的服务。

Like, you wanna be on this other retail flow, you're providing retail a great service by, like, giving them bigger size and not getting front run by HNT.

理论上很好，但在DeFi中，这显然是个愚蠢的想法，因为你如何证明你是零售客户？

Good in theory, but in DeFi, it's just obviously a dumb idea because how do you prove that you're retail?

一切都是匿名的，而且你没有进行KYC验证。

Everything's anonymous, and you're not KYC ed.

因此，这是对此概念的验证。

And so it's proof of concept for this.

我们基本上只是写了一个简单的Python脚本，用来向这些做市商请求报价。

We basically just span up this simple Python script that just, like, asks for quotes from these market makers.

他们给我们的报价，差了五个基点，报价有效期是六十秒、九十秒之类的。

And they were quoting us, you know, like, five basis points away of quotes valid for, like, sixty seconds, ninety seconds or something.

所以大多数时候，对做市商来说，这个单子很容易成交，但他们报的规模是十万左右。

So most of the time, it's, like, pretty good for the market maker to get that filled, but they're quoting, like, 100 k in size or something like that.

我们就说，好吧。

And we're just like, okay.

听好了。

Look.

我们等着价格变动就行。

We'll just wait until the price moves.

价格显然会变动，因为加密货币波动性很强。

And the price obviously moves, like crypto is really volatile.

当价格变动时，我们就说，好吧。

And when it moved, we're like, okay.

那我们就直接广播这笔交易。

Well, we'll just, like, broadcast this transaction.

那你打算怎么办？

Like, what are you gonna do about it?

这看起来简直太精准了。

And this seems, like, super pie sharp.

你甚至可以做得更好。

You can do even better.

你根本不需要等到价格变动才触发。

You don't even have to, like, wait till price moves to trigger.

这基本上是个免费的期权，而且期权还有时间价值。

It's basically a free option, and the option has time value as well.

所以你只需要等到期权快到期时，再决定要不要交易它。

So you literally just wait until the option is about to expire, and then you just decide if you wanna trade it not.

这样一来，它就变得更加稳定有利了。

So that makes it even more consistently good.

所以，我们其实就是这么做的。

And so, I mean, we just did this.

然后我想，我们并不是第一个这么做的人，或者也许我们是，但做市商反应了，他们说：好吧。

And then I guess this is we were not the first ones to do this, or maybe we were, but the market makers react, and they say, okay.

我们不再给你报价了，因为你让我们亏钱。

We're gonna stop quoting you because you are making us lose money.

所以你明显不是散户，他们就开始给你报非常宽的价，或者干脆不给你报价。

And so you're clearly not retail, and they just start to, like, give you super wide quotes or just, like, not quoting you at all.

然后你就换账户。

Then you just switch accounts.

你只是新建一个钱包，再重新来一遍。

You you just, like, fund a new wallet and, like, do it again.

从根本上说，我觉得这个策略没什么问题。

And, Fundamentally, think there's nothing wrong with this strategy.

我唯一有点担心的是，我们运行这个策略时所创造的主要价值，是向世界证明了这个RFQ的微观结构很愚蠢。

Guess a little sort of a concern I have is the main value we're adding when running this strategy is that we're showing the world, we're proving that this RFQ's microstructure is dumb.

资本应该被重新配置到更合理的事情上。

There should be an intellectual reallocation of capital towards working on something that makes more sense.

也许我们确实已经实现了这一点。

And maybe we sort of accomplished that.

我认为现在RFQ的做市商拥有了最后查看权，而不是零售用户。

I think now RFQs are the makers have last look instead of the retail.

正如你所说，我们已经停止了这个策略，但我认为已经发生了一些演变。

Like you said, we stopped running this strategy, but I think there has been an evolution.

但我确实认为，一旦你这么做了，RFQ的所有优势就都消失了。

I do think that once you do that, though, there's the whole benefit of RFQs goes away.

你可以看到推特上的讨论，改进中心限价订单簿是个难题，而我认为默认的RFQ并没有做到这一点。

You can see the discussion on Twitter threads, it's a hard problem to improve upon central limit order book, and I just don't think RFQs do it in default.

我想这是一个很好的例子，说明我们在DeFi中尝试各种方法，意识到这个领域多么不成熟，以及协议们根本没有深思熟虑过这些问题。

And I guess this is a good example of just us trying things in DeFi, just realizing how immature this space is and, like, how the protocols just, like, have not thought things through.

这正好为我们最终决定：嘿，引出了一个很好的过渡。

And this is, like, a nice segue into, like, us basically deciding, like, hey.

也许我们才是真正最适合构建一个服务零售用户、创建去中心化价格平台（如Gobi）的人。

Maybe we're actually the best people to build something that is actually going to service retail and, like, create a platform like decentralized prices, Gobi.

让我们顺着这个话题继续，因为这正是我想接下来讨论的，谈谈你最新的项目。

Let's take that segue because that's where I wanted to go next and talk about your newest project.

你仍然在运行高频交易簿，但你已经将大量智力资源转向了Hyperliquid这个项目。

So you're continuing to run the high frequency book, but you've pivoted a lot of your intellectual horsepower towards this project Hyperliquid.

它是什么？

What is it?

你为什么要构建它？

Why are you building it?

没错。

That's right.

我们之所以构建它，是因为在DeFi上交易时，我们感到非常困惑。

We're basically building it because when trading on DeFi, we were perplexed.

即使在2022年的DeFi中，也有大量的零售交易流量。

There's a ton of retail flow even in the DeFi 2022.

有大量的零售交易流量，但他们却在使用这些极其糟糕的协议。

There's a ton of retail flow, they're using these absolutely horrendous protocols.

他们支付了大量gas费，因为L1太差了，而且他们使用的这些协议设计也很糟糕，比如RFQ。

They're paying a ton of gas because the L1s suck and they're using these protocols where the design also sucks, for example, RFQs.

让我们惊讶的是，人们真的愿意使用这些东西，你从数据中也能看出来。

It was amazing to us that people actually want to use this stuff, and you can kind of see it in the data.

需求是存在的。

The demand is there.

我最初探索这个方向的方式，记不清FTX在这个时间线中具体是什么时候发生的，但肯定是在FTX崩盘之前，只是没早太多。

And the way I actually started exploring this, I don't remember exactly when FTX happened in this timeline, but it was certainly before FTX collapsed, but not that much before.

当FTX崩盘时，我认为叙事显然急剧转向了：‘天哪。’

Then when FTX blew up, I think, the narrative obviously shifted dramatically towards, oh, shit.

整个对手方风险的问题浮现出来，比如‘不是你的密钥，就不是你的币’。

There's this whole counterparty risk thing, like, not your keys, not your coins.

以前只是个梗的东西，突然间成了每个人脑海中的首要关注点，这进一步坚定了我们的信念：我们必须打造这样的东西。

This kind of stuff that used to be a meme was all of a sudden top of people's mind, and that just solidified our conviction that this was, like, something we need to build.

那么，该建什么呢？

And so what to build?

我认为我们在这一点上确实挣扎了很久。

I think we actually struggled with that a decent amount.

我们想弄清楚人们真正想要什么，以及市场上哪些需求没有得到很好的满足。

We kinda wanted to figure out what people actually wanted and what was not being serviced well in the market.

所以现在有大量Uniswap的克隆、创新或集成，比如聚合器、不同的曲线、不同的公式，以及各种调整，以让AMM机制发挥作用。

So there are a ton of Uniswap clones, innovations, or integrations, sort of like aggregators, different curves, different formulas, different adjustments you can make to make the AMM thing work.

因此，我们并不真正相信AMM。

And so we're not strong believers in AMMs.

我觉得现在有大量的愚蠢流动性，是由于对无常损失、挖矿收益等概念的错误和误导性叙事而产生的。

Like, I think there's just, like, a lot of dumb liquidity that is being provided due to this false, misleading narrative of, like, impermanent loss and or, like, yield farming and sort of remnants of that.

而且我们本来就不太相信这些，即使市场真的在需求这些，也已经有太多人在服务这个领域了。

And so we're not really strong believers in that anyway, and even if that were the thing that the market was demanding, there are so many people trying to service that.

如果我们再建一个，能带来什么新增价值呢？

What are we going to add by building one?

我们转而看向中心化交易所，思考：人们到底想要什么？

We kind of look towards centralized exchanges and say, what do people want?

价格发现发生了吗？

Does price discovery happen?

流动性在哪里？

Where does liquidity?

所有流动性都在永续合约里。

It's all in perps.

永续合约实际上是一种巧妙的创新。

Perps are just, like, actually ingenious innovation.

我认为这在传统金融中只存在了一分钟，但被加密货币普及了。

I think it was actually a minute in TradFi, but popularized by crypto.

让我们看看谁在去中心化的方式下做这件事。

And let's see who's doing that in a decentralized way.

基本上，没人做。

Basically, no one.

我的意思是，EYDX的订单簿是中心化的，但它们是最接近的。

I mean, EYDX is order book is centralized, but they're the closest you can get.

他们已经取得了一些进展。

They have some traction.

我们基本上在想，为什么我们不自己来构建这个呢？

We basically thought, like, why don't we build this?

所以我认为对交易者来说，你的卖点是：你喜欢币安，你喜欢Bibbit，你喜欢那些中心化但你又不想完全信任的东西。

So I think the pitch for traders is you like Binance, you like Bibbit, you like something that's centralized that you'd rather not have to trust.

会有一个叫HyperLiquid的东西。

There will be this thing hyperliquid.

最近，HyperLiquid在Klips Alpha上推出了。

There is this thing hyperliquid recently launched in Klips Alpha.

它为你提供相同的体验。

It gives you the same experience.

也许流动性还不足，但从根本上说，除了流动性之外，其他方面都一样：点差小、确认即时、Gas费用极低。

Maybe liquidity is not there yet, but, like, fundamentally nothing barring just the same liquidity, tight spreads, instant confirmations, epsilon gas.

基本上，Gas仅用于防止DDoS攻击，而链本身每秒可以处理数万笔订单，且无需排放。

Basically, like gas to the extent of, like, preventing DDoS, but, like, the chain itself can handle tens of thousands of orders per second without emission.

一切都是透明的。

Everything's transparent.

一切都是链上的。

Everything's on chain.

每一件事都是一个交易。

Like, everything is a transaction.

这基本上就是我们的愿景。

That is basically the vision.

我们最初瞄准DeFi，因为要做出这样的教育性阐述很难，而且我觉得很多人都在尝试这么做。

And we're targeting DeFi to start because it's hard to do to, like, make that educational pitch, and I think a lot of people are trying to do it.

教育人们，嘿。

Educating people that, hey.

你其实可以不用托管方。

You can actually there's this new way to, like you don't need a custodian.

区块链和智能合约就可以成为你的托管方。

A blockchain, a smart contract can be your custodian.

所以，这其实很难推销，也不是我们的核心优势，但确实有人现在就想做这件事，而这正是我们的目标。

And so, like, that is, like, a hard thing to sell and, like, not really our edge in doing, but there are these people who wanna do it today, and, like, that's what we're targeting.

我们基本上是在告诉他们：嘿。

We're basically showing them, hey.

在所有的DeFi协议中，大多数都不够认真。

Like, out of all the DFI protocols, most of them are not serious.

大多数只是贷款类的东西，有点像创可贴式的解决方案。

Most of them are just loans are something that, like, sort of works like a Band Aid solution.

也许它基于本地价格。

Maybe it's based on a local price.

随便吧。

Whatever.

适合投机者，但不适合需要真实流动性的专业交易者。

Good for degen gamblers, but not good for serious traders who want real liquidity.

但HyperLiquid脱颖而出，因为它正是为此而构建的。

But hyperliquid stands out because it is built with that in mind.

为了实现这一点，我们在技术上做了大量创新，因此花了将近一个季度的时间专心开发。

We had to innovate a lot on the tech to make this happen, and so we were heads down building for a good part of a quarter.

我们真的希望借助一些智能合约来让它实现。

We really wanted to make it work through some smart contracts.

我觉得我们最初被DYDX的无信任链下撮合、无信任结算模式所吸引，

Like, I think we were kinda sold on the DYDX model of trustless off chain matching, but trustless settlement.

但经过进一步思考，发现这种模式其实存在很大缺陷。

But I think upon further thinking, it's just pretty flawed.

系统的去中心化程度取决于其最中心化的组件，而不是最去中心化的部分。

The system's only as decentralized as its weakest component as its most centralized component.

因此，我们最终决定这不可接受。

And so we basically decided this was not acceptable.

这根本无法让我们实现真正想要的愿景。

This will not actually let us scale to the vision we actually want.

于是我们回头说，好吧。

And so back saying, okay.

我们必须做到完全去中心化。

We need to be fully decentralized.

这让我们几乎没有选择余地。

That leaves us very little choice.

我们需要自己构建一条区块链。

Like, we need to build our own blockchain.

我就这么做了。

And I kinda just did it.

我们秉持一种务实、不轻易接受既有假设的态度。

We're very much of, like, a no nonsense, don't take things for granted attitude.

所以人们说构建L1很难，但我们只是说，好吧。

And so people say it's hard to build l o ones, but we kinda just said, okay.

让我们找一个共识协议吧。

Like, well, let's find some consensus protocol.

不太理想。

Like, not great.

老实说，没想到它能运行，但它已经经过实战检验了。

Honestly, surprised it works, but it's been battle tested.

所以我们拿它作为基础继续构建，它帮助我们走到了今天这一步。

And so we took it and just built on top of it, and it's gotten us to where we are today.

你能再多谈谈这一点吗？

Can you talk a little bit more about that?

因为我知道，决定自建一个L1是你在Hyperliquid与其他去中心化交易所之间的关键差异，也是你策略中的核心组成部分。

Because I know deciding to build your own l one is a key differentiator between what you're doing with hyperliquid and other DEXs and is a crucial component to your approach.

首先，我们假设一些听众根本不知道什么是L1。

First, let's make the assumption some folks listening have no idea what an l one is.

你来解释一下什么是L1。

You're gonna just explain what an l one is.

其次，再次强调，为什么这对你们来说是一个如此重要且关键的决定？

And then second, again, why was that such an important critical decision for you?

L1，我觉得，就是整个叙事。

L one, I think, was, like, this whole narrative.

很多大型投资，比如Solana、Avalanche等等。

A lot of the big investments, Solana, Avalanche, etcetera.

这整个就是L1的原则。

This is, the whole, like, l one principle.

但实际上，这非常简单。

But, really, it's quite simple.

L1就是一个区块链。

And l one is just a blockchain.

它通常与基于智能合约的方法形成对比：后者是在另一个L1（比如以太坊或Solana）上构建你的交易所，由该L1执行智能合约。

It's usually contrasted with smart contract based approaches where you take another l one, whether that's Ethereum, Solana, and build your exchange as a smart contract that the l one executes.

这就是它的本质。

That's what it is.

我认为它如此重要的原因在于，这里存在一种奇怪的激励机制：人们想在L1上构建项目，因为你能获得风投或资金支持，而L1拥有大量代币，能为你提供这种背书和公关效应，因此这算是一种更稳妥的选择。

The reason it's so important, I think there's this weird incentive thing going on where people want to build on an l one because you get the VC slash funds, and the L1s have a lot of tokens, and you get that kind of backing and the PR, and so it's sort of like a safer bet.

显然，L1们非常努力地吸引人们在它们之上构建项目，因为L1本身是通用型的智能合约平台。

Obviously, like the L1s are really trying to get people to build on them because an L1 has a general purpose smart contract based.

L1如果没有人在其上构建应用，就没有价值。

L one has no value unless people are building on it.

因此，人们倾向于默认使用智能合约。

So there's bias towards defaulting to smart contracts.

而如果你看看所有基于Tendermint构建的Cosmos链，实际上没有人有动力去推动它们。

Whereas, if you look at Cosmos chains, which are all built on Tendermint, no one's really incentivized to be pushing those.

实际上没有任何价值积累。

No value actually accrues.

至少目前，像亚当这样的开发者并没有获得任何价值。

At least now, there's no value that accrues to Adam, for example.

我认为他们正开始找到实现这一点的方法，但从根本上说，这是一个自主权系统。

I think they're, like, starting to come up with ways to do this, but it's fundamentally, like, a self sovereign system.

因此，在评估你听到的这两种模式时，请记住这一点。

And so just keep that in mind in terms of assessment between the two you hear.

就我个人而言，既然我尝试过两者，我认为很难想象在智能合约平台上构建一个优秀的交易所，尤其是对于衍生品，如果你想要运行订单簿——我之前提到过，这是一个不错的模式。

In my personal opinion, like, having tried both is that it's hard for me to imagine building a good exchange for a contract platform, certainly for derivatives, certainly if you want to run an order book, which I talked about earlier, is a good model.

因此，也许这个观点的一个验证是，DYDX——可能是领先的项目——在五年后转向了构建自己的区块链。

And so maybe, like, some validation for this idea is that DYDX, which is probably the front runner, is pivoting to building their own blockchain five years later.

我认为对他们来说，可能是某种法律压力所致。

I think for them, it was maybe some sort of legal pressures.

我只能在这里猜测。

I can only speculate there.

但他们目前运行的系统显然不是去中心化的，所有人都知道这一点。

But the current thing that they're running is obviously not decentralized, and everybody knows this.

我想他们会根据自己的准备情况逐步淘汰这个系统。

And I guess they will sunset that when they're ready.

但从我们的角度来看，L1 才是构建优秀交易所的途径。

But from our perspective, L1s are the way to build good exchange.

也许举个具体的例子，如果你在运行一个智能合约交易所，你会受到智能合约协议运作方式的限制。

Maybe as a concrete example, if you're running a smart contract exchange, you're sort of constrained by how the smart contract protocol works.

因此在以太坊上，交易必须由用户操作触发。

So on Ethereum, transactions must be triggered by a user action.

所以，如果你想在永续合约交易所中执行这些基本操作，比如每八小时分配资金。

So then if you wanna do these things, these very basic operations on a perpetual exchange, such as distribute funding every eight hours.

这是将价格推向现货价格的机制。

This is the mechanism by which the price is pushed towards the spot price.

如果你试图在一个L1上构建订单簿，设计这样的功能会非常困难。

That's like a super hard thing to design if you're trying to build a order book on an l one.

假设你有十万笔未平仓头寸。

Let's say you have, like, a 100,000 outstanding positions.

在以太坊上，你需要更新的存储槽数量，根本无法放进一个区块中。

The number of storage slot updates you need to make on Ethereum, like, that doesn't fit into a block.

因此，你必须设计一个协议，来决定由谁来执行这个操作。

And so, okay, you have to design a protocol around, like, who does this?

你需要某种机制，比如拍卖，来决定谁有权限触发资金费率，谁获得奖励。

You need some pocket, like, some auction to, like, figure out is that privileged people who are allowed to trigger funding, Who gets the credit?

必须给他们支付费用，因为他们要支付Gas。

There's gotta be some fee to them because they're paying gas.

这不是原子操作。

It's not atomic.

会出现一种奇怪的情况：你大约每八小时获得一次资金，但根据人数多少，你可能会晚三分钟。

There'll be this weird thing where you get funding out approximately every eight hours, but depending on how many people there are, you might be three minutes later.

关键在于你该如何围绕这一点制定交易策略。

It's just how are you gonna run a trading strategy around that.

这只是一个非常基础的操作。

It's just, like, a super basic operation.

所有的永续合约交易所都需要这个功能。

Like, all per exchanges need this.

但如果你运行在L1上，这就很简单了。

But if you're running on a one, it's trivial.

你只需将其直接内置到共识协议中。

Like, you just bake it into the consensus protocol itself.

你只需说，好吧。

You just say, like, alright.

当你生成新区块时，你会执行任意代码。

When you're producing new block, you're gonna execute arbitrary code.

所以，如果这个区块距离时间零点正好是八小时的倍数，我们就触发这个机制，然后执行相应操作。

So if this block is a new multiple of, like, eight hours since time zero, let's just trigger this thing and, like, do the thing.

而且这要简单得多。

And it's just, like, so much simpler.

所以我认为，运营一个交易所比编写一个简单的智能合约更接近于构建一个L1。

And so I I think, like, running an exchange is a lot closer to building an l one than it is to writing some simple smart contract.

我知道你在谈论永续合约。

I know you're talking about perps.

我想稍微回退一下，谈谈去中心化交易所目前运作的传统方式，即通过不同费用层级的流动性池。

I wanna go back a little bit to the traditional way in which decentralized exchanges currently operate, which is via liquidity pools at different fee tiers.

所以，想要提供流动性的用户可能会为ETH/BTC交易对存入以太坊和比特币，并在1个基点、5个基点、30个基点或100个基点（我认为这是最高档位）的费用层级上提供流动性。

So someone who wants to provide liquidity might put in, I don't know, Ethereum and Bitcoin for the Ethereum Bitcoin pair, and they might offer liquidity at a one basis point fee tier or a five basis point or a 30 or a 100, I think, is how high they go.

但这些都是非常具体的资金池层级。

But they are these very specific buckets.

例如，我无法在15个基点的层级上提供流动性。

I can't, for example, offer liquidity at a 15 basis point tier.

这与订单簿模型非常不同，而你通过高频交易的经历对这种模型非常熟悉。

This is a very, very different model than the order book model, one in which you are intimately familiar with through your high frequency trading days.

你为什么认为你为Hyperliquid采用的订单簿模型，比当前去中心化交易所所使用的这种基点层级结构更优越？

Why do you think the order book model that you're adopting for Hyperliquid is inherently better than this b tier structure that current decentralized exchanges operate on?

基点层级这个概念很有趣。

The b tier thing is interesting.

如果你观察一下AMM，它们正在缓慢地向订单簿模式演进。

If you look at the AMMs, they're slowly trying to progress towards being an order book.

很多去中心化金融（DeFi）都让人感到沮丧。

A lot of DeFi is little frustrating.

这就像在重新发明轮子。

It's like reinventing the wheel.

也许在这个过程中会有一些创新。

Maybe there'll be some innovations along the way.

但从根本上说，流动性池模型既巧妙又像骗局。

But fundamentally, the liquidity pool model, it's both ingenious and scam.

所以它是在必要性中诞生的。

So it was born out of necessity.

如果你回到2018年，或者Uniswap刚建成的时候，除了每笔交易进行几次简单的算术运算、一到两次存储更新外，根本不可能做别的事情。

If you're in 2018 or whenever Uniswap was built, it's not feasible to do anything other than a few simple arithmetic operations, one or two storage updates per transaction.

用户对支付的Gas费用有承受极限，因此它是在必要性中诞生的。

The user there's a tolerance on much gas they're willing to pay, and so it was born out of necessity.

这就像一种计算限制。

It's like computational constraint.

于是他们通过巧妙地诱使人们向池中提供流动性，勉强让这个系统运转起来。

And so they, like, kind of managed to get it to work by basically tricking people into providing liquidity to the pool.

我觉得无常损失简直是个绝佳的营销手段。

I think impermanent loss was like a super good marketing ploy.

我觉得这几乎到了不道德的地步。

Borderline, I feel like unethical.

我不知道。

I I don't know.

我觉得这些人都是聪明人。

Think these people are smart people.

我很难相信他们不知道自己在做什么。

I find it hard to believe that they didn't know what they were doing.

但我认为他们是在欺骗人们，说：嘿，把你的东西放这儿。

But I think tricking people to say, hey, you put your stuff here.

你不是在交易。

You're not trading.

你不是在提供流动性。

You're not posting liquidity.

你只是把钱存进这个收益项目里。

You're just depositing into this yield thing.

是的，你可能会有亏损，但别担心，这是暂时的。

And yeah, you might have some loss, but don't worry, it's impermanent.

这绝对是值得怀疑的。

It's definitely questionable.

而且我认为现在人们正在对此感到焦虑。

And I mean, I think people are working up with this now.

你只需将价格建模为随机游走。

You just model the price as a random walk.

过去关于这一点有很多争议。

There used to be a lot of controversy around this.

我真的不知道为什么。

I don't really know why.

作为一个交易员来说，这太明显了。

It's, like, super obvious as a trader.

你只需套利这些池子就能赚大钱。

You just arb these pools and make a ton of money.

现在这是一项非常竞争激烈的交易，但它确实是一笔好交易。

That's a super competitive trade now, but, it's a really good trade.

那么，是谁在提供这种流动性呢？

And, like, who's providing this liquidity?

这不像订单簿那样是专业的做市商。

It's not professional market makers like an order book.

而是一群散户，可能把资金放到了那里，然后根本就忘了自己放了钱。

It's a bunch of retail that maybe put their funds there and, like, literally, like, forgot to put it there.

这本质上就是一种长期负期望值的行为。

It's just, like, this negative EV over time.

你只是在不断亏损。

You're just suffering.

他们搞些收益耕作的东西来激励流动性，但也许收益耕作一结束，散户就忘了他们的资金还留在那里。

They throw on this, like, yield farming stuff to incentivize liquidity, and then maybe the yield farming thing, like, dries up within the retail, like, forgets that their liquidity is still there.

我不太确定，但这种模式不可持续。

Like, I don't really know, but it's not a sustainable model.

人们可能会说，交易量还挺高的。

And, like, people might say, oh, the volumes are pretty high.

也许它有效，但这是因为一种巧妙的营销策略。

Maybe it works, but it's because of this ingenious marketing scheme.

随着时间推移，我预计流动性会持续呈下降趋势。

Over time, I expect the liquidity just to be, like, trending downwards.

当真正达到均衡时，你会发现流动性如此之差，以至于交易费用只能勉强覆盖相对于零售资金流的逆向选择损失。

And when it actually hits the equilibrium, you're just gonna find the liquidity needs to be so bad that the fees pay for the adverse selection relative to the retail flow.

如果算一笔账，这种水平的流动性简直糟糕透顶。

And that level of liquidity, if you do the math, is awful.

这就是为什么这些基于池子的模式根本行不通的核心论点。

And that's the fundamental, like, argument for why these pool based things don't work.

它的演变形式就是GMX或所有GMX的克隆版，它们不再使用恒定曲线，而是采用预言机价格。

An evolution of that is GMX or all the GMX clones, where instead of having this constant curve, they use an Oracle price.

它们用了各种技巧、限制和手段，以确保在交易发生时，预言机价格能相对准确。

They have all these tricks and sort of limits and things like that to get the oracle price to be relatively accurate when trades come in.

但即便如此，你还是开始看到一些著名的案例：有人在中心化交易所操纵价格，然后利用被操纵的预言机价格在GMX上进行交易。

But even then, you start to see these pretty famous cases of people manipulating a price on centralized exchanges and then trading on GMX against the manipulated Oracle price.

我认为所有这些都只是临时性的补救措施。

And I view all this stuff as Band Aid solutions.

我认为这项技术最终并不适合在一层共识或这类研究领域中使用，在这些领域里，你根本不需要做出这些妥协。

I think the tech is finally out of place in, I guess, like, l one consensus or this kind of, like, general area of research where you can just, like, not make these sacrifices.

你可以两者兼得。

You can have your cake and eat it.

你可以保持去中心化并运行订单簿，据我所知，从实证角度来看，这是人们发现的唯一能促进真实价格发现和真实市场的办法。

You can be decentralized and run an order book, which is, from what I can tell, empirically speaking, the only way people have found to encourage real price discovery, real markets.

拥有自己的L1的一个潜在问题是，它要求用户从某种法币入口或其他链上桥接资金进出L1，这可能会对价格发现构成风险。

One of the potential problems with having your own l one is it requires people to bridge money on and off the l one from some sort of fiat on ramp or another chain, which I could see potentially being a risk to price discovery.

由于资金进出平台的速度受限于桥接环节，平台上的价格发现可能不够高效。

The price discovery on the platform might not be as efficient because money moving on and off the platform is inherently speed limited by this bridging component.

我想听听你对此的看法。

I'm curious as to your thoughts there.

你认为在自己的L1上运行这种模式存在潜在风险吗？还是觉得这根本不是问题？

Do you see that being a potential risk in operating this as your own l one, or do you think that that's a nonissue?

这确实是加密货币整体的问题，而不仅仅是DeFi领域，因为即使你在中心化交易所交易套利，你的存取款也发生在区块链上。

It is definitely an issue in crypto in general, not even just in DeFi because even if you're trading on centralized exchanges when you're doing ARBs, your withdrawals and deposits are on the blockchains.

因此，如果网络拥堵，你仍然会遇到这个问题。

And so if things are congested, then you still have this issue.

但我们目前专注于永续合约。

But we're focusing on perps to start.

正如我所说，这是因为永续合约在机会上符合二八法则。

Like I said, it's because it's the eighty twenty here in terms of opportunity.

几乎所有的交易量都集中在永续合约上。

Almost all the volume is in perps.

永续合约的一个好处是，你可以再次应用二八法则，比如直接用USDC作为保证金，就此打住。

And the nice thing about perps is you can start with another eighty twenty, which is, like, let's just margin them with USDC and call it a day.

再添加几种稳定币来实现多元化，也并不难。

And not that hard to add a couple more stables in there to diversify across.

虽然存在稳定币风险，但总体而言，人们普遍愿意接受这种模式。

Stablecoin risk, but by and large, people are, like, pretty willing to get on board with this model.

好吧。

It's like, alright.

我把我的USDC存入这个桥、链或合约之类的，然后这让我能够对一大类加密资产表达我的观点。

I deposit my USCC into this bridge or chain or contract or whatever, and then this lets me express my opinion on a large class of crypto assets.

这挺酷的。

That's pretty cool.

所以在高波动性和价格发现时期，只要你有抵押品，就可以表达你的观点。

So in times of high volatility and price discovery, you can basically just as long as you have the collateral, you can express your opinion.

从这个意义上说，就是套利。

ARB in that sense.

这并不是套利。

It's not an ARB.

现货永续合约套利是一种统计性操作。

The spot perp ARB is a statistical thing.

你只是试图在现货和永续合约之间，以有利的价差获取资金费率。

You're just trying to harvest the funding rate at a profitable spread between spot and purpose.