Forecasting Crypto Market Regimes

Just wanted to let you know about the Block Works Digital Asset Summit, which is back in New York March 24th through 26. We'll have top speakers from leading asset managers, financial institutions, DeFi protocols, crypto companies, and policy makers all under one roof. Think of Black Rock, Coinbase, Robin Hood, and more.

Link: [URL]

Nothing said on Zerox Research is a recommendation to buy or sell securities or tokens. This podcast is for informational purposes only and any views expressed by anyone on the show are solely our opinions, not financial advice. Pocachio Ryan and our guests may hold positions in the company's funds or projects discussed.

Welcome back to another episode of Zurex Research. Today, we're joined by a very honorable guest, Luke. Luke, how you doing?

Doing fantastic. Thanks for having me, Baka and Danny.

Glad to have Luke on. One of the gigabrains, maybe the gigabrain of the Blockworks research team. So, pleasure to have you on and dive into a big thought piece that you've just recently put out.

Yeah, absolutely. So this past week put out our research report on forecasting market regimes with the SUSD term structure on Pendle and term structures and yield curves are everywhere in Tradfi.

You can think of the VIX futures curve or a equity index volatility or the 2-year 10-year treasury rate curve or the sofur futures curve. And how they're useful is they basically reveal the market's expectations for the path of a certain security price or interest rate forward out in time.

And the market prices it based off of the expectations channel. The curves can shift upward and downward. But crypto and the onchain economy has largely lacked having a term structure for most of its history.

But what we found is over the past year given the extent of the utilization of SUSDE on Pendle where you have several maturities priced across various points out in time into the future that creates a term structure for the underlying yield that has been a pretty reliable gauge of changes in the underlying interest rate and also correlations with movements in major asset price levels like Bitcoin.

I guess I know you kind of gave general overview over here, but maybe for the listeners who need to catch up a little bit in I guess in Tradfi, you mentioned that there's the VIX or I guess the VIX curves. Are there any other examples and how are these used in Tradfi as an investor or as a trader?

Like what would somebody look at these for with regard to signals and then we can move on to the crypto side of it?

Absolutely. So we can start with the VIX futures curve which is basically price of the VIX future across basically monthly expirations into the future. And this is a snapshot from March 16th 2020 when the VIX futures curve was in steep backwardation.

So basically the front month was priced at 72 which is incredibly elevated. But then if you look forward out in time over the course of the next year the implied volatility given that VIX futures price it was expected that volatility would collapse from the current market conditions down to you know 30 forward out in time.

So when you see market crashes and once the trough has been set in equity prices, this is generally what the VIX futures curve will look like. It will look like steep backwardation and the market pricing declining volatility forward out in time and that's a pretty bullish signal.

The flip side of this is the VIX futures curve in contango where the curve is upward sloping. front month might be at 15 and back month might be at 20 to 22 and that's where the the futures curve spends most of its time in a slightly positive contango.

You can see this in commodity prices like oil or treasury rates and treasury futures two-year 10-year sofer futures curve and basically it reveals what the market expects that underlying instrument or yield to be forward out in time.

That makes sense. And for this form of a yield curve to exist, there needs to be expiring markets. Right.

Correct. So the SUSDE instrument itself basically just pays the underlying yield into perpetuity. But when you list that instrument on Pendle at one maturity, you basically have a tradable market to price the implied yield of the instrument and the implied yield will fluctuate up and down.

Various premiums and discounts will be exhibit but that will basically reflect what the market expects the underlying instrument to pay into that maturity date. But when you list multiple maturities forward out in time, you can reveal how the market expects the underlying yield to change between those points in time.

Okay. to kind of set the mood in terms of I guess crypto these forms of markets where there were expiring tradable pools of assets or whatnot like they haven't existed because we've been primarily a perpetuals driven market. Correct.

Like we don't have this form of futures. I mean I guess Trafi you can trade Bitcoin futures but we have mostly perpetuals. It kind of lends itself to the idea that we're still early.

We didn't have dated expirations or the ability to to price to price loans or to price interest rates into set durations until Pendle V2 came out and really found its strongest success with this SUSD instrument.

And that that market has settled multi-billions in liquidity and multi-billions in volume. So it's certainly the deepest and most liquid and largest term structure you'll eventually get it for loans and money market rates but in the current state of the onchain economy today this SUSD term structure is biggest most liquid and highest signal from what we've seen.

One last question before I guess we jump into the SUSD term structure is you mentioned Pendle V2. Is there something specific about the version two upgrade that led to this being like this phenomenon coming out or was it more so that it grew enough that now we are able to observe it at like a liquid level?

Yeah, so there are a number of updates between V1 and V2 that ultimately made V2 the better AMM implementation for pricing interest rates. One being the basically as you as time decayed into maturity, the pool would automatically rebalance liquidity to towards where it would be most likely to to price the yield.

So it's basically, you know, time variant liquidity which made the pricing structure much more efficient. But then also V2 came out and was in production well before Athena launched the USDE instrument.

I don't really think they knew it was going to happen but when it when it came to market we've historically seen 30 to 70% of the SUSD supply end up on Pendle for a number of reasons that that we can go into later.

Yeah, I guess maybe to follow up there and and lead into like why SUSD is so critical as part of this like maybe you can start digging into exactly that question of you know the Pendle platform existed. You could have other assets listed there with varying maturities like why exactly was it not until SUSD came along and the term structure there for that particular asset at different maturities so meaningful compared to you know all the other assets listed prior to that.

Yeah absolutely. So first and foremost SUSD is an instrument with very low variance in the principle is delta neutral and has very high variance in the APY. So if you're an LP, you're basically just providing SUSD. So you don't have much risk in terms of that appreciating or depreciating.

Then on the other side of that, there's the reasons for why you'd want to trade it on Pendle. One being is the fixed yield, the PT leg or the zero coupon bond. If you can lock in 10, 15, 30% fixed on a low variance principal position, that is an incredibly attractive financial instrument.

And that's been exhibited in that the scale to which there's been demand for it, but then also the demand to utilize it as collateral on money markets. Then secondly there's the YT leg where given the variance in the underlying yield which is quite high coming from funding rates as the most volatile leg down to the share of reserves that's parked in a USDC or USDT supply positions.

That's a significant share, but it's dramatically less volatile than down to the USD TB basically T bill rate. So given that variance, there's a lot of money that can be made basically being long or short the YT at various points in time.

But we would also find that this specific YT is perhaps the the most precise and liquid and scalable hedge to the cost of carry on long Bitcoin or ETH positions whether those positions are taken out in perpetual futures or levered long on a if you're levered long a perp or you're levered long those assets on a your cost of carry can rise in in ways that are indeterminate to you.

So if you wanted to hedge that that cost the best place to do so would be the SUSD YT.

Yeah, I think all of that I think all of that makes a lot of sense. Luke, can you maybe going from there like let's talk through some of the data? you know, you've done some digging here and like talk us through some of the examples like looking at the SUSD term structure different scenarios like what are some of the initial findings and like how that relates to you know market regime other things like that take us down the rabbit hole.

Absolutely. So I think we can start by looking at slides two and three here. We start with two. So this is what the term structure looks like in steep backwardation. There's a snapshot from December 15th of 2024.

So basically the the front month was pricing right around 30% fixed, but then if you went all the way out to the back month, it was priced right around 21%. there's about an 8 and a half% spread between the back month and the front month implied yields.

So what this snapshot of the term structure at this particular point in time suggested was that the market was expecting declining funding rates or underlying yield at that point in time. Then if we look at slide three, the inverse this is a snapshot of contango contango from April 8th, 2025.

Where at this point in time the front month was right at about 7 and a half. The back month was at 83 and we had very modest contanga with a positive term spread of about 60 basis points. And what that revealed was the expectation that the underlying yield would rise which might be a bullish signal.

So then looking at slide four basically at any any point in time we have multiple maturities on pendle you can get the front month implied rate and the back month implied rate. And it's really when these two implied rates produce a large spread where we can get pretty compelling signal whether it's backwardation or contango.

So for the sake of the analysis we created this rolling term spread metric which we can look at on the next slide which basically measures the difference in the back month implied rate less the front month implied rate.

So large negative values on the yaxis reflect steep backwardation and the market pricing declining underlying yield. Whereas on the flip side, positive values reflect contango and the expectation of of rising yields.

And with this like we hypothesized backwardation is a bearish signal and contango is a bullish signal to be you know consistent with the intuition of what this sort of term structure would suggest.

So we then sought to test you know to what extent has this given the the year and a half of history we have now on this term structure to what extent it actually is accurate in forecasting the forward returns on Bitcoin over the next 90 to 120 days and the changes in underlying yield over the next 120 days or so.

So if we look at yeah the next slide basically ran regression of the term spread observations against the forward realized return skew of Bitcoin over the coming 90 day 90 90 days and on the x-axis we bucket every term spread observation into you know 2% spreads.

And the yaxis shows the mean forward returns of the observations within that bucket. And here the the relationship is is quite clear like large negative values or steep backwardation on the term spread which is the the left side of the axis has coincided with the most negative return skew on Bitcoin over the coming 90 days.

In the middle of the curve where the term spread is normally distributed right around 2 to 3% there's not much signal returns are pretty flat we see observations both to the upside and to the downside but then as we move right on the x- axis towards contango and positive term spread values we see the highest forward realized return skew for those observations which is you know it's a first punch at confirming the hypothesis that backwardation is bearish and contango could be bullish.

Is there I guess like a trady comparable to this form of I like when you were I imagine that while you were doing these findings you were also researching on the side to see if there are like traffic equivalents. Is there anything that you're like well this has been the case in this market from this period to this period or is do you know what I'm trying to say?

Like is it purely because we're relatively like you said early that there's some form of like findings that you can have here or is it more so like just a general markets phenomenon and something that you see in multiple markets. Does that make sense?

So I'd say the behavior of this term structure is it fits it it fits neatly into analogies to both the VIX and the sofur forward curve because it has a combination of both.

The impi the implied rates on this market are driven by long biased hedging activity. when you see large premiums to the underlying yield a lot of that is hedging activity to long biased leverage.

In the same way large spikes in the VIX are driven by large demand for downside put protection in equity index fall. Then there's the underlying interest rate component which is quite analogous to something like the sofur forward curve.

That makes sense. And then I guess to make this a bit more because I feel like we're discussing something cool that you found. How how do you make this applicable as an investor?

So I wouldn't I wouldn't necessarily regard it as a binary trading signal like long short when certain metrics are hit but given the variance of crypto markets both in terms of the price action of the majors and also variance in interest rates both on a funding rates.

We see the strongest signal in the tales of the distribution. So looking at the past year and a half of price action and the term spread overlaid with it when we have these large runups in the Bitcoin price and you know trailing 90-day returns look something like you know 50 to 100% that drives up funding rates and you see, you know, the right tail of the distribution of funding rates and those observations consistently coincide with backwardation being a bearish signal.

Concurrently when rates are are very very low and they're very very cheap kind of like they are right now that's when we historically see contango which is a signal for more of a bullish market outlook.

So it's not yeah as I said earlier not a trading signal as much as a tool within a comprehensive framework to get a better forecast of how the market is expecting the forward distribution of interest rates to realize. So yeah if we look at slide eight here can go more in more in detail on that topic.

Yeah so term spread percentile. So what we did here instead of looking at the magnitude of the term spread we assigned a percentile ranking to every term spread observation and then bucketed those percentile rank rankings into deciles and then recorded the mean and median return skus of every observation within that decile.

And yeah, similar similar findings to the previous chart where you know the left tail of the term spread percentile steep backwardation exhibits the worst return profile on both a median and mean basis. Then inversely on the right tail positive values on term spread and contango we see the right tail of forward realized returns and the most positive return profile.

I think that makes sense and from what I understand this is more of a tool to understand how other people are positioning as opposed to making signals giving signals to long or short. Every market to an extent is a prediction market.

The term structure reveals what the market has implied as the forward distribution. There are points in time that the market pricing is incorrect. Perhaps people are overpaying for cost of carry protection.

Or perhaps people are too optimistic on you know fixed rate return at 5% and you know that's too too cheap. the yield should be higher. There are points in time the curve can be incorrect. At those points in time, there's a strong incentive to trade against the curve and correct the mispricing.

Maybe more directly on this point around the tails of this distribution like I'm curious maybe to what extent or if you go deeper into the analysis like is there something to do with just like the mean reverting nature of a lot of markets that is at play here.

I think to your point earlier, you know, about contango versus backwardation and bullish versus bearish like as something of an indicator, but not quite, you know, not quite causation or anything along those lines, but more about, you know, sort of an after-the-act analysis.

You can kind of see like, oh, the market was sort of generally positioning itself based on the term structure towards this. But of course, like if we talk about the VIX, like the VIX can always go higher. Like you could think maybe it's the top but then it could go higher and and then at that point maybe there there is a trade there at that point.

But the indicator itself doesn't tell you something. Is that you know just mean reversion and then people jumping on the jumping on board to like a new trend that is forming as a result of a bottoming in sentiment and just you know the piling into certain trades or how you think about that?

Do you go more deeper here into the data to to play into that?

Yeah, perfect question Danny and that sets up the next three slides perfectly. So what we see at large is the behavior of the term structure is simply the back month is always offering a pointer for mean reversion in the future distribution of yields.

So if we look at slide eight which is yeah term spread verse underlying APY underlying APY is on X-axis term spread is on the Y ais. So what you see is steep backwardation large negative term spread values that coincides with the right tail of the underlying yield distribution.

Inversely, contango positive term spread values coincides with the left tail of the underlying yield distribution. So there's a strong inverse relationship here just R squar of 6 on this scatter regression where you know the the most negative term spread values and backwardation those occur when the underlying yield is in its it highest regime and you're realizing the right tail of yields.

So we look at next slide slide nine. So this is the probability of contango by underlying APY cortile. So again we assign a percentile ranking to the underlying yield observations. Then figure out you know how frequently do we observe contango in that cortile.

You can see contango is over represented in low yield regimes on the left tail of the underlying yield distribution. And then inversely on the next slide. How frequently do we observe back steep backwardation?

So term spread less than arbitrarily neg 5%. Steep backwardation is over represented in the right tail of the underlying yield distribution when the underlying yield's above its 75th percentile. In every single observation where the underlying yield has been above its 90th percentile, we see steep backwardation in 100% of those observations.

So what this suggests is basically the front month pricing on the on the term structure that's going to fluctuate the most with the current market conditions but the back month pricing is always going to reflect longer term expectations or where the underlying yield will settle.

So in cheap interest rate regimes and the back month prices a higher yield in contango that's basically pointing to mean reversion and then the same thing is true in the opposite in high yield regimes where the back month is at a steep discount to the front month that's just pointing towards mean reversion

And I guess in the case of these markets like when we think about USD and like funding rates for Bitcoin ETH crypto majors There's a lot of that would be as a result of during periods of relatively low activity th those those rates come down as a result of you know low activity trading activity there the funding rates come down or stagnant and then naturally at some point would lead way to periods of higher activity and over kind of you know more degen speculation and pushing up funding rates and that would drive that amplified yield for USD and that would be reflected here.

We're just sort of seeing like the back and forth between those two states of the market. Absolutely. Spot on. So the more speculation there is, the higher the yield is, which means you want to be a little careful.

Exactly. So the the term structure basically gives you the metric for, you know, when you want to be more careful. And we we ran the regression and analysis on you know given that the term structure produces signals at the tails and it correlates with the underlying yield you know what's the better signal is the term structure useful at all if you already know what the underlying yield is and what we found was that term structure itself even when you incorporate underlying APY as a control variable able produces signal order order of magnitude higher.

Yeah. Yeah. So, this kind of goes into the the hedging conversation we had earlier where it's actually a a false assumption to think that if the Pendle V2 AMM was perfectly efficient in pricing the underlying yield that there would be no premium or discount or that the premium or discount would net out to zero to the underlying yield.

Given the behavior of the instrument basically the PT the zero coupon bond they are short the entire right tail of the underlying yield distribution almost like selling a call and they receive in fact a premium for foregoing that exposure. Inversely the YT has convex and nonlinear payoff curves with respect to the underlying yield and it offers basically volatility protection and optionality.

So they pay a slight premium for that exposure. So more often than not we observe large premiums to the underlying yield on the pendle market and large premiums are they reflect speculation both in the pendle market itself but then also hedging long biased speculation whether you're levering up on pers or on a so there's signal in both the level of the premium and discount in itself.

Like we had that example of steep backwardation in December of 2024. If we look at that observation on this chart here, the front month was up to over a 10% premium to the underlying yield whereas the back month was kind of having none of it and pricing a a huge discount to the underlying yield at that point in time it was basically saying if you look out to if you look out to April or May and that maturity this yield's going to come down 10% is yeah what that structure was telling you

I have a question about how's makeup plays into this so we understand that clearly there's a relationship with the funding rates How do you see? Well, I guess maybe like first to give kind of an overview, they used to use staking rates. They used to do staking. They don't do it anymore. So, right now it's made up of T bills funding rates and I think that's it. Some some USDC and USDT supply.

Yeah, some a stuff like simple basic lending stuff. How do you see each of this is you might not have direct answer but I assume you do but how do you see each of these like free segments play into this? I think funding rates is relatively clear here. I guess like the lending markets and tables how they play into this and another question is do you think specifically in terms of you're doing a lot of exploration of like how term rates and like yield works here.

I'm curious if you found any links to Tradfi. I guess specifically maybe how interest rates on A work or like lending on a works within that. Do you do you get what I'm trying to where I'm trying to go here?

Yeah. Yeah. So I'll try and I'll try and tackle that. So Athena's strategy composition has has changed and you know the the balance of what they allocate to will fluctuate but there's some large share that's USDTB in T bills incredibly low variance in what that yield is going to pay.

Then the middle of that is the money market supply rate which is it has much higher variance than the T-bills but significantly lower variance than what they earn from funding rates and the pendle market the YTS those instruments are they are long volatility in their payoff profile when the the expectation of the right tail fattens through volatility which would come through primarily the funding rate leg of the strategy not T bills or the money market the YTS appreciate and that drives a premium to the underlying and increase in the implied rates.

Okay. So the rates if it was just T bills and lending would not fluctuate the way they do and they mostly fluctuate due to the funding rates which is why this is signalish. Is there any thought to and I'm not as up to date on Athena's go forward plans like future timeline with the product but if these other you know if T bills if money market funds become a larger and larger share of sort of the backing of the instrument like does the value of this term structure around SUSD kind of disappear?

Does that melt away at some point?

Yeah, I think there's there's a really good case for that argument actually where this term structure isn't it's not going to be the last one or the benchmark one. Rather I think it's it's the first one of scale and signal that's been producing the signal we'd expect from it.

But looking out in time, there are a number of reasons why it might not be that benchmark in the long run. One of which is being this market has historically been heavily incentivized through SATs emissions from Athena and those will eventually run out and they won't be able to incentivize this market anymore.

We go back to the volatility characteristic of the YT. What we found is basically the Athena price correlates extensively with funding rates and the three-month basis. High funding rate regimes, Athena price rises, low funding rate, it falls.

So if you're in a 5% underlying and implied yield regime and maybe the market is pricing a 1% premium from SATs, should we move to a 20% funding rate regime because ENA correlates to that that 1% kicker would would actually realize you know 4% kicker and the probability of that basically needs to get discounted back to today's implied rates.

So what it does is it fattens the amplitude of the right tail of that distribution which the YT is long and drives a slight premium to the implied rate. That's one reason. Then secondly, we now have Boros as of August of this year, which it still is only at you know, one one10enth 1/100th of the scale of what this term structure was at its high watermark.

But in the long run, I'd expect it to be a more capital efficient venue for directly pricing the funding rate. So, Athena is a basket of a few different yield strategies. What's traded and listed on Boros is, you know, one specific interest rate payment on one specific market. It's much more precise in that way.

And also, the nature of a swap market is the longs versus shorts are kind of more matched in their capital outlay to express their positions. Whereas on the V2 yield market, there a lot more capital needs to get put up to buy one PT than what does to buy one YT, if that makes sense.

Then the third thing that Bako you had this question earlier about you know how does this tie back to money markets is the the spring and summer listing of USDE type PTS on a was quite significant and did materially change the underlying demand to buy PTS on this market.

We can look at some of the charts here. PT collateral on top money markets. Yeah, it's it's hard to overstate how significant the demand on a was for these instruments. Basically you could borrow against them on Oiler and and Morpho for a period of time, but but not at scale without impacting your borrow cost.

But a holding, you know, multi multi-billions in stable coins ready to lend against these instruments, you could slam a ton of size into this this PT loop without impacting your borrow cost much at all. If we look at next slide here basically that at that point in time the PTS exhibited the highest utilization rate out of any major collateral type on money markets.

So for every every $1 in PT deposited on a they generated about 89 cents in active loans which is really high leverage and very high margin. At this snapshot, if you looked at basically the total total serviceable market of PTS that money markets could underwrite against, you saw you know 80 to 90% of that total supply of PTS be utilized as collateral on money markets which is enormous utilization.

So what's that doing is driving demand for the bonds which as you buy the bonds that basically lowers the implied yield all else being equal on on the Pendle market. Then concurrently like when you were looping these on Oiler or Morpho at that point in time they were charging you like 10 to 12% to borrow against the PTS but on a it stayed 6 to 7%.

Meanwhile there was you know 15% of fixed yield to be had on the collateral. So significantly more size could be put into that strategy and going forward should there be a high high rate regime again and a supports listings for it that demand for the instrument to loop it could materially change basically the premium that the bonds pay to the underlying and in turn the term spread value.

So basically in the long run I would I would expect that this term structure is not the last one of significance rather I I would expect we see something similar with the exact same basically conclusions to instead be originated on the Boros market once it hits sufficient open interest and volume.

C can you talk to that then on in terms of where Boros is at today relative to the size of these PT markets what kind of growth is demanded of Boros for that you know if for that to sort of become the ruling regime then yeah so right now Boros it's up to about 250 million in open interest and it's doing about I think just short of $2 billion dollars monthly.

So if you look at the month overmonth or week-over-week growth rates in both those figures and annualize it, extrapolate that trajectory. I I think we'll hit it within a year or two would be my my my best guess. So we're really not that far off. Yeah.

And maybe just as a final followup on that is like can you talk through maybe the the value like I would assume either perhaps some capital that was previously using SUSD and the related instruments to perform some sort of trade here might now be switching over to Boros even at reduced size like advantages advantages. Great point there.

So say you're you're long a boatload of Bitcoin pers on Binance 10x leverage. Historically in a pre- boros world, if you wanted to hedge your funding rate, you would do so that the only publicly available open market hedge to that position was the SUSD YT on this term structure.

But there you're paying for the funding rate, you're paying for the A rate and the T bills in addition to you're buying some SATs along with it. So it's not the most efficient. Instead if you're long BTC per on Binance, you can hedge that precise funding rate directly on Boros and you're not paying for SATs or the A rate or T bills anymore.

You're just paying for that specific funding rate period and cash obligation. And if you move up that market, there's a strong incentive to correct it on the other side for for someone to to short it or long it back to its underlying rate.

So, boros just gets makes it much easier on the traders end to isolate what they actually want to hedge at. Correct. Yeah. Right now leverage leverage is capped at around 3x on those biggest markets.

Management guides to much higher leverage caps 20 40 50x on those sorts of markets. So your capital outlay to control you know one unit of Bitcoin per hedge should go down dramatically.

One angle that I think the SUSD pts will continue to shine out over Boros is the DeFi composability and the ability to, you know, loop those high yield zero coupon bonds on a at scale. That being said, I I see a lot of reasons why the premiums will go down in the future.

There is a very smart man that I know that anytime you give him a yield opportunity, he says, "I can just get this by looping SUSD." So, I imagine that that is very very valuable for people.

I I guess like not particularly relevant to this just something that popped into my mind with Boros. Would you I don't know if this is something that you found out or something that you thought about throughout the analysis. Would you classify as like longing the Bitcoin funding rate on Boros as effectively longing VIX or no like the VIX of Bitcoin or in that category?

Obviously not because VIX is like a very different product, but I guess longing activity on crypto. It's I would say you're getting long funding rate volatility and the right tail of the funding rate distribution if you were to go long on Boros.

But what also has materially changed crypto derivatives landscape is IBIT options now which in open interest they're in the I believe the DECA deca billions in scale so you can get better implied volatility there for Bitcoin the spot asset rather than the funding rate leg yeah I guess I was thinking of more of an onchain option But that makes sense.

So more so we just have almost like more precise offerings for particular use cases. It seems like my takeaway is that some of the SUSD pts and some of these other you know SUSD as an asset might transition more to you know you you mentioned looping, you mentioned just earning the the yield there on that asset. maybe more, you know, catered and targeting directly at users seeking the, you know, the nominal yield value versus for people taking large trades and wanting to hedge funding rates like they might try to completely distance themselves from that asset, get rid of all the other complexities behind it and Boros is the more optimal way to do so.

Yeah. If you were historically hedging your your Bitcoin per with the SUSD YT and you know Athena says we're we're taking down our per allocation we're going to allocate more