I wrote about the 3AC (Three Arrows Capital) wallets in October 2022. It’s May 2023 now and these are finally going to auction! The first lots open are live, with live bidding starting today, Friday May 19th, 2023 at 5pm Eastern / 2pm Pacific.
In February, CoinDesk and others reported that the 3AC NFTs would be liquidated, linking to the infamous Google Doc — still live — listing the assets.
More recently, the liquidator Teneo transferred these NFTs to Sotheby’s, who conveniently put them in a wallet named “sothebys-grails.eth” — although also with a few non-3AC “grails” thrown in. Notably some Gold Fur Apes and Mutants.
Michael Bouhanna, Sotheby’s head of Digital Art — announced this collection of 415 NFTs (mostly from the 3AC collection):
https://twitter.com/michaelbouhanna/status/1656931392964493313?s=20
He followed up that “Part 1” consisting of seven NFT lots, will be sold this Friday.
Following up on my 3AC piece from October:
we show our most recent valuations for the 3AC collection, including the pieces currently on auction
we dive into comps, market activity since October, and add some color to these valuations
we dig into the methodology for pricing illiquid JPEGs, particularly Art Blocks
we feature some of the most valuable remaining 3AC NFTs in the Sotheby’s Grails collection — for future auctions
High art values are often clouded in mystery. But I think much of this is pretty straightforward stuff. You can see historical sales, valuation histories and closest comps for these items on the DeepNFTValue website — or you soon will for some missing features. We aim to demystify these auctions somewhat, for those of you following at home and wondering how they get to some of these numbers.
In October, here is how we saw the values in the 3AC wallet:
collection floor num paid value USD
Fidenza 87.83 30 Ξ3,451 Ξ6,414 $8,716,943
Punks 66.00 11 Ξ1,230 Ξ2,177 $2,958,281
Ringers 53.25 17 Ξ3,204 Ξ2,126 $2,890,093
Autoglyphs 147.59 3 Ξ788.52 Ξ716 $972,958
Archetype 14.20 13 Ξ628.27 Ξ365.66 $496,888
Deconstructions 21.06 6 Ξ382.90 Ξ180.05 $244,666
Subscapes 9.17 7 Ξ210.57 Ξ158.98 $216,034
Squiggles 9.84 1 Ξ70.00 Ξ121.99 $165,769
----
Total -.-- 88 Ξ9,966 Ξ12,261 $16,661,635
Going into the auction in May, it looks a bit different.
collection floor num paid value USD
Fidenza 43.08 30 Ξ3,451 Ξ3,146 $5,693,863
Punks 50.00 11 Ξ1,230 Ξ1,432 $2,592,441
Ringers 23.47 17 Ξ3,204 Ξ937.5 $1,696,459
Autoglyphs 155.0 3 Ξ788.52 Ξ606.56 $1,097,583
Archetype 9.08 13 Ξ628.27 Ξ230.66 $417,389
Deconstructions 21.06 6 Ξ382.90 Ξ134.29 $243,006
Subscapes 4.4 7 Ξ210.57 Ξ76.28 $138,040
Squiggles 8.5 1 Ξ70.00 Ξ105.38 $190,687
----
Total -.-- 88 Ξ9,966 Ξ6,669 $12,069,468
The collection is down almost 50% in ETH terms, although a lot less in dollar terms.
Breaking that down into components
Collections floors are down, in ETH
ETH is up in USD
Punks and Fidenzas have held up reasonably well (and worth about the same floor — a long term trend)
Ringers have fallen a bit, and Ringers premiums have especially fallen
Autoglyphs and Squiggles are doing fine, but other ArtBlocks type collections continue to fall on low volume
The value in this wallet continues to be dominated by a few big collections
Overall, the 3AC wallet is still worth around $12M just for these 88 NFTs alone. The Sotheby’s Grails wallet has 400+ NFTs in it. Confusingly, not all are 3AC NFTs. They also added a few others, notably Gold Fur Apes and Mutants — as well as some non-3AC Squiggles. You can see the Sotheby’s Grails wallet on DeepNFTValue here. We’ll be adding historical valuations for the ArtBlocks collections soon — Punks and the (non 3AC Gold Apes) are up already.
If you’d like to see valuations for individual ArtBlocks items, I have those in a publicly visible Google Doc.
Meanwhile, let me do three things
explain the methodology of these valuations — how we arrive at the numbers above
highlight the top items in each collection, show comps, etc
give our view on the 7x items currently at Sotheby’s Auction — due to go into live bidding in a matter of hours…
Valuation methods
We price the Punks, Bored Apes, etc with our normal machine learning methods, described in previous posts. These look at trait values, combinations, individual NFT comps, listing and bids as upper and lower bounds… all kinds of things like that. Great for collections with trait-based structure and a healthy bid/ask market for at least some of the collection.
However trait-based pricing doesn’t really work for Art Blocks.
Sales are too rare
Traits that drive pricing are fuzzy and harder to define
How do you score The Goose Ringer — a one of one that 3AC paid 1,800 ETH for, with no grail traits like Gold Fur Ape or Punk Alien traits?
Instead we price ArtBlocks a bit differently. Considering a few factors:
For the entry price, what part was “floor” and what part was the premium paid at the time?
Since then, how have floors moved? And what about premiums — are those higher or lower for this collection?
This gets you to “pretty good pricing” — now we look at comps. What are the five most similar items in the collection, and how are we pricing those? That way you can fix the original pricing, smoothing out errors, and adding pricing for items that never had a realistic sales price or at least not recent ones.
The dark arts get into how to blend these factors, what weight to apply to comps and recent sales… as well as how to model “premium decay” over time. We’ll follow up with full pricing and more details in a future post.
But that’s why in the spreadsheet above, and in details below, we focus on entry price, and how floors and premiums for the collection have moved since the previous sale.
Seven Sotheby’s Grails
The grails that will sell today can be seen here.
These include
0xDEAFBEEF — for which we don’t have pricing
All price commentary below predicts the full price paid by the buyer. My understanding is that the current bids don’t factor in the Sotheby’s 27% fee.
Autoglyph #187 — we see this as a pretty normal glyph, nothing special (or bad either). We have the price estimate at 205.7 ETH, glphys floor at 155 ETH, and estimated liquidation price for glyphs at 125 ETH. I’ll be honest, I’m not a huge expert on glyphs.
The currently bid is $230K ~ 126 ETH — which looks on the nose to me. It’s what a random glyph would sell for as a lower bound. I think our 205 ETH estimate is high to be honest. Expecting this one to hammer 140-180ETH, probably somewhere in the middle, near the current 155 floor on OpenSea…
Now the Fidenzas. We have two.
Fidenza #725. This is a really nice “Small Scale” denza, with otherwise a normal color palette, shapes and density, nothing else remarkable about it. We have it at 184.4 ETH with effective Fidenza floor of 43ETH — a healthy 4x floor premium. This sounds about right.
The current high bid is just $130K ~ 71 ETH. I expect this one gets at least 100-125 ETH but might be hard to get a higher price given we’re talking 4x floor and there will be 30x other Fidenza sold by the Sotheby’s Grails wallet alone.
Best comps are other “small scale” and “micro” denzas. See recent sales on the DeepNFTValue site:
Small and micro Denza are not going below 100 ETH, and they are rare. But 200 ETH does sound like a lot to pay for a JPEG these days.
My guess is sophisticated bidders are going to wait, and try to sneak in a bid in low to mid 100s ETH. We probably don’t reach 184 ETH fair value… but we might.
The other Fidenza #861 we have at 126.6 ETH ~ $230K. Current high bid is also $130K and will go higher. For this one the special feature is “White on Cream”
Few of these have sold, but another decent comp is “White Mono” which do sell. Typically in the low 100s ETH range based on today’s prices.
This piece definitely feels more like a “3x floor” than a “4x floor” like the other Denza. I’d be surprised if bids don’t get to 100 ETH or close to it. But a bid over 150 ETH would also be a bit surprising, although it’s a cool and rare Denza for sure.
I’m less of an expert on Ringers. Most will say we are low on these items, and perhaps so but Ringers floor and premiums really have not gotten a strong bid in a while. Perhaps this auction changes that?
Bought for 55 ETH
Our valuation 49.4 ETH
Current high bid $48K ~ 26 ETH
Bought for 115 ETH
Our valuation 41.4 ETH
Current high bid $45K ~ 25 ETH
I think Sotheby’s expected these to go higher than floor, and much higher than current bids. They will probably be disappointed.
Both of these Ringers are clearly above-floor. #375 is dominated by the “25 peg” design. Here are the 25-pegs Ringers that have sold recently.
The other Ringer #194 is dominated by the Red Background — also rare. Here is how those have sold.
It’s hard to interpret these charts without looking at floor movements. But it seems clear that both categories trade for somewhere between 1.5x and 3x of the prevailing floor. That’s a wide range, and our model is on the lower end of that. A bigger question might be whether it’s fair to estimate that a “floor” Ringer trades for 23 ETH now.
In any case, I’m curious to see where the hammer drops, but ~2x floor for both of these premium Ringers sounds right to me. A motivated collector perhaps pays a bit more…
Our model predicts ~90ETH for the pair. I’d think 80-120 ETH (including fees) sounds about right. Maybe on the higher end of that, would be my prediction. I’d be surprised if the total goes above 120 ETH although it only takes one buyer… or two bidders realistically.
Our model sees Punk 1326 as worth 82.5 ETH ~ $149K. The current bid is $110K USD, and we expect this to go higher, probably around our price, maybe a little below. It is unlikely that we see a high bid, given this piece isn’t particularly special, and 10x other Punks are in the wallet. I’d expect a 72-85 ETH final sale, including fees.
It was bought for 48 ETH two years ago.
This is running long, so I’ll save analysis for other 3AC Grails in a piece after the auction. But there’s one more piece we must discuss.
The Goose
In 2020, “The Goose” sold to 3AC for 1,800 ETH, a record price, and the second highest price ever paid on-chain for an ArtBlock.
This may be the grail of the Sotheby’s Grails collection. This single piece was something like 18% of the entire ~10,000 ETH that 3AC spent on NFTs.
Word on the street is that Sotheby’s is “starting this auction” at $3M USD. Presumably this means they have a bidder at $3M or expect to get one.
This number strikes me as highly unrealistic. $3M is 1,654 ETH at today’s conversion rate. The item at its peak sold for 1,800 ETH. Since then both floors and premiums are down…
To justify at 1,600 ETH value, would imply one of the following
Premiums have not fallen as much as we estimate
This particular item has maintained its premium while other Ringers and other grails have fallen
The 1,800 ETH price paid then was a very good deal at the time
None of these seem particularly likely. Our valuation is low — below 250 ETH. Most will say we are too low on The Goose. When we estimated it at ~500 ETH in October I was told by experts that the real value it would sell for would be about 1,000 ETH. Now I’m hearing ~500 ETH when we are 250 ETH. We’ll see, I guess.
What other NFTs are worth 250 ETH these days?
An “above average” Hoodie Punk
“Blue Beams” Bored Ape — or half a Solid Gold Fur Ape
5-7 floor Fidenzas
10x floor Ringers
2x floor Autoglyphs or one premium one (with a bit of change left over)
The point being 10x floor is a big premium, and that’s the low end. A long term collector might pay 20x floor (500 ETH) but unlikely that a speculator will in this market. I suppose that if Sotheby’s gets multiple bidders this might go somewhere in the 400-800 ETH range — still much less than the $3M rumored to be their anchor point. And I don’t think there are many funds or big collectors looking to spend $1M+ on ArtBlocks these days — much less an single piece.
That said I don’t think Sotheby’s will will sell The Goose for 250 ETH. My guess is that they keep it on ice for months, waiting for the market to materialize. Maybe Ringers will bounce back a bit after this auction. Eventually perhaps they relent and sell it for 400-800 ETH… or as part of a package with other NFTs thrown in. Not hard to imagine a bundle that gets a bigger headline number.
Lastly, the 10x+ premium for The Goose… is based largely on it being the best Ringer. Which in turn is based on it selling for 1,800 ETH to 3AC. Maybe that still matters now, but will it matter in five years? I think the uncertainty there pushed down the premium for long time buyers. It isn’t as obvious a grail as Alien Punks or other premiums that have gone for 1,000+ ETH.
In any case, I look forward to these several Sotheby’s Grails being sold, in a matter of hours. Once the live bidding starts at 5pm Eastern / 2pm Pacific, I’ll be following along at home.
Good luck to any bidders if you’re reading this!
I’ll follow up with a post-mortem on how this went down and how my predictions worked out.
More than that, looking forward to the price discovery, and re-pricing our models with this new data!
80 more NFTs in the chart, and 400+ more NFTs total in the Sotheby’s Grails collection…
Let us know
Drop a comment to this article, if you think we’re getting pricing wrong, have a question, or a helpful link for Art Blocks valuation and modeling. We’d love to see it, as other readers may as well.
The jewel -- Fidenza #725 -- went for 3x our price prediction!
https://twitter.com/ivan_bezdomny/status/1659721798328500227?s=20
Every other piece went for above our price, although three were close, and the others 50% and 100% above.
Bullish prices for the auction. Don't know how Sotheby's does it!