ServicesMarketsMarket makingTokenisationRWA infrastructureTreasury managementPrice feeds & oraclesIntegrityMarket surveillanceSecurity auditingPlatformCase studiesBlogVenturesTeamAboutContact

Case study · Market making · November 2026 · Vandergrid

Holding a tokenised gold peg on two exchanges

Our books held a tokenised gold product within a median 7.4 bps of its on-chain reference price on one exchange and 12.4 bps on the other, with both sides quoted in 100% of snapshots over the measured month. Since launch, the exchange price has shown no systematic premium or discount to gold: the signed median deviation is +2 bps over 261 days.

7.4 bps
median from reference, Exchange A
12.4 bps
Exchange B
100%
two-sided quoting
+2 bps
signed median vs gold, 261 days
01

How the peg works

The books follow an on-chain reference, and the reference follows gold.

Our peg engine reads the pool every 30 seconds and reprices both exchange books whenever they drift more than 0.2% from it. The leg we control is three to five times tighter than the leg we do not.

Exhibit 1The leg we control is the tightest in the chain

The peg engine reads an on-chain pool every 30 seconds and reprices both exchange books when they drift more than 0.2% from it. Our books track that reference to a median 7.4 bps on Exchange A and 12.4 bps on Exchange B. The pool itself tracks gold to a median 40 bps, and the exchange price tracks gold to 32 bps over 261 days. The leg we control is three to five times tighter than the leg we do not.On-chain poolread every 30sPeg enginereprices beyond 0.2%7.4 bps12.4 bpsExchange AExchange B32 bpsGoldpool 40 bpsthe reference, not gold itselfbooks to gold, 261 days
The peg engine reads an on-chain pool every 30 seconds and reprices both exchange books when they drift more than 0.2% from it. Our books track that reference to a median 7.4 bps on Exchange A and 12.4 bps on Exchange B. The pool itself tracks gold to a median 40 bps, and the exchange price tracks gold to 32 bps over 261 days. The leg we control is three to five times tighter than the leg we do not.On-chain poolthe reference, read every 30sPeg enginereprices beyond 0.2% driftExchange AExchange B7.4 bps12.4 bpsbooks to pool, medianGoldpool to gold 40 bpsbooks to gold 32 bps over 261 days
How the peg is wired, with the median absolute deviation on each leg. Source: our own peg records and public tokenised gold closes.
02

Tracking the reference

Exchange A sat within 25 bps of the reference 96.3% of the time, Exchange B 77.7%.

Both quoted both sides in every snapshot of the month. The difference between them is the book, not the peg: the wider market spread on Exchange B makes its mid a noisier estimate of value.

Exhibit 2One book sits inside 25 bps almost all the time

Share of snapshots within each band of the reference price: Exchange A within 25 bps 96.3%, within 50 99%, within 100 99.9%. Exchange B within 25 bps 77.7%, within 50 87%, within 100 94%. Both exchanges quoted both sides in 100% of snapshots.0%25%50%75%100%within 25bps96.3%A77.7%Bwithin 50bps99.0%A87.0%Bwithin 100bps99.9%A94.0%Bshare of snapshots inside the bandExchange AExchange B
Share of snapshots within each band of the reference price: Exchange A within 25 bps 96.3%, within 50 99%, within 100 99.9%. Exchange B within 25 bps 77.7%, within 50 87%, within 100 94%. Both exchanges quoted both sides in 100% of snapshots.0%25%50%75%100%within 25bps96.3%A77.7%Bwithin 50bps99.0%A87.0%Bwithin 100bps99.9%A94.0%Bshare of snapshots inside the bandExchange AExchange B
Share of 5-minute snapshots inside each band of the on-chain reference, over 31 days. Every snapshot, no sampling. Source: our own peg records.
03

Tracking gold

A signed median of +2 bps over 261 days: no systematic premium or discount.

Median absolute deviation is 32 bps. The monthly median tightened from 80 bps in the launch month to between 19 and 38 bps from the third month onward.

This comparison covers Exchange B only, because Exchange A publishes no public price history.

Exhibit 3The peg tightened after launch and has held

Monthly median deviation from gold, 2026-02 to 2026-10: 02 80, 03 51, 04 38, 05 25, 06 30, 07 19, 08 27, 09 35, 10 32 bps. It tightened from 80 bps in the launch month to between 19 and 38 from the third month onward. The worst single reading of the whole period, 915 bps, is in September, when the on-chain reference itself dislocated.030609080708Feb51532Mar38138Apr25188May30178Jun19147Jul27374Aug35915Sep32164Octworst, bpslaunch915 bps in Septhe on-chain reference itself dislocated; the books followed it
Monthly median deviation from gold, 2026-02 to 2026-10: 02 80, 03 51, 04 38, 05 25, 06 30, 07 19, 08 27, 09 35, 10 32 bps. It tightened from 80 bps in the launch month to between 19 and 38 from the third month onward. The worst single reading of the whole period, 915 bps, is in September, when the on-chain reference itself dislocated.030609080708Feb51532Mar38138Apr25188May30178Jun19147Jul27374Aug35915Sep32164Octworst, bpslaunch915 bps in Septhe on-chain reference dislocated;the books followed it
Monthly median and worst deviation from gold, on daily closes over 261 days. Source: public daily closes for the product and two tokenised gold assets.
04

Where it broke

In early September the reference itself dislocated, and the books followed it.

The exchange price moved as far as 915 bps from gold. That is the limit of a reference-following peg: it is only as good as its reference. Most of the remaining gap to gold sits between the pool and gold, a median of 40 bps, not between our books and the pool.

05

Big gold days

On gold moves of 4 to 5%, both books held inside 25 bps for over 99.5% of the day.

Deviation from gold is wider on big-move days than the period median, which comes from comparing daily closes struck at different times on different venues rather than from anything the books did.

Exhibit 4Big gold days did not break the peg

The 9 largest daily gold moves with a measured deviation, plotted against how far the exchange price sat from gold that day. The two days with an intraday peg record, on gold moves of about 4 and 5 per cent, held inside 25 bps of the reference for over 99.5% of the day. Deviation from gold is wider on big-move days than the period median, which comes from comparing daily closes struck at different times.0501001502002500%2%4%6%99.5% in band99.6% in bandBigger gold moves did not push the books off the reference.absolute daily gold move|bps from gold|Intraday peg recordDaily figure only
The 9 largest daily gold moves with a measured deviation, plotted against how far the exchange price sat from gold that day. The two days with an intraday peg record, on gold moves of about 4 and 5 per cent, held inside 25 bps of the reference for over 99.5% of the day. Deviation from gold is wider on big-move days than the period median, which comes from comparing daily closes struck at different times.0501001502002500%2%4%6%99.5% in band99.6% in bandabsolute daily gold move|bps from gold|Intraday peg recordDaily figure only
The largest daily gold moves with a measured deviation, against how far the exchange price sat from gold that day. Source: our own peg records and public daily closes.
06

How this was measured

Every snapshot in each window, with no sampling.

The gold reference is public daily close data for tokenised gold, which carries its own small premium to spot gold. Intraday peg records begin in late July 2026, so earlier large gold moves have daily figures only.

07

Reference data

Over 31 days (September to October 2026), every 5-minute snapshot, each joined to the nearest on-chain read.

Exchange AAExchange BB
Median deviation from reference7.4 bps12.4 bps
Within 25 bps96.3% of the time77.7%
Within 100 bps99.9%94.0%
Both sides quoted100%100%

Exchange B’s daily close against the median daily close of two major tokenised gold assets, over 261 days: median absolute deviation 32 bps, signed median +2 bps.

Vandergrid provides market making services and was paid for this mandate. Figures are from our own records and public market data. Not investment advice.

How we run a market making mandate