/ xRapid

Down the xRabbit Hole

Cale Robertson

An avid FPS gamer and game developer. ILP/XRP/Codius developer.

Read More

Sparked by a single post, I've decided to follow the white rabbit. In this case, the never-ending search for on-ledger xRapid transactions.

Not only will I be trying to find xRapid transactions; I'll also be doing my best to provide some analytics on the frequency/volume, potential owners of addresses, as well as the maximum tested size of transactions.

So where did I start?

I was scrolling through XRPChat (as I do most days) and found a very interesting comment in a thread about Mercury fx's use of xRapid (link):

The comment in question by user 'iLeeT' stated that this is the only confirmed xRapid transaction to date.

I asked for proof, and iLeeT provided the link to a post by none other than David Schwartz (Joelkatz) himself.
It explains that the above transaction was performed on the night of Consensus 2018 when they were doing their live xRapid demo, and can be found here.

The minute I finished reading David's comment is precisely when I saw the blur of a fluffy white tail whiz by.

I now had everything I needed to start my chase:

  • A source address
  • A destination address
  • A destination tag
  • Proof this was an xRapid transaction

The next step was to start querying the ledger based off the above information. I could have used the XRP Ledger directly, however it would have taken a considerable amount of time and effort. Thankfully, Wietse Wind has graciously made the entire history of the XRPL available via Google Bigquery.

With a few starting tips from the infamous galgitron (Twitter) and XRPGoat (Twitter), I had the basics of how to interact with the XRPL Bigquery database (which can be found here). This my friends, is the moment we dove head first into the rabbit hole to explore the hidden wonderland that is xRapid.


So now, it's time for the fun stuff.

As I stated earlier, this is the only verified xRapid transaction. So what can we gather from that? We can see it was sent from Bitstamp to Bitso. We can also see the destination tag is 25490962.

xRapid_tx

My first two assumptions are:

  1. As it was performed as a live demo, the accounts on each side are Ripple owned exchange accounts, not a customer's.
  2. The destination tags are static, and won't change every time a user makes a withdrawal.

Therefore all transactions to rfEu1Wnr7LxStoFx8DBdzgr8M16FBUbH3K with destination tag 25490962 belong to Ripple.

So first of all let's see how many transactions have been made to Ripple's Bitso account (bear in mind they could have multiple, or changed tags at some stage).

Using this query, we get the following results:

Bitso_ripple_summary

We can see that Ripple sent almost 12,000 XRP over 273 transactions, with the first being on the 16th Jan 2018, and the last on 26th June 2018.

Let's drill down to some interesting aspects of this account.

Using this query, we can see that quite a few transactions were made on the night of the live xRapid demo. Based on the similarity in total XRP sent, as well as timestamps, it appears they were demoing xRapid from approximatiely 4pm until 9pm (time zones could be incorrect though).

bitso_ripple_tx_times-1

Another query will show porvide data showing Ripple consistently perform transactions of similar sizes:

bitso_ripple_size_count_graph

The consistency of tx size is important further on in this research.

Something else of note about this account is the timestamps/ledger close times (some of your keen eyes may have noticed). As started before, this was all happening around Consensus in May 2018. At that stage, there had been zero xRapid partner exchanges announced. The first three partner exchanges were announced in this Ripple Insights post on August 16th 2018:

  1. Bittrex
  2. Bitso
  3. Coins.ph

That's right; Bitstamp wasn't even one of the exchanges listed! It wasn't until 18th December 2018 they were finally announced as an xRapid partner exchange, during an Ask Me Anything with Monica Long and Brad Garlinghouse when talking about xRapid.

If you have a another look, the first transaction from Bitstamp to Ripple's Bitso tag 25490962 occured on 16th January 2018. That's almost a year before Bitstamp was confirmed as an xRapid parter, and almost six months after the final transaction to this account. Obviously this delay could be due to NDAs, it could be due to unfinished testing, or a range of other reasons. However, one thing it does tell us, is that xRapid corridors can be open and running for a long time before they are made public knowledge.

That's about it for this account. They were only small transaction, and are not still actively being sent. Clearly this purely an account for testing xRapid, likely during demos to banks and other financial institutions.

So let's go searching for the bigger fish.

In order to do this I decided to get a summary of transactions by destination tag, based on total transactions, and total XRP sent between the xRapid partner exchanges.

First off the ranks, Bistamp to Bitso (query here):

bitstamp_bitso_ripple_byTag

Woah!
Now that's more like it.

So first things that stick out to me:

  1. The destination tag 30106572 has almost 1000 transactions totalling almost 20 million XRP
  2. Of these results, 7/10 had their first transactions within 2 days of each other (and the Ripple tag, which is #2 in the above list), suggesting they all belong to the same person (Ripple) OR this is the date that testing of this corridor began
  3. Almost all of these accounts/tags stopped receiving transactions around the same time, and every single one is no longer active. Again, this suggests they're either owned by the same person OR testing of this corridor was completed.
  4. Similar to point three; each one of these tags appears to have been active for the same approximate time period as the confirmed Ripple account

I felt like seeing the volume of XRP flowing through this corrider, so using this query I pulled the data showing the volume of transactions:

bitstamp_bitso_ripple_volume_graph-1

We can see that the (presumed) testing of xRapid from Bitstamp to Bitso started off very strong, with over 620,000 XRP being sent on the 17th January 2018, only five days after this corridor opened.
The most consistent time of testing appears to be during early to mid March 2018.

So how big were the transactions? Let's have a look:

bitstamp_bitso_tx_size_graph

In the above graph we can see that the most common transaction size is 20,000 XRP with a whopping 379 transactions. Please note I've converted floating point numbers to integers, meaning these numbers are rounded up the closes whole number. I've also excluded any values with less than five transactions from the above graph (I'll show some of them shortly).

A few thoughts on the above information:

  1. The consistency of transaction size matches that of the confirmed Ripple destination tag.
  2. The larger numbers appear too rounded to be xRapid (although there were days during this time period where XRP was worth $1 USD and $2 USD flat.
  3. Following on from point #2, I believe this could be testing of their xRapid transactions where only the 'sell' half of the transaction is made, emulating an FI who is already holding XRP. This means they could be testing by sending x amount of XRP rather than x amount of USD.
  4. If these are xRapid transactions, there was clearly enough liquidity when XRP was priced between $1-$2 USD for transactions of up to 30,000 XRP (between $15k to $30k USD)

As the above graph excludes transactions that occured less than five times, here's a list of the top transactions by amount:

bitstamp_bitso_tx_size_table

Most of the top ten transactions by size actually look more like natural xRapid transactions than the most frequent, purely based on the close-to but not quite rounded numbers.
Based off this, it's possible that an xRapid transaction requiring over 100k XRP has occured.

There's more information that could be extracted from this XRPL address, however we'll move on seeing as hasn't been active in over 6 months. Luckily, I know another Bitso XRPL address, which may be linked to more accounts. So...


Bitstamp to Bitso (wallet #2)

From this query, we get the following list of destination tags are related information:

bitstamp_bitso2_byTag

Bingo! 3/10 of the top destination tags have sent a transaction in the last 24 hours (at time of writing)! Two of which are the top accounts by both transaction count and total XRP sent.
Another one for those with keen eyes; yes, the top two accounts in the list have identical destination tags as those in the now inactive Bitso XRPL account. They also became active just after the previous XRPL account we explored went inactive.

What does this tell us:

  1. These tags still belong to Ripple
  2. Ripple are still making xRapid transactions / tests from Bistamp to Bitso
  3. Bitso changed which XRPL address they use as their hot wallet

Below you can see the graph for transaction volume across this payment corridor (query):

bitstamp_bitso2_tx_volume_graph

That's a heck of a lot more volume than our previous Bitso address.

Things of note:

  1. Maximum volume in a day was 3,520,040 XRP on 29th December 2018
  2. Second highest volume was 1,367,134 XRP on ..today!??

That's right, as of today 28th March 2019 when I've gone back to try finish this post (everything above that graph ^ was written and data-collected a few days ago).

Let's look at this graph by destination tag (assuming the top two accounts are responsible for both spikes):

bitstamp_bitso2_Volume_byTag

So both massive spikes were from destination tag 96604252, one which hadn't been on my radar until now.
I re-ran my query showing the summary of top tags, and we can see 96604252 has skyrocketed from #5 with 44 transactions, to #1 with a whopping 866 increasing their total XRP sent to 10,000,000.

Old top 3:

bitstamp_bitso2_old_top3

New:

bitstamp_bitso2_new_top3

Here are the transaction sizes by tag 96604252 over the last few days:

bitstamp_bitso2_marchTagVol

In that graph we can see a lot of very similar transaction sizes, but slightly varied, and very high frequency. This to me, looks a lot like xRapid. At the current market rates on Bitstamp for XRP/USD, these transactions are in $500 USD amounts. I'd say that would be quite a common value to be transferred home to love ones back in Mexico at the end of each month.

We can also see the transactions from this tag are being sold on the market, as the Bitso XRP/MXN volumes match perfectly the the earlier volume chart.

Big picture:

bitso_market_long

December 2018 spike:

bitstamp_bitso2_marketVolume_28Dec

Today's spike (March 28 2019):

bitstamp_bitso2_marketVolume_28March

What we can see from above is that the huge volumes of XRP being sent from Bistamp to Bitso are definitely being sold on the other side as there is a direct correlation in volume (and the size of volume also matches).

For anyone who plans to dig into xRapid, the Bitso account rLSn6Z3T8uCxbcd1oxwfGQN1Fdn5CyGujK, destination tag 96604252 definitely seems like one to watch.

Anyway, for no we'll look at some other corridors.


Bitstamp to Coins.ph

Again, we first look at the summary of transaction count, size and dates, organized by destination tag:

bitstamp_coinsph_byTag

From the above table we can instantly see tag #5 will be one to dig deeper into. We can also see that all but one destination has been active this month, a majority of which we in the last few days. This gives me high hopes for this corridor.

Using this query, we'll check out the volume by tag, over the length of their lifespan:

bitstamp_coinsph_tx_volume_graph

As expected with it's 5288 transactions, the destination tag 5 (blue line) was dominating the volume and frequency of this corridor since inception. It had a peak daily volume of ~260,000 XRP. The second most active tag (by total XRP sent) was 1130 and also had a maximum single day volume of ~260,000 XRP.

The elephant in the room; yes that one outlier on the right made 299 transactions, totalling ~1,400,000 XRPP over the last 10 days, with it's peak daily volume of ~461,000 XRP.

A few things to note from the above table and graph:

  1. Almost all activity (and activations) started on or around 8th August, exactly one week before the Coins.ph corridor was announced as live on xRapid
  2. The rest of the activations occured over the last month or two, with the most activity in the last few days
  3. The last transaction of tag 1130 was the day after Bitstamp was named as an xRapid partner. My assumption is this particular tag was purely for testing, and likely Ripple owned

I want to start chasing a few individual transactions, or destination tags that I think may belong to non-Ripple parties (Mercury Fx Ltd for example). Before I do so, I'll quickly summarize the other markets.

Some of this may seem slightly rushed, but I've been writing this up for almost two weeks now, and a lot of my information is going to be old news as I'm seeing similar data being spread on Twitter over the last few days, so I'm hoping I finish this piece before it's entirely obsolete. (Edit: I think I found some exciting stuff. Keep reading!)

Also after some digging, it appears there is minimal to no transactions going back the other way, eg. MXN to USD, or PHP to USD, so I will be excluding that direction from this research.


Bittrex to Bitso (address #2)

Bittrex was named as the first "preferred digital asset exchange for xRapid transactions that move through US Dollars". So naturally we should look at these corridors with Bittrex as the source address.

A summary of top transactions ordered by destination tag:

bittrex_bitso2_byTag

Things of note:

  1. Volume is significantly less than that coming from Bitstamp, despite being announced as partner first.
  2. Almost all the accounts were created the month leading up to the xRapid partner exchange announcement.
  3. A decent number of the top accounts are still currently active.
  4. The only noteable destination tag being sent to at Bitso by both Bittrex and Bitstamp is the known Ripple tag 25490962

Transaction volume over time split into destination tags:

bittrex_bitso2_Volume_byTag

And from 1st August

bittrex_bitso2_Volume_byTag_august

Again we see a large outlier in the first graph, tag 28160779with over 1,200,000 XRP. If we look from the 1st of August instead, we can see a much more even spread of volume occurs. A majority of these tags seem to have gone active within a week of the xRapid partner exchange news.

Unlike the Bitstamp corridors, the volume coming from Bittrex to Coins.ph appears to have dwindled over the last few months, with no recent continualt activity, nor a spike.

Most frequent transaction by amount of XRP sent:

bittrex_bitso2_byTag_amounts

Notable facts:

  1. The most frequent transaction was 1499 XRP
  2. There is quite an even spread of transaction size
  3. Sub-100 XRP transactions are the most frequent demographic

Bittrex to Coins.ph

Summary of top destination tags:

bittrex_coinsph_byTag

Similar to the Bittrex -> Bitso summary, it's clear that there is signifinly less volume flowing into xRapid corridors from Bittrex than Bitso. It also appears that almost all the destination tags at Bittrex became active after the exchange was announced a partner. As we know, this was not the case for Bitstamp, which was active almost a year in advance.

Volume by destination tag:

bittrex_coinsph_byTag_volume

Here we can see that there weren't many outliers, and transfers were somewhat consistent over the course of the lifespan of this Bitso account, however much lower in frequency and size.

Unlike the Bitso corridor however, we can see there is a new spike in the last few days, similar to that in the Bitstamp to Coins.ph corridor.

As you can tell, I've somewhat lost my train of thought, along with my motivation to dig insanely deep into this data. This has probably had to do with being so busy at work, struggling to write this over the last two week, as well as being slapped a $150 invoice for my rather mediocre use of Google's Bigquery (didn't realise it was so expensive).

In light of the above, instead of summarizing the volume's etc of the above. I'll try find the best suiting transactions to match those made by FI's live on xRapid.


The Hunt

Mercury FX's biggest transaction

January 17th 2019 marked quite an exciting time for XRP fans, when Mercury FX announced they'd sent their biggest xRapid transaction to date, boasting "UK-based Mustard Foods saved £79.17 and 31 hours on the transaction".

An important piece of information is that the transaction was sent from the UK, into Mexico, and had a value of £3,521.67.

First let's look at all the transactions from Bistamp and Bittrex to Bitso:

Bittrex:
bitstamp_bitso2_mercuryFX1
Bitso:
bittrex_bitso2_mercuryFX1

I chose to look at transactions between the 14th of Jan (Monday) and the 18th (day after announcement) so that I had any transaction sent that week. The next step was working out the current value of XRP per British Pound at that point in time.

On the 16th of January, the day before the Twitter post from Mercury, the GBP to USD conversion was 1.29. Meaning this transfer was worth $4542.95 USD (which I'm assuming is the pairing traded against on either of the two source exchanges).

According to Bitstamp, XRP was trading $0.324 USD on the 16th of January, making the transaction we're looking for in the realms of ~14,000 XRP.

Unfortunately none of these transactions matched the amount we were looking for. However, it wouldn't be unlikely that the transaction was made the week prior to the announcement, so let's change our search to the week prior:

mistamp_bitso2_mercuryFX1_found

Would you look at that. A transaction for just over 1,200 XRP. A little lower than what we were looking for, but the rates may have been different.

After checking the GBP/USD (1.28) and XRP/USD (0.3664) pairs for the 8th of August instead, we arrive at a total of 12,300 XRP. Would you look at that; almost exactly the size of the above transaction! Taking into consideration fees and slight variation in XRP price during the day, we can account for the 6 XRP difference, and determine it's clear this is the white rabbit we've been searching for.

I have no doubt that this is infact Mercury FXs xRapid transaction, and their Bitso tag is 78980571.

For those who are interested, here is the hash and a link to the first discorverd FI xRapid transaction:
10A2406937A647C76FE9062BC8BD370ECF999E7B5E44B0785C77E5703D705322

Unfortunately I had a look and there are only two transactions to this Bitso tag, suggesting that Mercury have multiple destination tags under their name at Bitso.

mercuryFX_Bitso2

However there are another two from the previous Bitso address, using the same tag (which as we know from Ripple accounts, have stayed consistent):

mercuryFX_Bitso

This is consisten with Mercury FX stating they were testing xRapid in Q1 of 2018.

Moving on.

Mercury FX's first Phillines transaction

On the 5th of March 2019, Mercury FX announced they'd been the first to ever send a commercial xRapid payment through the Phillipines corridor.

I started searching for this transaction too, but with no facts to go off regarding transaction monetary value, I didn't have anything I could say with confidence or evidence was surely it.

So I'll leave you with a brief summary of findings or assumptions.


Summary

  • Bitstamp -> Bitso is by far the largest corridor by volume
  • Both Bittrex corridors are significantly smaller than Bitstamps
  • Barely any volume going back into US (as we'd expect)
  • Ripple own a lot of exchange accounts and are still actively using them
  • Ripple Bitso addresses by tag:
    • 25490962 (confirmed by David Schwartz comment)
    • 30106572 (determined by research)
    • 96604252 (determined by research)
  • Highest destination tag by total transactions and total XRP
    • 96604252
  • Found direct correlation with xRapid XPRL and exchange market volume
  • MercuryFX transaction found:

If you've made it this far, I thank you for having the patience to make it through this horrible mess that is my train of thought during research.

I hope some of this information has been useful, enlightening, exciting, or sparks someone to do a far more thorough job (be careful though, Bigquery is expensive!).

Farewell!


Did you like this post by Cale Robertson?

Send some love: