Jump to content
Sign in to follow this  
ICookinBrine

Post of the Year 2018

Recommended Posts

Is there any interest in a year-end list of the top posts to XRP Chat? If so, voting may be unwieldy, so I suggest a nominating process followed by a decision from the moderators.

This idea was sparked when I was reading a comment in the Island of Misfit Toys (i.e., Off-Topic) and thought there was a lot of quality content that may have been overlooked throughout the year elsewhere.

Edited by ICookinBrine
Typo, correction

Share this post


Link to post
Share on other sites

jbjnr with his opening post in the thread "Analysis of Ripple's XRP Sales" is a must read:

On 11/17/2018 at 7:30 PM, jbjnr said:

I would like to present here my findings on a slightly more detailed examination of sales from the ripple XRP-II wallet (rHjJwY4maqRyUxCnJ9bNNQ2Tva9fe9T8ud) and also the 3 wallets rHjJwY4maqRyUxCnJ9bNNQ2Tva9fe9T8ud, r49yezViZ8N6FtwwKg9byDYtJ3UDuJi1zv, r3epSojEkn6pFRwjxkFjwJLPAxC8Hj9ZTK that we believe are ripple controlled and have been supplied to my by @xp3215233 who deserves your thanks for the numbers I have extracted and plotted here. I will refer to the 3 wallets hereafter as RP1, RP2, RP3 to indicate that we think they are ripple controlled. In addition, I will also include some details of sales from the "BearWhale" account and Jed's tacostand account as a tag-along bonus at the end.

There is no particular agenda to the analysis presented here, other than to further our understanding of the XRP ecosystem and broaden our knowledge (in the spirit of DYOR). The analysis I presented prior to the Q3 results release was a bit sloppy and I thought I would try again with a slightly more rigorous approach.

Disclaimer : Although the numbers/dates extracted from wallets can be considered 'facts', the actual ownership/control/motives of wallets and transactions through them is mostly unknown and suppositions are based on the collective guesswork and historical evidence gleaned from resources such as xrpchat and previous ripple forums. Other than the XRP-II wallet which is known to be owned and controlled by the ripple subsidiary XRP-II and is licensed to sell XRP, I do not actually know for certain who controls any of the wallets mentioned in this post, and therefore any discussion or conclusions made here must be considered speculation. Nothing else that you read here should be taken at face value and you should even consider the possibility that everything presented here is forged material, made up just to manipulate the market.

Let's start by looking at all the sales from the 4 (ripple controlled?) wallets going back to last year. Note that in the following plot, transactions between accounts are not included, only those that leave them are shown.

ripple-wallets-source.thumb.jpg.f0a5071d06f9e38b5657a88fcee6136c.jpg

The plot shows payments from the 4 wallets in separate colours to distinguish and identify their origin, overlaid on the total sales volume on exchanges (as returned by queries to https://min-api.cryptocompare.com/data). It is immediately clear that the 4 wallets are being used to make  transactions with different 'sales' patterns. (The word sales is quoted here because in general, I assume that transfers from wallets are for the purpose of sales, but in many cases they may not be, however the term 'sales pattern' is a nice one to describe activity). Note that the plot uses linear scale (left) for the (black) volume from cryptocompare, and logarithmic scale (right) for the payments as they span a wide range of values and are best viewed thusly.

The patterns become slightly more interesting when we look at the destination wallets of the payments, as shown in the following plot where we see the same data coloured by destination rather than source. I have tried to cross reference the destination account with the names of wallets used by exchanges. Again, I stress that the names presented here may not be correct, however, the chances are high that some/most of them are valid and what we care about is the pattern, rather than the specifics of each payment.

ripple-wallets-dest.thumb.jpg.46fb766a3828f983d43a9178adcd546e.jpg

The coloured dots represent payments and the solid line above is the daily sum of those payments, all are again overlaid on the daily volume from cryptocompare and a logarithmic scale is used for payments and their daily sum. The logarithmic plot compresses the sum line considerably and so it is not immediately obvious that their is any correlation between the transfers from ripple and the daily volume.

Let’s have another look, but this time we will add another filter, we will remove all payments that went to addresses/wallets that do not have a destination tag. We do not know the addresses of all exchanges, but most exchanges require the use of a destination tag to separate client accounts. (GateHub is one notable exception where each user has an individual wallet).  Filtering by tag gives us a reasonable approximation of payments that are going to exchanges and we can see the list of destination accounts is significantly reduced.

ripple-wallets-dest-tag.thumb.jpg.467aad2ed98b1500bd34d45811c94339.jpg

Some things are obvious from the plots of payments: The sheer number of payments from these wallets during the second half of 2018 is quite staggering, and the fact that a very large number of them appear to be going to exchanges is very interesting. There do appear to be correlations between payments/sales and trade volumes, though those correlations change subtly over time and it is quite evident that in early April 2018 the sales strategy of ripple changed dramatically. It is of course possible that ripple uses other wallets for sales and simply changed at this time to one that we are tracking, from one that we do not know about.

The pattern of payments changes on 12th April 2018 so let us examine more closely the pattern from that time onwards, excluding the payments from the XRP-II wallet. (This wallet is believed to be responsible for “direct” sales rather than programmatic and careful observation of the first plot shows that it makes less frequent, but often very large payments that do not look like programmed sales to an exchange).

RP-1-2-3-zoom.thumb.jpg.28777533680655cf42a9f46835723589.jpg

This plot shows us payments from RP1,RP2,RP3, only to accounts using a destination tag and it is now quite clear that there is a correlation between payment volume and exchange volume. There is also a mysterious gap in payments between 12th and 19th of September this year. The correlation is a little difficult to see due to the log/linear comparison, so lets compute the correlation time shift and normalize the plots to see how they compare and overlay them using linear scale for both.

cross-correlation.thumb.jpg.e2775c901deb866e8dcf674731b1f666.jpg

The correlation is good, the cross correlation function has a peak that appears with a time shift of 1 day. This plot shows the ripple sales volume (blue) time shifted by 1 day (earlier = to the left) and overlaid on the exchange volume (black). Normalization of the two volumes give a factor of 29.19 difference between the ripple sales and the exchange volume - this seems to me to be a fairly remarkable result.

correlation-text.jpg.15d58ffb3365600b8bf5a8bd86a67c85.jpg

It tells us that ripple are selling XRP at approximately 1/29.19 = 0.034 = ~3% of daily volume based on the previous days exchange total. This is far higher than I had expected and higher than reported in their quarterly sales figures.

The figures declared by Ripple are as follows

text-ripple-sales.jpg.40113df5a6bdc1c2bf75922b7b287a90.jpg

Note that we do not know if the figures for 2016-Q4, 2017-Q1 are direct or programmatic, but I put them in the direct column since it would appear that their programmatic sales have been better coordinated more recently than in the past.

Whilst the correlation is good (with time shift we have Pearson coeff 0.88), with the peaks and troughs lining up extremely well, there are some deviations. Around the region of Jul-Aug on the graph, the (normalized) volume sold by ripple appears slightly higher than the exchange volume, so it is possible that they have been adjusting the amounts slightly over time on a month by month basis to see how the market is affected. It’s also possible that the volume being used for calculations is coming from another source (i.e. not cryptocompare) and is slightly different, it is also possible that I am including sales from wallets that are not part of their programmatic sales, or other wallets exist that I am not aware of. Certainly we can see that when the trade volume rises and the xrp price rises, the programmatic sales increase the next day and this in turn brings the price back down. The ongoing bear market (at least for XRP) has not been assisted by the selling pressure of 3% per day - especially when we consider that significant parts of the other 97% of trade may be of suspicious activity on largely unregulated exchanges.

How do the sales we have extracted compare in terms of $$$ with those declared by ripple in their quarterly reports? We assume here that the price that the XRP is being sold for is the mean price on the day it is transferred, this is at best an approximation, but since the transfers are happening every day (at least for RP2 since April), we must assume that the XRP is actually entering the market and not simply accumulating in wallets on the exchange.  

test-RP-1-2-3-sales.jpg.9d00ffacd3572237ed3059b86dca5354.jpg

edit: ERRATUM: The table shows $$$ amounts for the quarters, but the right column total is the XRP amount, the $total is less. Apologies. .

The numbers here are much larger than those declared by ripple (and shown earlier). I have experimented with different combinations of summation from different wallets, using different combinations of high/low/closing price, but have not been able to find a set that reliably reproduces the figures from ripple.  What hypothetical conclusions can we draw from this

  • Ripple are telling outrageous lies about the amount they sell? This seems highly improbable. Everything we know about the company indicates that they are “Very Serious People” in the right sense of the words (if you pardon me parodying a phrase partially popularized by Krugman).
  • Ripple are making the payments we see on the ledger, but they are not all ripple sales. This is my own best guess. They have setup bots to make transfers to exchanges. Many many exchanges, far more than I imagined, and now that this infrastructure is in place, they are using it for their own programmatic sales, but reusing it for sales from staff/founders wallets. Why not sell some fraction of daily volume direct from ripple, and include some of the other XRP they need to offload. This way, they only have to worry about 1 set of scripts and can do the accounting of who is owed what internally within the company.
  • Ripple are looking after us. True that they are ‘dumping’ massive quantities of XRP into the markets which is certainly keeping the price down and preventing wild speculative fluctuations - from the sums above, we are looking at around $500m (edit: not $1B as written previously) in total since 2017-Q4. This is interesting because it is ‘new money’ entering the markets. This is not wash trading between day traders on exchanges, but continuous drip feeding of XRP into the markets that is building the liquidity that will be needed when large scale adoption of the “internet of value” kicks off.
  • Ripple are looking after us. Look closely at the graph of sales for September 2018. Ripple sales stopped for 1 week between 12-19 of the month, apart from a small amount that was probably setup on a different trading algorithm from the main programmatic sales that started after April 12th. Why did sales stop that week? I chose to overlay the price on the volume correlation in the plot above because on Sep 11th the closing price (as cited by cryptocompare) of XRP was around $0.26, the lowest value of the year and I suspect that ripple decided that it would be a good idea to stop their sales temporarily to help the market. Sales stopped and the price began to recover. In fact by the end of the week  it shot up dramatically, partly based no doubt on news about xRapid adoption, as there was much talk prior to swell - but the correlation between a sudden doubling of XRP price right after the cessation of ripple sales seems to me to be extremely interesting. Correlation does not imply causation and we cannot be sure, however, should ripple cease sales again for some reason and a similar effect take place, we will have better evidence to support our hypothesis. (If anyone from ripple is reading this, then please feel free to temporarily shut down sales so that we can see if the price spikes again).
  • The market is healthy. We have only shown what we believe are programmatic sales in the correlation and summations, but XRP-II is also making direct sales that add to the total. In addition there are ongoing sales from the bearwhale and Jed and possibly many many more wallets that we are not aware of, that are all being absorbed by the market and it appears to be holding up very well.

Up to this point we have looked at the sales from RP1, RP2, RP3, but XRP-II which we believe to be the direct sales account. Here are the payments from XRP-II to accounts with a destination tag, the total is around 264 million XRP paid (not the $ amount since we do not know the sale price).

text-XRP-II-sales.jpg.31668bdd527c84d0b01fd1c556d97ddf.jpg

If we remove the restriction of only counting payments to tagged accounts, then the figure jumps to a tidy 3.8 billion XRP transferred since Q4 2016. Not bad at all.test-XRP-II-no-tag.jpg.4c2c18f9bf4e4d9eea3577b38b4ee881.jpg

Let’s take a small diversion and add Jed and the bearwhale to our analysis to see if it gives consistent figures.

jed.thumb.jpeg.86d7add985012bf296245ba611003938.jpeg

Payments from Jed’s accounts do not correlate well with exchange volume unless we only consider Feb 2018 onwards. There are some gaps in sales where trading appears to have temporarily stopped, however if we consider sales from 2018-02-01 (Feb 1st 2018), we have a very good correlation with volume with a time shift of 13 days and a normalization coefficient of 221 which equates to a daily selling of 1/221 = 0.0045 or 0.45% of daily volume. This appears to be consistent with the agreement which appears to allow 0.5% of daily volume based on the average volume of the period 2 weeks prior (please inform me if I have used the wrong agreement details).

The BearWhale remains our most ardent foe, with semi random transfers to exchanges that happen when the volume increases. Presumably the sales then take place in steady chunks on the exchanges, and we cannot track them individually, but we can assume that the sale happen over the days and weeks following transfers to exchanges. Using 24 account that are believed to be BearWhale related, we see the following pattern.

bearwhale.thumb.jpeg.713668802438f25ac716b9232689a5dd.jpeg

Out of curiosity, if we sum the sales from BearWhale, Jed, and Ripple and assume they are all traded at the price on the day, what do we get.

ripple-jed-bw.jpg.f75f5dc73ff0f1a19b68e560f353b85a.jpg

edit: ERRATUM: The table shows $$$ amounts for the quarters, but the right column total is the XRP amount, the $total is less. Apologies. .

The sums are quite impressive.

Note that Jed's tacostand is not included in the figures above as it send to a non tagged account, so please add $50m (not $131m as written previously) to the total.

tacostand.jpg.7c0a96af2fe634fa9d56f9dd21f4bb11.jpg

edit: ERRATUM: The table shows $$$ amounts for the quarters, but the right column total is the XRP amount, the $total is less. Apologies.

To conclude this post, I am rather stunned by the scale of xrp sales that are going on, whilst many of the numbers presented here may be incorrect, and not attributable to ripple, jed or the BearWhale, the fact remains that huge amounts of xrp are being purchased and we are still well over double the price of this time last year. I am very confident that if and when xRapid is adopted at scale, we will see further improvements in the market.

 

Share this post


Link to post
Share on other sites

NightJanitor's description of Proof of Work (overlooking the language for this one, tsk-tsk):

8 hours ago, NightJanitor said:

"proof of work"

It's like if Henry Ford had died just after inventing the assembly line and then his retarded cousin took over the factory and ordered all the workers to bunch up, at the exact same station, to do the exact same task, and then he picks the best looking part, throws all the rest away (and pays the one worker who created the one part) - and then goes:   "Alright, everyone to the next station!"

I still can't get over how otherwise supposedly very smart and engineering-minded people think this is so brilliant, but it's really not, so maybe they're just susceptible to ********.

I dunno.  Maybe I'm the idiot.

But I doubt it.

Share this post


Link to post
Share on other sites
On 12/4/2018 at 3:49 PM, ICookinBrine said:

jbjnr with his opening post in the thread

There were a number of brilliant posts last year but that one was definitely impressive and I said so at the time.

I know I shouldn't be voting if I need to judge the winner but I want to add my own favourite. It's a thread rather than a post and is so deep and complex that even a second read was not enough to unpick all the themes (I mean, it includes a recommended reading list and lecture series so I'll be going back to it multiple times this year). Credits go to @KarmaCoverage @tar and @Wandering_Dog:

 

Share this post


Link to post
Share on other sites
On 12/4/2018 at 10:45 AM, ICookinBrine said:

Is there any interest in a year-end list of the top posts to XRP Chat?

I think this is a great idea. Even as a basic metric, the top 5-10 posts that gained the most reactions? Is it possible to pull something like that together?

Share this post


Link to post
Share on other sites
Sign in to follow this  

×