Jump to content
Guest

MGI, ODL, Bitso and a theory about the price

Recommended Posts

12 hours ago, NightJanitor said:

Thanks.  Great thread.  These spread charts are very, very interesting!

Thanks friend !
We are a very small group of esoteric chart watchers lol - I just like figuring out complex and innovative systems.

I am ALMOST done with my next chart to display the spread samples I've been accumulating (every ten minutes) the last few months. I will start a new thread for this perhaps, hopefully it will be done by this weekend...

Share this post


Link to post
Share on other sites

Just a quick interim note, the last five hours of Bitso trades have been double the recent prime-time average - over $282k USD per hour of volume.

My best guess at this time puts the baseline (spread cost) friction at the range of 0.8% to 1.2% - I will post updated charts at the EOD (UTC) metrics later today.

(or you can watch along too at the live link: https://wp.www-net.com/xrp2mxn-demo-test-page-01/ ) - ten minute updates on the 2nd minute.

Share this post


Link to post
Share on other sites

Quick note: Right now the order books are showing a spread of ZERO at $20k and a PROFIT of 0.2% under. 

The Bitso bids are about 25% deeper than the past averages. And the hourly volume is higher than the recent norms except for two hours. 

We'll see how today plays out.

Share this post


Link to post
Share on other sites
On 1/23/2020 at 10:06 AM, JASCoder said:

(15 minutes later) The profit window is shrinking (see table below), but it's looking like the Bitso "surplus" has vanished... 

1067032438_ScreenShot2020-01-23at10_03_42AM.thumb.png.e18b60d9dd2b9193df608b52c48f2fa0.png

Screen Shot 2020-01-23 at 10.03.52 AM.png

Thanks sharing and keeping this thread up to date! Can you explain in layman terms the significance of rebalancing and surplus in terms of it's affect on price and spread and how to think about this?

Share this post


Link to post
Share on other sites

Quick note for my record here, a minor discovery on my data collection for calculating the spreads...

When I checked this morning's data, I was astounded by the numbers being TOO favorable to be true, so I did some digging and verified the FX rate used by my scripts, cross-referencing with another source. That other source was significantly different, so after making my coffee, finally, I drilled down into my server logs, and low and behold, it was THEN the correct rate was acquired...  My timer was not making the poll on the API at the right time, rookie blunder.
I am chagrined.
FACEPALM :dash1:

This is very frustrating, as all the data I've been creating and recording since mid Dec2019 is inaccurate during a portion of each day. I will have to find a work-around before I put online the historic charts. But first I will finish the live charts (ALMOST DONE!!) and of course fix the FX rate synchronization. 

Maybe my dad was right, I should have been a lawyer, not a coder.

heh.

NOT !!!

Share this post


Link to post
Share on other sites
Guest
1 hour ago, JASCoder said:

Maybe my dad was right, I should have been a lawyer, not a coder.

So you are saying that the fixed figures will now show ODL will drop the price of XRP?

scared ghost GIF by kirun
 

Nah....   just joking.  Thanks for doing it,  thanks for finding it,  thanks for sharing it.

Your Dad was definitely wrong.   :) 

Share this post


Link to post
Share on other sites
1 hour ago, Mpolnet said:

Thanks sharing and keeping this thread up to date! Can you explain in layman terms the significance of rebalancing and surplus in terms of it's affect on price and spread and how to think about this?

Thanks for your question, friend. I'm happy to have found a group sharing my interests in this very obscure topic :)

I am planning to work up a diagram to help explain our understanding of how ODL's processes work, and your questions are a helpful feedback on what to cover.

For now, my best "quick answer," (anyone else please feel free to chime in!!) :

For simplicity of discussion, think of the "corridors" (eg. BS/Bitso) as analogous to two fresh water lakes, separated by a hill, and connected by a pair of one-way pipes that can move water in just one direction, with one able to flow South, and one North.

For the case of Bitso, the North-bound pipe remains effectively dry, while the South-bound is very active, exhibiting a varying flow rate. For the case of say a USD/EUR corridor, we'd expect to see more bi-directional flows. So now extrapolate the water as capital (value), where with BS/Bitso in the North we have USD fiat raining into that lake, and in turn flowing into the South lake via the pipe.

Maybe there's a cloud burst of water in the North, but the flow thru the pipe South remains constant.

In the South, that lake is evaporating, it's hotter down there, and the water is being removed from the lake, to vanish into the atmosphere. 

So, at BS (the North lake), USD fiat is traded for XRP tokens. This affects Bitstamp's order book by consuming the lowest priced "asks," that is, offers to sell XRP for USD. The immediate result at BS, is the effective price of XRP rises, as the supply of the cheapest offers to sell XRP are purchased (trades). 

Meanwhile, AS FAST AS POSSIBLE, the newly acquired XRP held in the local BS account is transferred to the Bitso account and sold (at market) to the "bids" on their order book for MXN fiat. This event right here - the buy-XRP and sell-XRP price differences - is the friction, or spread costs of the event. 

The price of XRP at Bitso will be affected by the consumption (selling off the XRP) of the buy-XRP bids, driving down the local price of XRP on the Bitso exchange. The supply of XRP then grows at Bitso, depressing their local price.

So, what is it that is "evaporating" at Bitso ? The MXN fiat is leaving the system (given to recipients to buy food and pay rent). So if the supply of MXN fiat at Bitso is gone, there can be no market to buy more XRP trying to be sold at Bitso. Meanwhile, in the North, at BS, a MM (Market Maker) there has acquired USD fiat into their account by trading away their XRP. 

At this point, we hit some of the unknowns, such as how fast is the arbitrage occurring, how is MXN fiat on account at Bitso getting replenished (re-balanced with the surplus of USD fiat in the North?). In a perfect bi-directional system, you'd see demands for both fiat and XRP matching up at both end-points, and capital (fiat) movements of fiat would not require re-balancing across the borders...

These are the mysteries we're striving to learn :)  

Share this post


Link to post
Share on other sites

@JASCoder have a look at these accounts.

https://bithomp.com/explorer/rNgHXPviCpf5G11EUFCMs4YR33s3puBv8Y

https://bithomp.com/explorer/rModuA9GBakB1hUKZxXYF3sQP7a9qbpooR

I was looking at bitso trade data and can see that there has been a significant net purchase of XRP since MoneyGram started operations.(I don’t have the numbers in front of me, but 100’s of million XRP/month) It looks like the 2 wallets above, take the accumulated XRP from the ODL exchanges and then send it back to Bitstamp to complete the loop. I’m guessing the market maker is using the traditional banking rails to fund their MXN at Bitso, they accumulate XRP and send back to Bitstamp to cash out. 

Share this post


Link to post
Share on other sites
3 hours ago, EasterBunny said:

@JASCoder have a look at these accounts.

https://bithomp.com/explorer/rNgHXPviCpf5G11EUFCMs4YR33s3puBv8Y

https://bithomp.com/explorer/rModuA9GBakB1hUKZxXYF3sQP7a9qbpooR

I was looking at bitso trade data and can see that there has been a significant net purchase of XRP since MoneyGram started operations.(I don’t have the numbers in front of me, but 100’s of million XRP/month) It looks like the 2 wallets above, take the accumulated XRP from the ODL exchanges and then send it back to Bitstamp to complete the loop. I’m guessing the market maker is using the traditional banking rails to fund their MXN at Bitso, they accumulate XRP and send back to Bitstamp to cash out. 

Very cool info Bunny !
Thanks.

I did two spot checks, of hours 14 and 23 for 24JAN, in the transaction history - and reconciled them with my live chart's hourly volume metrics (see my char's link), and both were right in the ballpark (whew!!).

Cheers

Share this post


Link to post
Share on other sites
12 hours ago, EasterBunny said:

For reference.

It has been growing quite consistently and is clearly part of a rebalancing account for mxn-usd corridor. My guess is that the account belongs to a payment provider using ODL for mxn-usd taking advantage of the cheap(ish) xrp at bitso to make the return path. All these corridors have to have return paths. What mades the rModuA9GBakB1hUKZxXYF3sQP7a9qbpooR account different from the other ODL ones is that the xrp flows though it from bitso->bistamp rather than directly from bitso->bitstamp.

 

I consider it part of the ODL traffic and include it in my analysis. Today's graph (based on data until midnight UTC) show the ODL traffic is still increasing and we have hit the region of $6million in a day now. This is much faster growth than I had been expecting, but the data is still quite noisy and another 6 months' worth of this will be very informative.

image.thumb.png.4870c8cee2df6a7388a3a569b718e8a5.png

(I don't yet know what to make of the account rNgHXPviCpf5G11EUFCMs4YR33s3puBv8Y - it crops up in my statistics as an interesting one, but I don't know anything about it).

 

13 hours ago, JASCoder said:

The price of XRP at Bitso will be affected by the consumption (selling off the XRP) of the buy-XRP bids, driving down the local price of XRP on the Bitso exchange. The supply of XRP then grows at Bitso, depressing their local price.

This is only true to a certain extent. What people on xrpchat seem to ignore or forget - is that there are just as many people/companies potentially wanting to buy xrp with MXN at bitso and sent it back to the USA for USD (you only need, in principle, one payment provider on ripplenet with a few thousand customers in mexico who import from the USA to make this happen). The whole point of these corridors is to match flows in both directions - this actually puts buy pressure on both sides of the market. (It's only half as big since as the spread is being split in a manner of speaking by both sides benefiting from the rates).

 

Share this post


Link to post
Share on other sites
4 hours ago, jbjnr said:

This is only true to a certain extent. What people on xrpchat seem to ignore or forget - is that there are just as many people/companies potentially wanting to buy xrp with MXN at bitso and sent it back to the USA for USD (you only need, in principle, one payment provider on ripplenet with a few thousand customers in mexico who import from the USA to make this happen). The whole point of these corridors is to match flows in both directions - this actually puts buy pressure on both sides of the market. (It's only half as big since as the spread is being split in a manner of speaking by both sides benefiting from the rates).

Very astute of you to catch this, JB. After re-reading my post, I thought to clarify this oversimplification, as obviously the MM buying the tranche will be unlikely to immediately put it up on offer, especially at the swamped endpoint eh ?

I've watched the arbitrage processes work very fast, in one case of a ten minute span, not only 100k additional XRP got sold, but the bids and pricing was restored. This was back when the hourly averages at Bitso was under 500k XRP volume.

My new live chart of the spreads will hopefully shed more insight into the process. As this friction is a key component impacting trade decisions.

Please review my approach, and share your feedback:

  1. Generate a spreads matrix of a range of USD amounts every ten minutes 
  2. Aggregate the samples into a simple average and plot cost points vertically along timeline (X-axis)

If you peruse the transaction history of the Bitomp display shared by Bunny, you'll observe the to-Bitso tranches are coarse in their granularity, I find surprising considering the exposure to potential price fluctuations.

I have scripted my code to enable tuning of the polling interval for the sampling, to enable future tweaking. Perhaps this parameter should be driven algorithmically, factoring for the XRPL activity ?

Cheers 

Share this post


Link to post
Share on other sites
On 1/25/2020 at 6:11 PM, JASCoder said:

Please review my approach, and share your feedback:

  1. Generate a spreads matrix of a range of USD amounts every ten minutes 
  2. Aggregate the samples into a simple average and plot cost points vertically along timeline (X-axis)

The answer to that really depends on what is it you're trying to do or know! I do no know what your motivation for examining the order books is.

I try to track ODL activity because it's one of the key parameters in my models of xrp price, so I'm interested in tracking the total volume. Other parameters that are useful are the spread between bid/ask across the corridor - however, I can assume it will lie between (say) 5 and 50 basis points and that's good enough for me to chart parametrically. Knowing how much xrp market makers are holding (and or offering) in different price brackets would be very useful to know. I'm not certain that that is something that can be directly inferred from the trades, but it is reasonable to expect that the amount being held will be in some way proportional to the flow rate of ODL through an exchange, so maybe it is possible to derive some metric from the data you are collecting by analyzing the change in bid/ask as ODL trades are passing through. I have not thought about this deeply ...

On 1/25/2020 at 6:11 PM, JASCoder said:

I've watched the arbitrage processes work very fast, in one case of a ten minute span, not only 100k additional XRP got sold, but the bids and pricing was restored. This was back when the hourly averages at Bitso was under 500k XRP volume.

I will ponder this.

Share this post


Link to post
Share on other sites

@jbjnr Thanks for your feedback - I feel we're on the same page, but I'm a coder not a trader :) 

Please share your knowledge with me: I've seen the word "spread" used in the context of an order book, as a metric only factoring for the difference between the lowest bid and highest ask price of an order book. Is there another term properly used for the baseline cost estimation of an ODL paired trade (i.e. X amount of fiat spent to acquire XRP, then same XRP sold for fiat) this is the slippage maybe ? (Same term if a gain is realized?)

Have you by chance already seen my BS/Bitso DoM charts and Cost Estimates tables in other posts here ? I'm concerned about getting the terms and labeling right, your input is greatly appreciated.

Thanks for your time.

Share this post


Link to post
Share on other sites
Guest
3 hours ago, JASCoder said:

I've seen the word "spread" used in the context of an order book, as a metric only factoring for the difference between the lowest bid and highest ask price of an order book

I think you have that backwards.   The spread is between the highest bid (what someone will pay) and the lowest offer (what someone will sell for).

Slippage is how much more than the original lowest offer you ended up paying after the order book offers were eaten up by your purchase.   Or visa versa if you are selling.

You probably know all that so apologies but I thought it might help someone at some point,

Share this post


Link to post
Share on other sites

×
×
  • Create New...