Jump to content
Molten

MGI, ODL, Bitso and a theory about the price

Recommended Posts

5 hours ago, jbjnr said:

Does anyone know anything about account rModuA9GBakB1hUKZxXYF3sQP7a9qbpooR

I've started seeing significant traffic on bitso->rModuA9GBa but I do not know what exchange (or bot trader) it is. It might be a rebalancing of the MXN via some other corridor.

edit: NB traffic over the last few days has been rising quite steeply

Just a guess here, but with the BS->Bitso volume creeping up, I'd expect some surplus of the bought XRP acquired at Bitso to re-balance some of itself back to its origin, BS. Might we see some correlation with BS's order book ASKs matching up shortly after the scored transfer in the ledger ?

I am working on capturing and visualizing some of the DoM metrics at the two exchanges, watch for a new thread started soon when I have something to post. It could help us spot such a pattern as I hypothesized.

Cheers m8.

Sorry, I'm still getting up to speed here, what does "NB" stand for ?

Share this post


Link to post
Share on other sites

Quick update on this morning's (American time zones) activity...

On my "ODL Live charts" page I see a very confounding depiction:

For the UTC hours of 16-18, it shows XRP hourly volumes traded of: 2.4/.7/.5 million.

FOR THE SAME HOURS MXN prices (Blue/Green/Purple == Lows/Ave/Highs in MXN) moved like so
(in a tight range of about 0.06 MXN):

1575532533_ScreenShot2019-12-06at11_54_19AM.png.18787cc88ea78a4cdb8535ab59a11043.png

I am struggling to suggest any theories for this counter intuitive behavior heh...

Share this post


Link to post
Share on other sites

I've an evolving theory what might be going on with the BS/Bitso corridor...

After a few weeks of monitoring its daily trade patterns - which have really varied widely a number of times, almost like live experiments being tried out - and then reconciling with price fluctuations and their order books (see below), I've become convinced there must be some XRP "inventory management" behavior in the ODL system.

That is to say, the process acquires and holds some amount of XRP, and executes its cross-border trades at more optimal times. If you look at the hourly volume spikes I've posted prior here, there is no way the current order book would efficiently accommodate such high volume trades. 

png_download55023.thumb.png.56f6b154711659fb22f0d2ec05a45a85.png

This hunt for puzzling out ODL's secrets continues... 

Share this post


Link to post
Share on other sites
On 12/8/2019 at 2:02 AM, JASCoder said:

I've become convinced there must be some XRP "inventory management" behavior in the ODL system.

That is to say, the process acquires and holds some amount of XRP, and executes its cross-border trades at more optimal times

This looks a lot like the implementation of Bob Way's patent 

 

Share this post


Link to post
Share on other sites
On 12/7/2019 at 5:02 PM, JASCoder said:

If you look at the hourly volume spikes I've posted prior here, there is no way the current order book would efficiently accommodate such high volume trades. 

png_download55023.thumb.png.56f6b154711659fb22f0d2ec05a45a85.png

Thanks for this! Could you elaborate a little more on the highlighted sentence above in your post? Apologies in advance for my confusion. Does DoM help us determine the available liquidity based on existing orders?

Share this post


Link to post
Share on other sites

@Dario_o Thanks for the referring link, this will take me some time to re-read (a few times), I think it's directly relevant for sure mon.

One thing he mentioned - re RL not wanting to help clue in competitors - could explain why we're not seen any in depth descriptions of the details of the process. 

I have to again mention here that it's not stabilized yet, this is still an on-going work in progress. 

If you saw this morning's hourly spike event, and the corresponding MXN pricing at that same time, it's very confusing - there must have been a large ASK that was executed. NEED MOAR DATA lol

Share this post


Link to post
Share on other sites
16 minutes ago, Mpolnet said:

Thanks for this! Could you elaborate a little more on the highlighted sentence above in your post? Apologies in advance for my confusion. Does DoM help us determine the available liquidity based on existing orders?

Howdy friend - I'm always delighted to share with interested people :)

Check out below the single hour spike (at 16:00) of XRP buys at Bisto - then note how in the stale DoM chart above it's only is showing about 1.3m XRP worth of bids - NOTE: I'm only showing top 50 bids at Bitso, I need to bump that up a lot I'm now thinking... I am actually now capturing and saving the DoM data to a database, and soon will be able to "scroll" back in time to see how the DoM profile looked at any given moment (I'm saving ten minute window snapshots). That will make for some interesting data mining.

To answer your question, ponder this (somewhat simplified description) situation:
The "ODL-bot" is holding 500k surplus of USD fiat on their account at BS.
It monitors the DoM shown above (ASKs at BS, BIDs at Bitso) - waiting for a trade window to appear which is in range of its rule criteria.
When the two exchanges report a qualified pair available, it makes the buy at BS, then "deposits" it to their Bitso account, then sells it for MXN (it's actually more messy than this, since we know there must be some loose XRP inventory being held in accounts at both exchanges).

I hope this helps make it more clear ?

I am now trying to code a script to calculate the fully executed costs of these two events, for several different amounts of fiat - this info I think will be very helpful in monitoring both the efficacy of the process, as well as the evolution of the two markets as spreads shrink and the pools deepen. Right now I'm a bit stuck, as the slippage numbers I'm getting seem way too high. Hopefully I'll have a breakthrough soon so I can share this with everyone here, and start charting that as well.

LMK if you have any follow up Qs or comments. Cheers ! 

png_download87159.thumb.png.ee586f6e2bfd15d723e7a4f8545aea5c.png

Share this post


Link to post
Share on other sites

Thanks for sharing all this.

I’m a bit confused...  why is the postulated ODL bot doing this? (Trading this lump of USD for MXN Pesos)

And may I ask if you are perhaps falling into the trap that the Ellis theory people do...  assuming the static image of the order book is fixed when in fact it’s a dynamic system.  When you say ‘the order book cant accommodate that big trade’ isn’t it possible that the MM respond real time to the big trade by adding in orders to match just after it arrives?

Thanks for any light you can shed on my confusion.

Share this post


Link to post
Share on other sites
7 minutes ago, Tinyaccount said:

I’m a bit confused...  why is the postulated ODL bot doing this? (Trading this lump of USD for MXN Pesos)

( if the following is obvious to you, then i have failed to understand your question )

As MGI takes in USD fiat from their agents and customers, that fiat eventually gets deposited into MGI's bank account. Additionally, some of the electronic bank transfers by customers also accumulate in MGI's bank account. I'm not privy to all the amounts and steps and so forth - but suffice it to say MGI is acquiring USD fiat by the millions - and paying it out in Mexico.

So, in Mexico, MGI is paying out to recipients in Pesos to complete the remittance event. So as MGI's MXN capitol gets depleted, it needs replenishing of more Pesos to maintain ongoing operations.

Does this make sense, or have I totally misunderstood your question ?

Regarding your question on "Ellis theory" :

Yes, the order books are very dynamic, it's a torrent of data - all I can do is take snapshots of it to perform a sampling and discern general trends and patterns. In fact, I've even posted here previously about my witnessing very fast "re-balancing" of the order books after seeing large trades on the Bitso side. Much of the MM actions and capabilities are effectively hidden from us, and can only be guessed at with additional monitoring.

For myself, puzzling these dynamics out is one of my goals in this hobby project :) 

I'm now starting to wonder if my hourly charts are not granular enuf. Digging down into this rabbit hole keeps revealing more branches the deep we goes !

Share this post


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

So, in Mexico, MGI is paying out to recipients in Pesos to complete the remittance event.

Thanks for your reply.  I don’t think this is correct.  MGI are not paying our recipients.  Essentially MGI took dollars and then flicked it into the ODL system.  The dollars they took are slightly higher than the ODL requirement so MGI have dollars left over on hand.  That is their profit.

Now the entity ‘paying out’ in Pesos is whoever wanted to buy those incoming ODL XRP.  (Not strictly true...  The ODL system does the actual payout,  but the Pesos came from anyone who wanted XRP).    It could be literally anyone.  It probably isn’t MGI unless they are vertically integrated and want to be Market Makers as well.

So given that,  I don’t see why there is a stack of Dollars in Bitstamp needing to go to Mexico as Pesos.  The two ends are not necessarily the same entity.  It IS possible that at this early stage there actually is only the one entity at both ends and that your analytics has identified that fact.  But it is not a necessary condition.  Certainly it doesn’t need to be MGI.

So I don’t know if we are in agreement or not...  but I do not see an absolute requirement to move USD to Pesos as you do.

Share this post


Link to post
Share on other sites
48 minutes ago, Tinyaccount said:

Thanks for your reply.  I don’t think this is correct.  MGI are not paying our recipients.  Essentially MGI took dollars and then flicked it into the ODL system.  The dollars they took are slightly higher than the ODL requirement so MGI have dollars left over on hand.  That is their profit.

There's certainly a lot of mystery to the process, so we are left to make guesses, looking for confounding or affirming evidence.

And just to be clear, I don't say I know everything about how all this works,
it's why I've been researching all I have time to pursue :)

Firstly, let me start with a simple use case:
Say I enter MGI agent's store, and hand them five $100 dollar bills.

Certainly you know that fiat isn't - itself - flicked into the ODL system right then and there ?

I'd assume at the end of the day, all the day's unwashed currency is counted up, put in a canvas bag with a filled in deposit slip, and dropped in the night box at their local bank's branch ?  

But meanwhile, MGI must have a messaging system already in place, which will affirm to the Mexico based MGI outlet that it's ok to hand over the equivalent value in Pesos to the recipient who is picking up the "transferred" funds ?

( i'll stop here for now, so we can sync up and hopefully get on the same page before proceeding )

Share this post


Link to post
Share on other sites
6 minutes ago, JASCoder said:

Certainly you know that fiat isn't - itself - flicked into the ODL system right then and there ?

Hehe...  yeah.   
 

7 minutes ago, JASCoder said:

But meanwhile, MGI must have a messaging system already in place, which will affirm to the Mexico based MGI outlet that it's ok to hand over the equivalent value in Pesos to the recipient who is picking up the "transferred" funds ?

Ah ok...   perhaps this is the bit you are missing.

MGI in Mexico probably do not hand out the pesos.  What actually happens (if my understanding is correct) is the Pesis arrive in the Recipient bank account.  
 

Quote

must have a messaging system already in place,

That messaging system is Ripplenet.  The ODL transfer moves destination fiat money into a bank account.  In cases where MGI want to physically hand over notes at the destination country (as requested by their customer) then ODL will transfer the fiat into the MGI bank account in the recipient country.  The counter staff can check it’s arrived,  and when it does (inside a couple of minutes at most)  hand over notes from cash in hand.
 

So MGI do not have to take a pile of USD and convert it to Pesos to top up their Mexican end.  The value was truly transferred and transformed into the destination fiat.  So this is why I’m puzzled that you think they need a bot doing this for them.

If there is a balancing bot it is much more likely used by a Market Maker who is serving both ends of the corridor.  As I said earlier,  it IS possible that MGI are vertically integrated and are operating as market makers...  but that’s not really germane.

Thanks for your patience and interest and the work you are doing.

Share this post


Link to post
Share on other sites
7 minutes ago, Tinyaccount said:

MGI in Mexico probably do not hand out the pesos.  What actually happens (if my understanding is correct) is the Pesis arrive in the Recipient bank account.  

It's great to find someone as yourself to discuss this obscure topic with, thanks for your time and patience too :)

Maybe we'll luck out soon and an actual SME will pop in and spell it all out to us haha.

Indulge me if you will, let's see if we can better drill this down, to the process' component steps...

Our BS->Bitso use case is a closed system, the capital flow is only moving in one direction. That is to say, MGI's bank account (in the US) is both receiving fiat deposits, and making local electronic transfers to their Bitstamp account. I think we're in sync up to here.

So, the next steps maybe go like:

1) Asynchronously, some amounts of the fiat sitting in the BS account are traded for XRP.
2) Also async, some amount of XRP is "withdrawn" from BS account and deposited into Bitso account.
3) Trades are made at Bitso, selling XRP for MXN, which boosts the local fiat account balance there at Bitso.
4) Fiat in Bitso account gets transferred to local Mexican based MGI bank account.
5) MGI performs various bank-to-bank transfers to recipients' (and agent?) accounts.

So far, are we in sync at this point, or have do I have some errors or omissions ?

Share this post


Link to post
Share on other sites
43 minutes ago, JASCoder said:

It's great to find someone as yourself to discuss this obscure topic with

I do enjoy trying to understand it all...  but I am far from the best person to discuss this with.  Perhaps others will see and chime in as needed.

 

43 minutes ago, JASCoder said:

So far, are we in sync at this point, or have do I have some errors or omissions ?

I agree with what you listed as numbered points up to 3.  However I think point 4 is problematic.  I think we need to determine if this is a ‘straight into recipient bank account’ call it Case A,  or a ‘cash out at MGI counter’ transaction call it Case B.

Currently I think it’s most likely that the bulk of traffic in the Mexican situation is straight to recipients account.  This is based on Ripple saying one of the great things about ODL in the USD/Mex$ corridor is that the Mexican internal banking rails are real time.  The money leaves Bitso and arrives in the recipient account very fast.

Of course the strength of MGI is its local presence so they can provide over the counter Pesos to the recipient if desired (case B )

So if this is a Case A then the subsequent steps 4 and 5 are not required.

If case B then instead of your 4 and 5 I would add simply 

4.  Counter staff watch the MGI internal bank account till Pesos arrive and then give to recipient.

No bank to bank transfers required.

 

Edited by Tinyaccount
Emoji popped up instead of ‘Case B’. Moved bracket to fix.

Share this post


Link to post
Share on other sites
45 minutes ago, JASCoder said:

That is to say, MGI's bank account (in the US) is both receiving fiat deposits, and making local electronic transfers to their Bitstamp account. I think we're in sync up to here.

Oh I missed something.  No I don’t think MGIs bank account is making local electronic transfers to Bitstamp.  That is handled inside Ripplenet by the ODL software.  All that MGI do is initiate an ODL transfer...  the software does the rest.  
 

So it’s not a MGI bank transfer...  it’s a Ripplenet action.  The result is the same but the mechanism is not...  it’s not a ‘local electronic transfer to Bitstamp account’.

Perhaps you already agreed with that and just worded it differently than I would have.

Share this post


Link to post
Share on other sites

×
×
  • Create New...