Why we need MM program

[tulo]

I may be biased as MM but I explain to you my MM experience.

My little bots are moving around 2-300.000 USD equivalent in 24h. Not a big investment but still...

Out of this movements I'm paying every day around 400-600 USD as transfer fee! It makes 12k USD I'm paying to gateways every month! And this for a very little income in comparison, even going negative some time. It's a while I'm planning to stop the bots because gateways are eating all the income, reducing Ripple capitalization by 150.000 USD equivalent each month (in two years now we'd have 3.6M more value issued on Ripple). Of course the gateways have to gain with their work and IOUs and the transfer fee is fair, but by introducing MM incentive, it would be possible to reintroduce that capital into Ripple.

But again I'm biased

[Morty]

Will you be using your bots on BitStamp next week to take advantage of the 0 fees? 

[tulo]

3 minutes ago, Morty said:

Will you be using your bots on BitStamp next week to take advantage of the 0 fees? 

My bots are on Ripple, not Bitstamp (well, I have other bots for bitstamp). The 0 fees are only on the private markets USD-XRP and EUR-XRP of Bitstamp.

[Graine]

There's nothing 'tiny' about moving 200k of value a day. And for 12k/month of semi-passive income you can live in any part of the world happily. That said, I hope Ripple will offer you a solution. Though maybe gateways themselves would be more approachable. 

[Professor Hantzen]

1 hour ago, tulo said:

I may be biased as MM but I explain to you my MM experience.

My little bots are moving around 2-300.000 USD equivalent in 24h. Not a big investment but still...

Out of this movements I'm paying every day around 400-600 USD as transfer fee! It makes 12k USD I'm paying to gateways every month! And this for a very little income in comparison, even going negative some time. It's a while I'm planning to stop the bots because gateways are eating all the income, reducing Ripple capitalization by 150.000 USD equivalent each month (in two years now we'd have 3.6M more value issued on Ripple). Of course the gateways have to gain with their work and IOUs and the transfer fee is fair, but by introducing MM incentive, it would be possible to reintroduce that capital into Ripple.

But again I'm biased

I'm curious about how your bots work.  I thought about automating market-making for a while, but a couple of problems I had was how to make the bots unpredictable enough that people couldn't use their behaviour to trade against, and also what to do to manage the situation of being left "carrying the can" at the wrong end of large price swings.

Maybe something could be done to hold some gateway fees in escrow, and then distribute these to the MM's responsible?  I mean when it's been well-established (somehow...) that those accounts were indeed MMing.  Some metric such as sustained high volume trading roughly equally on both sides of a given market for a period of a month or something, could then qualify for a share of the gateway fees collected on those assets for that month?

I guess gateways might not be too enthusiastic about the idea, but it could potentially increase revenue for them, by causing market making to be more attractive to people, and the resultant greater liquidity may encourage adoption of the ecosystem as a whole which would obviously be good also.

Edited January 12, 2017 by Professor Hantzen

[tulo]

7 minutes ago, Professor Hantzen said:

I'm curious about how your bots work.  I thought about automating market-making for a while, but a couple of problems I had was how to make the bots unpredictable enough that people couldn't use their behaviour to trade against, and also what to do to manage the situation of being left "carrying the can" at the wrong end of large price swings.

Making the bots unpredictable is useful only if you operate them in illiquid markets, and that's my main motivation of not releasing the code. Another way is to make them "almost" independent from other user offers, so as to consider only funded offers and with a volume high enough to bother you, so that the "attacker" should put some money into it and take a risk that is not worth.

The "carrying the can" problem is real and almost impossible to avoid. That is the main motivation of losses.

11 minutes ago, Professor Hantzen said:

Maybe something could be done to hold some gateway fee's in escrow, and then distribute these to the MM's responsible?  I mean when it's been well-established (somehow...) that those accounts were indeed MMing.  Some metric such as sustained high volume trading roughly equally on both sides of a given market for a period of a month or something, could then qualify for a share of the gateway fees collected on those assets for that month?

To see MM accounts is very easy. Just check https://charts.ripple.com/#/active_accounts on 7 days base, and the accounts close to 50/50 buy-sell are most likely MMs. But there are more robust way to check.

[warpaul]

Hi Tulo,

We truly appreciate the effort you've put into market making on RCL - it doesn't go unnoticed!

RCL is definitely a unique trading environment, so I tip my hat to anyone with the knowledge and skill to profitably add liquidity to the market.

You may have noticed that we have been testing a market making program on RCL for the last several months and we have learned a lot in the process.

We're about to deploy a new feature on RCL called TickSize, which will help to eliminate bots fighting over the tip of the order book (a problem we routinely encountered while testing the MM program).

Larger tick sizes have also been known to benefit market makers by:

• Reducing spreads
• Promoting liquidity
• Increasing price discovery 
• Decreasing transaction spam
• Decreasing ledger churn and metadata sizes
• Bringing more scalability and transparency to order books

These are all benefits that will help improve our efforts to expand an institutional market making program to a wider audience.

Hope that helps to clarify things a bit more  

[tulo]

9 hours ago, warpaul said:

Hi Tulo,

We truly appreciate the effort you've put into market making on RCL - it doesn't go unnoticed!

RCL is definitely a unique trading environment, so I tip my hat to anyone with the knowledge and skill to profitably add liquidity to the market.

You may have noticed that we have been testing a market making program on RCL for the last several months and we have learned a lot in the process.

We're about to deploy a new feature on RCL called TickSize, which will help to eliminate bots fighting over the tip of the order book (a problem we routinely encountered while testing the MM program).

Larger tick sizes have also been known to benefit market makers by:

• Reducing spreads
• Promoting liquidity
• Increasing price discovery 
• Decreasing transaction spam
• Decreasing ledger churn and metadata sizes
• Bringing more scalability and transparency to order books

These are all benefits that will help improve our efforts to expand an institutional market making program to a wider audience.

Hope that helps to clarify things a bit more  

Sorry, I don't get it. Did you try a market making program intended as a bot that creates offer as market maker or did you experiment for the Market Making incentive program?
 

BTW I like the TickSize feature (I still have some doubts about it) but that sincerely doesn't help at all the market makers ... it only helps the ledger and its storage (which is a good thing).

[fh13]

On 12.1.2017 at 10:24 PM, tulo said:

To see MM accounts is very easy. Just check https://charts.ripple.com/#/active_accounts on 7 days base, and the accounts close to 50/50 buy-sell are most likely MMs. But there are more robust way to check.

Wow I'm quite impressed. I checked out a few account and one had only 1000 XRP and 1200 USD. Still he was able to have a 7 day trade volume of about $70k, which translates to a profit of about $100 considering the average spread.