Forums  > Basics  > Explaining market making  
     
Page 1 of 1
Display using:  

Jurassic


Total Posts: 415
Joined: Mar 2018
 
Posted: 2021-04-25 21:52
I just cannot understand the strategy behind market making.

Ive tried reading things on internet forums (wso/quora/reddit) but it isnt helping. Online it says that traders calculate a fair value for the asset. But how do you do this in reality? Then they apply a spread, but how do you know how wide to quote?

Ive read a number of sources about JS interview questions regarding estimating a quantity and then giving a confidence window. Im guessing that this is meant to simulate market making in some way. This is all obvious stuff in superficial situations like for estimating the number of gas stations in the states, but I dont see how you translate this into a live financial asset. In the equivalent situation in a financial market, the number of gas stations would vary in time, how does a market maker know whether to move the fair value higher or lower? Surely after a move higher or lower and/or an event where a transaction takes place, shouldnt the bid ask natually get skewed? Im not sure how these exercises relate to the job.

What is the general strategy on market making? Maybe I also use the wrong terminology and dont know the difference between flow trading and market making. Im not looking for academic papers on the subject because they dont seem to correlate to what a human market maker in a bank could possibly think through.

tabris


Total Posts: 1287
Joined: Feb 2005
 
Posted: 2021-04-26 05:20
I am not sure what you are reading and I think you are confusing yourself with the technicalities. The way I would define the strategy for market making is someone that is warehousing risk for a price. Flow trading and market making definition is probably a bit more nuanced and for all practical purpose you should just think they are similar. Sure, there are ways to think in terms of what you should charge to warehouse risk, like fair value and spreads but all you should be doing is think in terms of being properly compensated to "store" this risky asset between the time I take in a position to the time when it is out of my inventory.

Dilbert: Why does it seem as though I am the only honest guy on earth? Dogbert: Your type tends not to reproduce.

Jurassic


Total Posts: 415
Joined: Mar 2018
 
Posted: 2021-04-26 08:51
I get that but I just cant think how this works in actuality.

ronin


Total Posts: 689
Joined: May 2006
 
Posted: 2021-04-26 09:27
> Online it says that traders calculate a fair value for the asset. But how do you do this in reality?

I wouldn't necessarily worry about what it says online. But you are on the right track.

Only, it's not 'fair value'. It's more like 'price I can get out at'.

There are different types of market making, but that is the basic component of all of them.

Say plain vanilla high frequency market making. You need your order to be first in the queue at a given price level, and you need sufficient volume behind you in the queue. So when you get filled, you can turn around and scratch to the next guy behind you. The trading logic is something along the lines "when you get filled, wait for the opposite fill as long as there is plenty of volume to scratch. If the scratch volume starts disappearing, scratch and move on".

Or, say, pairs market making. You are market making some illiquid etf, but there is also a liquid future with the same exposure. So you are happy to get filled on the etf at a wide spread, then pay the smaller spread on the future to get out of the delta. You are still carrying some basis risk between the etf and the future, and hopefully you are making enough to cover it. And that comes in millions of flavours - single stocks, single stock futures, index etfs, index futures, proxies like commodities or bonds or fx, and then any combination of any of those against any other combination of any of those.

Or, say, options market making. You will never get flat on any single given option contract, but you can aggregate your deltas, gammas, vegas, skew deltas, smile deltas and what ever else you need to worry about, and hedge them across the surface.

And so on.

"There is a SIX am?" -- Arthur

Jurassic


Total Posts: 415
Joined: Mar 2018
 
Posted: 2021-04-26 13:50
The only thing I understand there totally is plain vanilla hft market making. However, in OTC markets like govt bond trading we dont what the order book looks like. I dont see how you can quantity for pairs market making how much you need to compensated for the basis risk?
For things like options market making, its obvious to me as to how to flatten all your greeks, but I cant see what the spread should be for doing this. I suspect options market making is just like bond market making but with another dimension.

If you have a 10y bond that just traded at $1000, you can quote 995x1005. How does this price change?

EspressoLover


Total Posts: 490
Joined: Jan 2015
 
Posted: 2021-04-26 16:29
You're going to think I'm being sarcastic. But honestly, if you want to grok market making on an intuitive level, watch some of those pawn shop shows on the History Channel.

Good questions outrank easy answers. -Paul Samuelson

tradeking


Total Posts: 32
Joined: May 2016
 
Posted: 2021-04-26 17:36
Markets are competitive. If you mark up prices too much (i.e. quote too wide a spread), then there will be few takers at your prices. Not to mention there could be other market makers competing for the takers if it's an open market. If you you mark up too little, the markup would not be enough to make the trade worth the risk (i.e. your sharpe suffers). The most direct way to gauge if you are quoting too wide or tight is to see the feedback from the market and adjust your quotes accordingly. Experience and having a feel for the market helps too.

Jurassic


Total Posts: 415
Joined: Mar 2018
 
Posted: 2021-04-26 19:27
@EspressoLover et al. I already understand market making on a high level, its just the details Im struggling wiht.

Ok so if you arent getting making trades you narrow your spread and vice versa. But what makes you change your mid? How do you know you wont just sell into a rising market and vice versa?

sloppy


Total Posts: 9
Joined: May 2011
 
Posted: 2021-04-26 19:57
If you are smart about mid, you are not a market-maker; you are a spread-crossing prop trader.

You don't need to know where mid is going to be a market-maker, that's the whole point. You only need to have some notion of fair within the spread.

Jurassic


Total Posts: 415
Joined: Mar 2018
 
Posted: 2021-04-26 20:30
@sloppy that was going to be my next if you always know where mid why not just set up a fund

So if you have a notion a fair range, how can you come up with that?

ronin


Total Posts: 689
Joined: May 2006
 
Posted: 2021-04-26 20:48
> in OTC markets like govt bond trading we dont what the order book looks like. I dont see how you can quantity for pairs market making how much you need to compensated for the basis risk?
For things like options market making, its obvious to me as to how to flatten all your greeks, but I cant see what the spread should be for doing this. I suspect options market making is just like bond market making but with another dimension.


@jurassic,

Honestly, I think you need to organise your own thoughts a little bit. I have no idea what this paragraph above means. You don't understand bonds, but you do understand options, because options are just like bonds with an extra dimension. But you don't understand bonds. And you do understand options.

Seriously, what?

You have all the basic ingredients. The principle is pretty simple. You make a bit of money on each trade, and you get a bit of risk on each trade. The money you make adds up, and the risk diversifies away. Bits that don't diversify, you give back.

The details of what you have to make per trade, how things diversify etc depend on so many different things. Your universe, your capital, your risk limits, your risk tolerance, etc.

> How do you know you wont just sell into a rising market and vice versa?

Well, you don't. But if you are quoting a 1,000 symbols, you expect to be getting filled on roughly half the bids and half the asks. If you see you are filled on 800 asks and only 200 bids, you are net short 600 and the whole market is moving up. So you buy, say, the index future for the size equivalent to your 600 lots to flatten things out. You made 1,000 half spreads (200+800 where you got passive fills), paid back 600 on the future. Net, you made 400 half spreads. Maybe you lost some on the initial move, so maybe you actually made only 100.

And then you have some idea of how that works in your particular market, like how long to get a fill etc, and based on that you do some sums and comit some capital.

"There is a SIX am?" -- Arthur

tabris


Total Posts: 1287
Joined: Feb 2005
 
Posted: 2021-04-26 23:14
OTC markets works the same way earlier as I described. It doesn't matter where market last traded or where mid is. You will make your market based on current inventory, future predicted flows, and time spent in inventory. If market traded at 1000 and I think some fund is going to come in later in 15 minutes and even buy at 1005, you would be willing to move your bid from 995 to 996 997 998 999 and even 1000 because you are willing to risk 5 in a span of 15 minutes. Sometimes OTC markets depth and liquidity are hidden and you just have to work with hedging apple prices with oranges and bananas to figure out both the liquidity and depth.

Dilbert: Why does it seem as though I am the only honest guy on earth? Dogbert: Your type tends not to reproduce.

Jurassic


Total Posts: 415
Joined: Mar 2018
 
Posted: 2021-04-27 07:37
@ronin well my line of thought was that the spread in certain markets is function of the cost of the basis risk between the product traded and the related hedging instrument. My thought was that for bonds you can hedge along the yield curve (1d) and for options you can hedge on the surface (2d).

Ok I get that when your bids and asks from the customers are imbalanced you could get with an index future products. However, is there always an index future product which hedges it? Also wouldnt it be easier in that to start moving your bids and offers in some direction?

I also would have thought how long to get a fill would be pretty lumpy and maybe completely different on some months to the next.

@tabris im not sure how you can predict whether someone is going to buy a certain product in 15 mins time?

"you just have to work with hedging apple prices with oranges and bananas to figure out both the liquidity and depth"

how does hedging help with correlated products help you figure out the liquidity and depth?

ronin


Total Posts: 689
Joined: May 2006
 
Posted: 2021-04-27 09:50
> However, is there always an index future product which hedges it?

No. You could just give back 600 single stocks, picked based on what ever logic makes sense.

> Also wouldnt it be easier in that to start moving your bids and offers in some direction?

Yes, that is how execution logic generally works. If you have plenty of time, quote wide and hope for the best. If you have less time, quote tighter. When time runs out, cross. Whether you think of that as a bias on your market making or a superimposed flattening tactic is up to you.

> I also would have thought how long to get a fill would be pretty lumpy and maybe completely different on some months to the next.

It depends on what you are trading. For most things, seasonality is mostly intraday. Check the other thread about intraday vol and vwap. I'm sure heating oil and stuff like that has a yearly seasonal component as well, but that's not really my universe.

Everything is lumpy. That's just life. Generally, the bigger the book, the less lumpy it gets.

"There is a SIX am?" -- Arthur

Jurassic


Total Posts: 415
Joined: Mar 2018
 
Posted: 2021-04-27 10:23
"You could just give back 600 single stocks"

But why wouldnt you want to do that? Does it cost anything to you to give them up?


"Whether you think of that as a bias on your market making"

Im not really sure what this means?

It says somewhere on this forum that an (MBA) could be trained up in a month to be a flow trader, https://www.nuclearphynance.com/Show%20Post.aspx?PostIDKey=187911, which implies this job isnt that hard, but all I can see is potential strategy difficulties

doomanx


Total Posts: 115
Joined: Jul 2018
 
Posted: 2021-04-27 10:50
'Does it cost anything to you to give them up'
Single stock liquidity sucks. Index ETFs are very liquid.

'Im not really sure what this means?'
It's just a comment saying there are different lines of thinking that lead you to the same basic principle of adjusting your quotes.

There is not a fixed mapping of qualifications -> whether you can do xyz job or not. People have different backgrounds. There are some fundamental concepts that don't require any particular skillset vital to success in market making (or most jobs), which imo ronin has done an excellent job explaining here. If you get them, more or less technical background will likely change the nature of your activity but the point is the same - you understand the economics behind why making a market can generate revenue and you are able to come up with good ideas about how you might capture that profit.

did you use VWAP or triple-reinforced GAN execution?

ronin


Total Posts: 689
Joined: May 2006
 
Posted: 2021-04-27 13:16
> But why wouldnt you want to do that? Does it cost anything to you to give them up?

Because, in that scenario, you got 600 undiversified shorts. It does cost - that's the 600 you are giving back. You made 1,000, gave up 600, net you made 400. And your logic may well be "give back the 600 that are cheapest to give back", or something like that. Or it might be something else.

> (MBA) could be trained up in a month to be a flow trader,

Man, did those stereotypes about fx traders just disappear? Like tears in rain?
Suddenly I feel old...


"There is a SIX am?" -- Arthur

tabris


Total Posts: 1287
Joined: Feb 2005
 
Posted: 2021-04-27 23:12
"how does hedging help with correlated products help you figure out the liquidity and depth?"

I think honestly you have zero experience in working in the markets and then you are trying to figure out all the details/nuance/technicalities of something that might or might not be relevant to you. These information will become more glaring when you actually do the job...

But anyway lets go over an example in the bond market... Imagine back in the pre Lehman days where you don't have an electronic platform to see where some 12year bond is trading and all you do is ask a broker. You pick up your phone, ask for a market/size, and that is all the information you will get (theoretically anyway). But, you do know that there is a futures market for 10year bond so you can back out some type of possible market with liquidity and depth using the stack on the futures.

Ok, then you would ask some basic question like but what if theres no futures, then I would say it works the same on other OTC products. You can collect all the bonds you have information on in the broker... with spread and depth and extrapolate. So previous example you can go back to the broker and ask for 11yr bond, 10yr bond, 9yr bond, 13yr bond, 14yr bond, 15y bond... etc etc...


Dilbert: Why does it seem as though I am the only honest guy on earth? Dogbert: Your type tends not to reproduce.

tabris


Total Posts: 1287
Joined: Feb 2005
 
Posted: 2021-04-27 23:14
"Man, did those stereotypes about fx traders just disappear? Like tears in rain?
Suddenly I feel old..."

I still remember when FDAXHunter used to describe a group of FX traders as a den of thieves... funny

Dilbert: Why does it seem as though I am the only honest guy on earth? Dogbert: Your type tends not to reproduce.

prikolno


Total Posts: 94
Joined: Jul 2018
 
Posted: 2021-04-27 23:14
> If you are smart about mid, you are not a market-maker; you are a spread-crossing prop trader.

If I follow what you're saying, I disagree with this statement.

It's hard to be a good market maker these days without being "smart about mid".

Now you don't have to follow my approach but I usually like to think of how to build alphas for a taking strategy first, because that makes it easier thereafter to build a market making strategy.

Maggette


Total Posts: 1325
Joined: Jun 2007
 
Posted: 2021-04-28 07:50
Well they don't strike me as a bunch of monkeys
https://www.xtxmarkets.com/

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

ronin


Total Posts: 689
Joined: May 2006
 
Posted: 2021-04-28 13:57
The stereotype was about driving trucks for a living.

And that guy on their website looks like he's just delivered a washing machine...

"There is a SIX am?" -- Arthur

EspressoLover


Total Posts: 490
Joined: Jan 2015
 
Posted: 2021-04-29 19:12
A market maker is just a business. A business which sells the product of liquidity. Like pricing any product, there's more than one way to skin a cat. A slick-talking used car salesman will use gut feeling and intuition to set prices. Carvana will prices its inventory based off an algorithm built by PhD data scientists. There's room enough in the market that both approaches can co-exist in a stable ecosystem.

Good questions outrank easy answers. -Paul Samuelson

Jurassic


Total Posts: 415
Joined: Mar 2018
 
Posted: 2021-04-30 13:00
I still only get the overview of this business. I just cant see what human market makers are analysing when moving their spread and changing the width of the spreads

Maggette


Total Posts: 1325
Joined: Jun 2007
 
Posted: 2021-04-30 13:57
"And that guy on their website looks like he's just delivered a washing machine..."

LOL!

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...
Previous Thread :: Next Thread 
Page 1 of 1