Small prop trading firms: Are they worth it?

[physguy]

Hi everyone,

I got an offer from a small and relatively unknown prop trading firm/market maker which also does algorithmic trading. The role is for trader. This is the normal trader, not so much quant stuff. Things like having 9 monitors in front of you, following the news and changing parameters in the algorithms live. Also working morning or night shifts to cover different markets.

The problem is that it is a small company. One office, around 50 people. The salary is also nowhere close to the great numbers from the big companies, could be around 50% of what companies like Optiver, SIG etc. are paying for the GRADUATE programs. This is in Europe btw.

My questions is : is it worth it to take the offer so I can make a move to something bigger in 2-3 years? As far as I understand, the big companies will never hire from such a place, and I guess that the big banks don't really hire from small places either? So what are in general the exit opps from a trading firm like this?

I work in quant risk now, and I am not sure whether the move is worth it.

 

[achirikhin]

>>changing parameters in the algorithms live.

Run away

 

[physguy]

>>changing parameters in the algorithms live.

Run away

Hi, thanks for the response. Could you please elaborate a bit? I did not really understand your comment

 

[achirikhin]

Hi, thanks for response. Could you please elaborate a bit?

Channing parameters of algos "live" is a very very very poor (model) risk management.

You should change them very infrequently, ideally (in the limit) never.

If somebody in the firm told you they do that, it is doubly dumb. Firstly doing so. Secondly, revealing it.

 

[physguy]

Channing parameters of algos "live" is a very very very poor (model) risk management.

You should change them very infrequently, ideally (in the limit) never.

If somebody in the firm told you they do that, it is doubly dumb. Firstly doing so. Secondly, revealing it.

Ok, I see what you mean. Maybe I also didn't explain in detail. What they told me is this: The algorithms are running, but in general something may happen that needs to be taken into account suddenly. For example you are at the night shift and news are coming that Covid pandemic broke all over the world. Then you need to stop the algorithm manually from running. Apart from that, the role had also "discretionary alpha generation". For example maybe you see that every Friday at 10 o' clock there is some inefficiency in the market etc., and they benefit from that.

Overall, this is what they explained to me. I don't know whether this changes anything.

 

[achirikhin]

Ok, I see what you mean. Maybe I also didn't explain in detail. What they told me is this: The algorithms are running, but in general something may happen that needs to be taken into account suddenly. For example you are at the night shift and news are coming that Covid pandemic broke all over the world. Then you need to stop the algorithm manually from running. Apart from that, the role had also "discretionary alpha generation". For example maybe you see that every Friday at 10 o' clock there is some inefficiency in the market etc., and they benefit from that.

Overall, this is what they explained to me. I don't know whether this changes anything.

No, it only confirms the suspicion. There should be automatic circuit breakers for such situations. A human, at most, should have an opportunity to override them from being triggered.

Very weird.

 

[physguy]

No, it only confirms the suspicion. There should be automatic circuit breakers for such situations. A human, at most, should have an opportunity to override them from being triggered.

Very weird.

Ok, thanks a lot for your time!

 

[QuantCoach]

Especially if you’re talking about the OMM space, manually adjusting parameters is not a red flag… it’s one of the primary jobs of traders even in the big prop shops you mentioned (Optiver, SIG, etc.). The algos are much less systematic than you think and there’s a lot more discretion involved, it’s not weird at all.

 

[physguy]

Especially if you’re talking about the OMM space, manually adjusting parameters is not a red flag… it’s one of the primary jobs of traders even in the big prop shops you mentioned (Optiver, SIG, etc.). The algos are much less systematic than you think and there’s a lot more discretion involved, it’s not weird at all.

They definitely do market making, mainly in ETPs I think

 

[achirikhin]

Especially if you’re talking about the OMM space, manually adjusting parameters is not a red flag… it’s one of the primary jobs of traders even in the big prop shops you mentioned (Optiver, SIG, etc.). The algos are much less systematic than you think and there’s a lot more discretion involved, it’s not weird at all.

It is a red flag to me, having dealt with some "trading firms" that do that. I am not here to prove anything to anyone, but one cannot be half-pregnant.

 

[QuantCoach]

All I'm saying is that from personal experience, many quant traders at Optiver, SIG and JS (at least their options team) manually move the system settings (width/sizing/etc.) and pricing params multiple times within a day. As OP mentioned, we especially need to do so in times like Covid; even OMMs with more systematic approaches like CitSec have Peng saying the systems are calibrated well to operate during standard situations and one of the primary responsibilities of the trader is to step in when things are not so normal. Maybe it's different for banks and other asset classes but it's not so wise to rule out prop shops solely because they change parameters live, otherwise you're ruling out most of the largest OMMs.

 

[achirikhin]

Well, I assume those firms know what they are doing, as they are not out of business. However in a smaller and more of a start up firm such behaviour can be a sign of possibly excessive and not systematic risk taking. This is what I observed.

Nobody disputes that traders need to step in from time to time, but the question is what amount of human interaction still allows to consider operation algotrading or not. You don't need to be algo to be systematic, of course.