SERVING THE QUANTITATIVE FINANCE COMMUNITY

 
User avatar
trackstar
Posts: 26820
Joined: August 28th, 2008, 1:53 pm

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 3:57 pm

A contribution to the original post:

Here are a few things to consider in the lifecycle of start-ups and tech innovation. It's not a argument for or against the use of MSMs, but just to keep you grounded in the every day realities.

Start-up Idea and Team - business plan, articulation of business idea, tech description, market size, estimates of financial needs, pitches to various VCs (who are not all created equal).

If the company gets funded:

- Tech Dev
- Hiring
- Milestones (not just the end target, as you note, but many steps along the way, with regular reporting to the VCs and the Board (which at least one VC rep from each VC firm that has invested will sit on)
- Hits milestones, needs more money - new capital raise
- Doesn't hit milestones, but still high potential and viewed favorably - new capital raise
- Doesn't hit milestones - trouble all around - tough capital raise or maybe death valley...
- Beta testers, early adopters, champions
- Eventual launch and exit (IPO, trade sale, or bring in strategic partners who then back the firm themselves)

What else can happen:

- Founder/co-founder disputes and sometimes exits - forced or voluntary - this is disruptive and replacement of key personnel takes time and brings new risks
- Challenges in hiring enough talent quickly - competition is fierce
- Team issues - culture and integration - scaling at human level
- Natural challenges in tech dev on a given timescale in any environment - knowns, known unknowns, unknown unknowns, etc.
- Issues with suppliers - any weak spots, like shortages in a critical material?
- Competitive landscape - changes, new entrants, new product launches, new tech directions to take in response
- Market shifts or changes - for the positive or the negative w/r/t this company's product/service offering
- Follow on funding - same VC or same and new investors too? Dilution issues, and board expansion - can be smooth or can become a rocky road
- Exit time: Public markets - is the IPO window open, hot, cooling, closed?
- For acquisitions - is it a vigorous competitive market for acq. in this space, or in a lull? Who else has been acquired recently? Is there a healthy appetite or a glut already ?
Other factors - surprise obsolescence, public policy or regulatory issues that emerge, failure at certain trials (in pharma e.g., FDA clinical trials - Phases I, II, III).

And more... 

How much is quantitative and how much is qualitative and how to address that in an MSM? Cool problem to study and seek to solve.

To help in your research, I'd recommend reading some start-up cases studies and some of the better VC literature (check SSRN).

Not popularized accounts, but more staid, fact-driven narratives.

Very interesting topic - good luck!
Last edited by trackstar on November 17th, 2019, 4:07 pm, edited 2 times in total.
 
User avatar
katastrofa
Posts: 8518
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 4:03 pm

Is the US or Australia health care country budget enough, Cuchulainn?
Not if Trump can help it; in fairness, you mentioned social sciences, so I had to start somewhere.
I would prefer more beds in hospitals.
Social science experts are commonly employed in the government and business projects. I prefer personalised medicine to more death beds. In the case of Australia everybody gets an equal chance to get the hospital bed (at least in theory).
BTW, Sycamore is Google's invention.

Anyway, as you'd put it, your confused rambling is off-topic and not adding any quality to the discussion.
Last edited by katastrofa on November 17th, 2019, 4:31 pm, edited 2 times in total.
 
User avatar
trackstar
Posts: 26820
Joined: August 28th, 2008, 1:53 pm

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 4:15 pm

It's moving, but only 1 minute QPU time/month.
I'll need a bigger boat to do a social network analysis of this site, down the years.
dwave-ocean-sdk OK 
Remark: It would be nice to run on PC before burning my monthly qbit allowance!
LOL. Looks like it is Optimization time then.  Write in C?

In the early 2000s, the question was how to optimize over distributed network of (not so many nodes).

Now easily have hundreds of thousands of *cores* humming along...

On the Top 500 list, #1 is Summit by IBM - over 2.4 million cores.

Top 500 List - June 2019

data from chart:
Cores
Rmax (in TFlops/s)
Rpeak (in TFlops/s)
Power (kW)

[table][tr][td]Summit - IBM Power System AC922, IBM POWER9 22C 3.07GHz, NVIDIA Volta GV100, Dual-rail Mellanox EDR Infiniband , IBM 
DOE/SC/Oak Ridge National Laboratory
United States[/td]
[td]2,414,592[/td]
[td]148,600.0[/td]
[td]200,794.9[/td]
[td]10,096[/td]
[/tr]
[/table]
It makes me happy to read this list. : )

BTW - #3 on the list Sunway TaihuLight has over 10 million cores, but you can see that significantly more is not necessarily the Best.  
Communications overhead may be a lurking monster...
 
User avatar
trackstar
Posts: 26820
Joined: August 28th, 2008, 1:53 pm

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 4:42 pm

BTW, Sycamore is Google's invention.
Yes - thanks - just realized I had not finished my post about that - have now added an article and attribution to Google for Sycamore.

Multitasking here, with a very insistent but happy cat on my lap. : )
 
User avatar
katastrofa
Posts: 8518
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 5:02 pm

BTW, Sycamore is Google's invention.
Yes - thanks - just realized I had not finished my post about that - have now added an article and attribution to Google for Sycamore.

Multitasking here, with a very insistent but happy cat on my lap. : )
Aww, kiss him from me :-)
 
User avatar
trackstar
Posts: 26820
Joined: August 28th, 2008, 1:53 pm

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 5:09 pm

BTW, Sycamore is Google's invention.
Yes - thanks - just realized I had not finished my post about that - have now added an article and attribution to Google for Sycamore.

Multitasking here, with a very insistent but happy cat on my lap. : )
Aww, kiss him from me :-)
I will!  Smokey is a girl, but she accepts kisses graciously...

You will see further additions to the Google space in previous post.  

Quantum fairness and balance in posting on W forum. ; )
Last edited by trackstar on November 17th, 2019, 11:07 pm, edited 1 time in total.
 
User avatar
katastrofa
Posts: 8518
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 6:20 pm

Oh, I know Smokey is a girl! I just thought you were cheating on her with another cat since you didn't mention her by her name :-) (Careful with stray cats - I haven't been able to get rid of folliculitis from my chin and belly for a month after I got licked!)
 
User avatar
Alan
Posts: 9867
Joined: December 19th, 2001, 4:01 am
Location: California
Contact:

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 7:57 pm

There is also the popular terminology surrounding burn rates, etc. He makes no mention of this, but it is sort of inherently implied. For example, you receive some capital, and there is some future date at which you must complete your innovation otherwise you will have "burned through" all of your capital, at which point you are ready to go to market. What is the probability of hitting that target spot on or exceeding it at any point in time? In theory there is a distribution of potential outcomes revolving around that target date.


Before we cross the inescapable black cat event horizon, a few comments on the original question and `burn rates'.

First, as other have pointed out, apparently MSM's are just different language for what are traditionally called "Continuous-time Markov Chains".

Second, are they useful for venture finance?

 I think if you turned it around, and asked: does venture finance suggest interesting problems for a course in stochastic processes, the answer would be yes. As an example, and related to your "burn rate" issue, I had some discussion in a recent book on the Stefan problem, which is a classic physics problem related to the water-ice phase transition. I made some online remarks here. As it turns out, I was also able to relate this problem to the venture capital burn-rate problem: 
LewisVol2.StefanExample.pdf
(50.18 KiB) Downloaded 9 times
 
But, could a venture firm improve it's performance by adopting a formal stochastic process model for some aspect of what they do? I suppose anything is possible. But generally, I'm with trackstar in suggesting that (beyond elementary stuff like diversification), most of the issues are qualitative. 
 
User avatar
Cuchulainn
Posts: 60501
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 9:35 pm

First, as other have pointed out, apparently MSM's are just different language for what are traditionally called "Continuous-time Markov Chains".
Indeed. I looked very quickly but that is what I would also have said.

So, it is a special case, then all the already found results can be applied. No need to concoct a new name (what ever happened the existing literature to see what has come before?)

What about modelling CTMC using matrix ODE?
Last edited by Cuchulainn on November 17th, 2019, 10:12 pm, edited 1 time in total.
http://www.datasimfinancial.com
http://www.datasim.nl

Approach your problem from the right end and begin with the answers. Then one day, perhaps you will find the final question..
R. van Gulik
 
User avatar
Cuchulainn
Posts: 60501
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 10:00 pm

Oh, I know Smokey is a girl! I just thought you were cheating on her with another cat since you didn't mention her by her name :-) (Careful with stray cats - I haven't been able to get rid of folliculitis from my chin and belly for a month after I got licked!)
How many times have you typed the word 'cat' here?

274, 192 of which originated in OT.

I have 159, but 95% are replies to kats' posts which as we all know by now contains copious allusions to felines.

trackstar has some catching up to do, a measly 205.
Last edited by Cuchulainn on November 17th, 2019, 10:38 pm, edited 3 times in total.
http://www.datasimfinancial.com
http://www.datasim.nl

Approach your problem from the right end and begin with the answers. Then one day, perhaps you will find the final question..
R. van Gulik
 
User avatar
katastrofa
Posts: 8518
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 10:27 pm

@Alan "First, as other have pointed out, apparently MSM's are just different language for what are traditionally called "Continuous-time Markov Chains"."

Multi-state models mentioned by the OP don't need to be Markov or continuous in time. I wrote that that was the case in medicine (addressed by OP), where they are indeed usually i) continuous in time ii) Markov - and iii) hidden/hidden Markov - because we are observing the real (markov) process - the disease progression, via another (Markov) process - the medical observation. IMHO when talking about hidden states, it is convenient to differentiate between the model and the process/chain, and in medicine the distinction is material.
Though, if you disagree, I won't scratch and bite =^o.o^=
Last edited by katastrofa on November 17th, 2019, 11:03 pm, edited 2 times in total.
 
User avatar
katastrofa
Posts: 8518
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 17th, 2019, 10:55 pm

Oh, I know Smokey is a girl! I just thought you were cheating on her with another cat since you didn't mention her by her name :-) (Careful with stray cats - I haven't been able to get rid of folliculitis from my chin and belly for a month after I got licked!)
How many times have you typed the word 'cat' here?

274, 192 of which originated in OT.

I have 159, but 95% are replies to kats' posts which as we all know by now contains copious allusions to felines.

trackstar has some catching up to do, a measly 205.
In my confused delusioned orations i make copious allusions to felines.
Attachments
C0AFCFB0-0729-45A2-8BA5-7A01A762FB16.jpeg
 
User avatar
Alan
Posts: 9867
Joined: December 19th, 2001, 4:01 am
Location: California
Contact:

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 18th, 2019, 3:22 pm

@Alan "First, as other have pointed out, apparently MSM's are just different language for what are traditionally called "Continuous-time Markov Chains"."

Multi-state models mentioned by the OP don't need to be Markov or continuous in time. I wrote that that was the case in medicine (addressed by OP), where they are indeed usually i) continuous in time ii) Markov - and iii) hidden/hidden Markov - because we are observing the real (markov) process - the disease progression, via another (Markov) process - the medical observation. IMHO when talking about hidden states, it is convenient to differentiate between the model and the process/chain, and in medicine the distinction is material.
Though, if you disagree, I won't scratch and bite =^o.o^=

Fair enough. So apparently then the term encompasses any stochastic process [$]J_t[$] on a finite number of states [$]J \in \{1,2,\cdots,N\}[$].  
 
User avatar
Cuchulainn
Posts: 60501
Joined: July 16th, 2004, 7:38 am
Location: Amsterdam
Contact:

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 18th, 2019, 4:54 pm

We use Latex and Word. We had to install Aurora as addin. Then you can import Latex into Word document.

// oops, wrong thread, reply to kats.
http://www.datasimfinancial.com
http://www.datasim.nl

Approach your problem from the right end and begin with the answers. Then one day, perhaps you will find the final question..
R. van Gulik
 
User avatar
katastrofa
Posts: 8518
Joined: August 16th, 2007, 5:36 am
Location: Alpha Centauri

Re: Multi-state probabilities/models (MSM) and regression applied to venture capital and innovation

November 18th, 2019, 8:23 pm

We use Latex and Word. We had to install Aurora as addin. Then you can import Latex into Word document.

// oops, wrong thread, reply to kats.
Thanks. In the case of WM it turned out to be a funny miscommunication. I'm sorry if that sounded as if I questioned their common sense :-)
ABOUT WILMOTT

PW by JB

Wilmott.com has been "Serving the Quantitative Finance Community" since 2001. Continued...


Twitter LinkedIn Instagram

JOBS BOARD

JOBS BOARD

Looking for a quant job, risk, algo trading,...? Browse jobs here...


GZIP: On