SERVING THE QUANTITATIVE FINANCE COMMUNITY

Search found 241 matches

  • 1
  • 2
  • 3
  • 4
  • 5
  • 17
by willsmith
November 23rd, 2015, 9:55 pm
Forum: General Forum
Topic: Correlations during crashes
Replies: 1
Views: 2346

Correlations during crashes

We all know the theory : correlations between asset returns in crash scenarios tend to be higher than during uninteresting times. Can anybody suggest some standard references where this has been quantitatively tested and analysed? Thanks.
by willsmith
April 9th, 2013, 10:34 am
Forum: Book And Research Paper Forum
Topic: Best resources for commodity swaps?
Replies: 3
Views: 9224

Best resources for commodity swaps?

What are you swapping? It would probably be useful to model the underlying dynamics of the two assets - different commodities have very different spot price behaviours.
by willsmith
April 9th, 2013, 10:31 am
Forum: Trading Forum
Topic: does the term "neural network" have any specific meaning?
Replies: 3
Views: 8608

does the term "neural network" have any specific meaning?

<r>I studied neural networks quite heavily in my first degree in computer science, so I'd argue that neural networks have a very specific meaning, quite clearly defined at <URL url="http://en.wikipedia.org/wiki/Artificial_neural_network">http://en.wikipedia.org/wiki/Artificial_neural_network</URL> T...
by willsmith
April 9th, 2013, 8:00 am
Forum: General Forum
Topic: Expected range of a random walk?
Replies: 10
Views: 9219

Expected range of a random walk?

If it's brownian motion, isn't it simply proportional to sqrt(t)? I.e. 2 * vol * sqrt(t)?
by willsmith
April 9th, 2013, 7:58 am
Forum: General Forum
Topic: Value at Risk using implied volatolity
Replies: 5
Views: 8357

Value at Risk using implied volatolity

<t>Just thinking qualitatively : you can get implied vol for an instrument by looking at derivatives on that single instrument. To get implied correlation, wouldn't you need to look at derivatives priced off both instruments? Without that, I can't see how vol tells you anything about correlation. </t>
by willsmith
January 2nd, 2013, 1:39 pm
Forum: General Forum
Topic: Calendar spread option commodity question
Replies: 15
Views: 11585

Calendar spread option commodity question

Other well known models for the oil term structure are the "Gibson & Schwartz 1990" model and the "Gabillon 1991" model. Full references are in http://commoditymodels.com/recommended-papers/
by willsmith
October 25th, 2012, 7:49 pm
Forum: Economics Forum
Topic: Oil bubble again?
Replies: 40
Views: 47519

Oil bubble again?

<t>My opinion:The oil price is back to being controlled by OPEC. OPEC was dominant in the market in the 1970's-80's, lost market share afterwards but is back to having sufficient market share that what they say, goes. And OPEC's share is only going to rise, they have far more remaining reserves than...
by willsmith
October 25th, 2012, 7:42 pm
Forum: General Forum
Topic: U.S. to Get Downgraded in 2013
Replies: 7
Views: 10335

U.S. to Get Downgraded in 2013

If nobody thought it possible, it wouldn't already be on negative watch at some rating's agencies.It's hardly new news, is it ? "Moody's Warns of US Debt Downgrade - ABC News" - 11 Sep.
by willsmith
October 25th, 2012, 7:35 pm
Forum: General Forum
Topic: PhD not worth the time says Economist
Replies: 11
Views: 11996

PhD not worth the time says Economist

<t>I just did a PhD and every day I had time to have breakfast then go for a swim with my young son. Try doing that in an IB.In terms of expected $ payoff there's more chance that it's positive if you do the PhD early-career.But if you do the PhD mid-career - like me - you will enjoy it more (chance...
by willsmith
October 25th, 2012, 7:31 pm
Forum: Student Forum
Topic: Longsaff&Schwartz and fmincon in matlab
Replies: 3
Views: 14566

Longsaff&Schwartz and fmincon in matlab

<r>There's discussion of the longstaff and schwartz model here:<URL url="http://www.wilmott.com/messageview.cfm?catid=34&threadid=87942I%27m"><LINK_TEXT text="http://www.wilmott.com/messageview.cfm? ... d=87942I'm">http://www.wilmott.com/messageview.cfm?catid=34&threadid=87942I'm</LINK_TEXT>...
by willsmith
October 18th, 2012, 1:06 am
Forum: Trading Forum
Topic: Stock Friday returns are generally high?
Replies: 14
Views: 12364

Stock Friday returns are generally high?

<t>Quiet fridays are observed in all sorts of data. I've seen it in electricity prices (cheaper on friday than any other weekday) and also if you plot visitors to websites, friday is lower than every other weekday.Interestingly I found the "friday afternoon effect" in electricity prices to be more p...
by willsmith
October 18th, 2012, 1:00 am
Forum: Trading Forum
Topic: How implied volatility of options is determined?
Replies: 11
Views: 12380

How implied volatility of options is determined?

Implied by options prices. Basically you keep choosing implied vols and plugging them into the black-scholes option pricing formula, with all other parameters taken from the market, until yuor options price produced by black-scholes equals your option price observed in the market.
by willsmith
October 18th, 2012, 12:53 am
Forum: Technical Forum
Topic: Instable VaR question
Replies: 9
Views: 11146

Instable VaR question

In the spirit of VaR maybe if you have n daily samples you should not use the mean() at all but the 90th percentile, for example.
by willsmith
October 18th, 2012, 12:48 am
Forum: Technical Forum
Topic: time series volatility question
Replies: 14
Views: 11288

time series volatility question

<t>I've seen series with positive autocorrelation where monthly vol and daily vol are greatly different, specifically shipping indexes. Because they are not directly traded (rather polled from a panel like the infamous LIBOR) there's no efficient market theory to prevent autocorrelation. If I recall...
by willsmith
October 18th, 2012, 12:43 am
Forum: Technical Forum
Topic: Simulating different regimes
Replies: 3
Views: 10282

Simulating different regimes

<t>The standard way is to use Markov models. Define a few regimes (ideally only 2) and the probability of switching from regime A to regime B in any period. The probabilities become a matrix, with rows all summing to 1. Calibrating is difficult if it's not fully observable from the data which regime...
  • 1
  • 2
  • 3
  • 4
  • 5
  • 17
GZIP: On