# cfai book 1 quant question (confidence intervals)

so… my brain is freezing here reading 12, p 348, question 26. (end of chap. questions) the default spread coefficient is shown as 3.04, the t-stat is 4.52… so the text calculated standard error for this simply as 3.04/4.52 = .673 and got 3.04+or- 1.31908. is this correct?

I don’t have the book handy but, the t-stat = (coef - 0) / std error, here you are given coef and t so std error is just the coef / t-stat = 0.673 so 3.04 +/ -1.96(0.673) ??

great. i got so used to the book giving me the standard error in these problems that i forgot how to actually calculate it. embarrassing. at least it’s early on. Thanks slouis.

no sweat. I miss quant. I just read about calculating nominal and real-term financial projections in order to prepare a DCF valuation of an emerging market company. yikes, what are the questions on that LOS going to look like? The growth duration model is interesting though.

slouiscar How is the preparation going? I know that you were getting ready to place bets based on your new quant knowledge. Any other “real world” applications so far?? What all have you covered?

Hey, Prep is going. I haven’t really started studying but I plan to finish the CFAI readings in early Jan. Right now I am halfway through Equity Val, I like to go in order on the “walk through”. The NCAA FB models were a lot of fun. I attempted to regress all kids of neat stats against margin of victory. w-l records, strength of schedule, points allowed, points scored, etc. are all available and updated weekly going back 5 years or so on the NCAA FB website. The model that got me excited was margin of victory = bo + b1 home team winning % * home strength of schedule + b2 visitor winning %, * strength of schedule + Ei. n = 50, R^2 was .38, sig F, t- stats all over 2.5. Here was the interesting thing… the results of the model, margin of victory, had a high correlation with the spread. That is, I found that if the model said that this week E(margin of victory) = -22 for #24 Cincinnati at Syracuse, well, when I look up the line, it is Cinn -20! Each week if there were 40 games, for 30 of them the E(mv) would be so close to the point spread that I wouldn’t want to put money on it… For the other 10, I would keep track. Unfortunately, as you would expect, the results were average, turnovers and injuries all fell neatly into the error term. In the end the results weren’t significant but the exercise was a great Excel learning tool. In about five mouse clicks you can take your data and generate a worksheet with all the regression statistics with an ANOVA table just like in the textbooks. Lately I have been toying with the Gordon Growth Model, DDM using formula/functions in Excel and in javascript. When I fine tune and debug maybe it will be more useful than a gambling tool. By the way, did I read that you have a new addition coming in March? Congrats.

slouiscar Wrote: ------------------------------------------------------- > Hey, > > Prep is going. I haven’t really started studying > but I plan to finish the CFAI readings in early > Jan. Right now I am halfway through Equity Val, I > like to go in order on the “walk through”. > > The NCAA FB models were a lot of fun. I attempted > to regress all kids of neat stats against margin > of victory. w-l records, strength of schedule, > points allowed, points scored, etc. are all > available and updated weekly going back 5 years or > so on the NCAA FB website. The model that got me > excited was margin of victory = bo + b1 home team > winning % * home strength of schedule + b2 visitor > winning %, * strength of schedule + Ei. n = 50, > R^2 was .38, sig F, t- stats all over 2.5. Here > was the interesting thing… the results of the > model, margin of victory, had a high correlation > with the spread. That is, I found that if the > model said that this week E(margin of victory) = > -22 for #24 Cincinnati at Syracuse, well, when I > look up the line, it is Cinn -20! Each week if > there were 40 games, for 30 of them the E(mv) > would be so close to the point spread that I > wouldn’t want to put money on it… For the other > 10, I would keep track. Unfortunately, as you > would expect, the results were average, turnovers > and injuries all fell neatly into the error term. > > > In the end the results weren’t significant but the > exercise was a great Excel learning tool. In > about five mouse clicks you can take your data and > generate a worksheet with all the regression > statistics with an ANOVA table just like in the > textbooks. I really wish I had better excel skills. Something I will have to work on in the future. > > Lately I have been toying with the Gordon Growth > Model, DDM using formula/functions in Excel and in > javascript. When I fine tune and debug maybe it > will be more useful than a gambling tool. Keep me updated. > > By the way, did I read that you have a new > addition coming in March? Congrats. Yeah…another boy coming in early March. He is already altering my life by way of my current study schedule!

Newborns, ouch. But boys are a piece of cake. Throw my son in a safe room with a big empty box and a wiffle ball bat and you can close the door and check back in an hour or two. If you want schedule altering, spend a day with my 4 1/2 y.o. daughter. It takes 50 minutes of tears and drama for her to get dressed if she doesn’t feel like wearing a particular color that day. Of course I know her mom, so I should have seen that coming. It is like having twins. Either way, it is time well spent as you know. … now that I think about it, due “early march” - 280 Days = early May 2007. I can’t remember if I even talked to my wife last may other than to explain to her why I was an obsessed, “direct method” zombie.

Ah yes…daughters. I haven’t had one yet (hope to in the future). I am sure that lack of sleep will play a big role in my prep for level II. My wife was pretty annoyed at the end of the prep for Level I with just one boy (She was mad I was always on AF!)…I can’t imagine what it is going to be like this time. At least if the newborn won’t sleep I can read him stories about pension accounting. That should knock him out in about 5 minutes.

mwvt9 Wrote: ------------------------------------------------------- > At least if the newborn won’t sleep I can read him > stories about pension accounting. That should > knock him out in about 5 minutes. LOL. good luck with that. Your baby will be a CFA candidate right from the first hour of his birth.

My second one is going to be delivered this Thursday. The newborn needs to take mid-night mike for several months. I need to integrate my study with the feeding time as I will wake up at mid-night. Remember that even the one who feed the baby is your wife, she (your wife) will wake you up anyway. Reading the pension accounting may not be able to get the baby to sleep as his father could fall into sleep earlier than the baby. My experience is to dance with the baby and listen to the music together. For the progress of Level II study, I have done 2 Schweser books - speed-reading and doing the end-of-chapter exercise. As I don’t know how much time is left for studying after the baby is born, I want to finish first round of study as fast as I could.

mwvt and jogging. best of luck to you. i had my first a week before the level 1 test… i bet it is a lot more difficult when you have one already running around on top of that! it’s harder for me to study now with a 6 month old at home because i want to come home and spend time with her instead of reading quant in a conference til 8pm. but i guess it’s all for them anyway.