Diary

Wed, Jan 4'

7:15 Start new daily price sheet write Meg a Prispr. card 9:15 to Dr. Hills for new crowns Drive by Microsoft 1:15 At home. lunch Lookover Sum 87 - current quarterly reports to see what other ways Microsoft is diversified. And decide not to include VL adjustment for '87 stock options since they don't adjust fiscal '88 3:15 Update sheet But figure that it is best compromise to use the middle ground earning figure which excludes not only tax benefit of NSO [...] but also expense of it But even though I have some success creating the quarterly instances, I cannot get the fourth quarter except by working back from annual. Also realize this would be incompatible w/ forecast S & P uses the most conservative figures. 5:15 So I go back + use those which UL also uses now, but at least I understand mechanics of it now. two lessons: complications of interpreting earnings. [...] can't undo what it once published. Finish making it look good. Dinner. put together "Page 3" to compare UL forecast P/E + actual postfacto P?E. Play around w/ charts. Print all. 10:45 to bed. One [...]

thu Jan 5 '89

Woke up thinking about these numbers and what more to add. Not much

7:30 --Improve formats of table a bit. Play with charts more 9:30 Go to Failla Go to library. Into world + UL. Lunch at Blocks. Go to Starbucks + pick up WST at Failla 1:45 Make charts look better, consistent. Print 'em Refine. Finally they look great: Backup. Make notes of what to add to "Routine" 5:00 ----Put awat UL + Infoworld clippings. 5:45 ----Yesterday + today's mail. Fay Ex hands. 7:00 Look over strategems newsletter on ExxEd Have ideas on comparing MSFT to other stocks and write these ideas down. 7:45 Go to gym 8:00 Arms Think that I can divide stocks into 4 categories depending on [...] And for DIV O, P/E low or H. Steam + Shower a bit 10:00 Go home. Look at Vanguard Funds list + this pretty much describes the layout. So I can just use UL to see what thypothetical funds on [...] would be lie, and compare those to MSFT. To neighbors - Nah To Tubs? Save it for Saturday. Play with UL database. Make matrix showing characteristics of stocks in each type of augfund. 1:00 To bed. One [...]. 1:30 But get up to try selection based on Div 2:15 Back to bed. 4:00 [...] Wake up to pee + realize that this is a further integration of comparing bondss w/ Div paying stocks Div paying stocks with Nondiv (Growth) Growth stocks as a group with MSFT Back to bed ... SNOWING HARD

Fri Jan 6' 89

8:00 Think about these comparisons more 8:45 [...] ---10:00 To Studio 904 for coloring + cut "Start "Doc ordered" 12:00 Lunch at Umbertos. Go to Univ Audi for Wiper blades. Come home to put them on. See Christina + guy living in Brunois. Call Mom + Dad while hand warm up. Finish wiper blades. 2:15 Investigate statistical output of the three different funds I've envisioned. Try to decide whether to use average or medium for each value, trying to sense the dispersion and balance for lack of it in each Histogram. When it came to P/E, I decided I should compete 1/Aug (I/PE, + I/PE, + ...) and I used Excel for this. I found that the little P/ES, [...] , really throw things off. Because these are wierd, anyway, I decide to always exclude the bottom 5% of stocks sorted by P/E, which leaves MIN P/E of 6 for December. Compute 1/Aug(I/PE ...) for each, fund and find it comes close to median. Also justify using median of yield. Do ratings + Estimates, other Marker data. 7:15 Dinner. Examine [...] of S.T. ERS history and projections. Write it np. Make list of what to keep in sheet. 8:30 WSW with John Templeton. 9:00 Make the sheet. 3 pages, 4 values per page, eachHD, LD, ND, MS. [...] figures for December. Try charting page one. Do so with lots of different patterns, and then go for bar chart. Got it to look good on printer too. 12:45 To bed. One [...].

Sat Jan 7 '89

8:15 -----. watch wsw: templeton again. washer not fixed yet 9:45 Add "statistical background" page To monthy equity report. Refine the borders a bit. Washer still not working. Laundromat. Refine sheet a bit more. Great. Play with Horiz + Vert charts Bar + Column. Placement of legend. Lunch at home. Decide how to [...] values onto 2 pages. All charts for 2:30 Think about how to compare P/E with higher earning or dividend growth. If I make a column showing this, what could I discard? 26-wk % chg, since its sum of 2 13-wk chg? Check. Go home. Keep thinking about how to compare What to compare with what? All compare to MSFT, due to my perspective Or with [...] , since its the most stable. Now will I rearrange columns? Should I shown [...]? See how timeliness varies: HD, LD, ND. Note I get extra precision, one decimal, using tabular reports to compute average timeliness. Financial strength is a fundamental rating. Beta is a pricing rating. Diff pages? What should I use as the growth to compare against P/E? Earnings growth Div yield + Divd growth Div yield + Earnings growth. 3:30 Dinner at home. Pretty Spartan. Golden girls. 7:30 Realize I can chart P/E + Growth, so maybe I don't need to decide on one element to compare too. The picture will be the comparison. Graphing is just another formatting. So maybe I do not need another column, for comparison, although maybe for timelines Try log log, but it only allows increments of 10 on scale. Why? ... Do it with regular axes.

but i still don't know if I should use earning, divd, or Prj Av for Growth. I look at Pri Av on MSFT 1/4ly report from VL and see that they have 35% EPS growth which they run for 2 years and get $5.50/sh in 1991-93 and mult by P/E of 20, getting 110/sh, but then they take square root (110119), why four years it only counting 2 years of 35% growth. And from previous experience with it, I know VL use diff # years for diff stocks ... So bug [...]. Maybe don't even put it in a column. I still think about how to compare MSFT against ND group when they have Fin Str difference of Aus C++. So I select only stocks having FinStr > = B+, so A is avg, but this makes little difference to P/ES! Or even to PrjAvRet. So stick with orig ... But Fin Str is important to me, as it reduces risk, if not increasing Pr, AvRet.

Looking at timelines, I hate that all of the funds are between 2.6 + 3.3, so pretty even, and this is good as it indicates they are well priced, S.T.

Try to convince myseld that d+g should equal EPS growth theoretically, as it seems to empirically , but I cannot. [...] 12:45 Should I go out 70 Eagle or Tubs. Go get glass of wine to cook off [...], and see its been snowing. Streets are white , so stay in. More thought wine makes me groggy. 1:45 To bed. One Am it ....

Sun Jan 8'89

In bed, decide to use Div+G(divd) for return on HD+LD because G(divd) falls between Past EPS growth + Prj. EPS Growth. For ND, do something similar, but Div is 0, so assume G(divd) is halfway between past 5 yr EPS growth + Prj EPS growth. For MSFT, just pick some number.