Archived from groups: rec.games.empire (
More info?)
"Markus Armbruster" <armbru@pond.sub.org> wrote in message
news:87oedi3e2t.fsf@pond.sub.org...
> Quote from
http://www.angelfire.com/empire2/bungholio/PZ4-TheMakingOfPZ4.html:
> BUNGY's SEMI-HIDDEN FORMAT:
>
> - = [ Empire Power Report ] = -
> as of Mon Feb 28 21:33:21 2005
> sects eff civ mil shell gun pet iron dust oil pln ship unit
money
> 2 ?? ??? ?.?K ?? ? ? ??? ?? ?? ?? ? ?? ?
??M
> 5 11 91 11K 322 490 21 4 1 3 502 0 18 0
17K
> 3 ?? ? ?.?K ?? ? ? ??? ?? ??? ?? ? ??
? -???
> 1 ? ??? ?.?K ??? ? ? ? ? ? ? ? ? ?
??K
> 4 ? ??? ?.?K ??? ?? ? ? ? ? ? ? ? ?
??K
> ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
> worldwide ?? ?? ??K ??? ??? ?? ??? ?? ??? ??? ? ?? ?
??M
>
> ORIGINAL FORMAT:
>
> - = [ Empire Power Report ] = -
> as of Mon Feb 28 21:33:21 2005
> sects eff civ mil shell gun pet iron dust oil pln ship unit
money
> 2 19 100 3.9K 19 0 0 76 19 57 38 0 12 0
12M
> 5 11 91 11K 322 490 21 4 1 3 502 0 18 0
17K
> 3 37 0 7.5K 37 0 0 148 37 111 74 0 11
0 -143
> 1 3 100 2.1K 111 0 0 0 0 0 0 0 0 0
25K
> 4 2 100 2.0K 122 60 4 0 0 0 0 0 2 0
25K
> ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
> worldwide 72 47 26K 611 550 25 228 57 171 614 0 20 0
12M
>
> Unfortunately, I think I bumped heads with Markus on this when trying
> to get it incorporated into the official server (Sorry Markus!).
> I won't get into details on it. What I think I need to do is run a
> game with this option, and see if players like or dislike it as-is. I
> wanted this option because I wanted something inbetween the
> way-too-acurate standard power report and the total-paranoid-hidden
> game..
>
> End quote.
>
> Permit me to go into details
>
> Your proposal is a mixup of two ideas, one possibly good and one
> probably not. The possibly good one is to mess with the way power
> rounds --- many players are unhappy with the amount of information it
> discloses. The other one is to mess with the way power prints
> numbers. Let me elaborate.
>
> Consider a non-negative number x (my argument can easily be extended
> to negative numbers, it just complicates its presentation, so I
> don't). Your masked power report prints x as certain combination of
> question marks, decimal point and a suffix character.
>
> Such a string represents any of a range of numbers. You get its
> inclusive lower bound l by replacing the leftmost question mark with
> the digit 1, and all the others with 0. The exclusive upper bound is
> 10*l. Observe that l is a power of ten.
>
> In other words, your funny string contains the following information:
>
> l <= x < 10*l
>
> Conversely, all x in this range print the same funny string.
>
> Therefore, printing l instead of the string gives *exactly* the same
> information. Instead of l, you could choose any other representative
> for [l..10*l), say 5*l.
>
> Mapping numbers to your funny string actually consists of two
> operations: a rounding operation (truncate to the next power of ten)
> and a number formatting operation (print '?' instead of digits).
>
> Your funny formatting operation a bad idea for two reasons:
>
> 1. It breaks tools parsing power gratuitously.
>
> 2. It only works for one particular way of rounding. What if we find
> that rounding to powers of ten is too coarse, and want to try
> powers of 4, or powers of 10 times 0.25, 0.5, 0.75 and 1? Can't do
> that if we foolishly chose to print '?' instead of digits.
You are overlooking the obvious good points: the change was simple
and the result is masking, not rounding (beyond that which is already
present in the standard code). You saw Ron's code, it was only
about 10 lines and it was professionally implemented - not spaghetti
code.
The effect of the question marks masks the detail. This is more
classification by powers of 10 rather than rounding. I agree that any
digit can be inserted for the ? but that may be misleading by making
players think that it is a form of rounding, which has other implications.
If you are concerned about the 'gratuitious' breaking of tools, simply
replace all '?' with a '3' to obtain an effective log10 rounding.
You are concerned about 'what if we want to try something else?'
Well, we are trying a first attempt. Let us learn from trial-and-error
before you pronounce failure.
Tom