Comments/Ratings for a Single Item

Grand Chess and family is less sensitive to openings than FIDE-Chess, but
it is not insensitive, as Altair/Achernar/Deneb are not insensitive but
definitely less sensitive to openings. The possibility of a wide variety
of openings without inmediate risks of a clear loss of the game is
important in every game, and in some of them the sensitivity to openings
may play an important role in this variety. Further investigations on this
topic should be good.

Perhaps we should attempt a Sensitivity Scale from [1] Shatranj to [10]
Wormhole Chess - a variant where early checkmate is possible, if not
probable!  My 'Lions and Unicorns Chess' variant also contains some
opening traps leading to long forced lines - too long for Zillions to
calculate in advance.

Or the sensitivity scale could be based on the minimal number of moves(not
 turns) needed to threaten an opposing piece.

For example:  FIDE might be considered equal to 2.  White moves then Black
 can threaten.  1. e4 Nf6(or d5)

This will be a fairly easy value to quantify for each game and it involves
 a prove-able dynamic.  Thus limiting the potential for un-intended error
 based upon bias.

Or the value might be the minimal number of moves to threaten an opposing
 goal piece.  This could be difficult to quantify since it might assume
 poor performance by one player and superior performance of another.  And
 would this value be applied regardless of the potential of capture of the
 threatening piece?

The idea could be to calculate the minimal number of move before one
 player is forced to perform a move in defense.

Larry is on the right track, but the moves to attack a piece measure will
give false results in some games. 

In Decima, for example, White's usual opening move attacks a Black piece,
yet the game is less sensitive than FIDE chess. 

It may be that number of moves to attack a goal piece is not always
accurate, though no examples spring to mind. Clearly more resarch is
needed to decide the best measure.

David hit one of my biggest pet peeves. And it's this: Don't use 1-10
scales. They are biased. Go ask everyday people (ie, without any relevant
background) 'on a 1 to 10 scale, what is average?' And they'll say
'five.' Well, it makes sense, right? But it's wrong. Average is five
and a half. And this makes people give ratings that aren't really what
they are thinking.

I'm sorry if this is too mathematical for your taste. It's just that I
needed to say that. Please use 0 to 10 scales, those work nicely.

Michael Nelson writes: 'It may be that number of moves to attack a goal
piece is not always accurate, though no examples spring to mind. Clearly
more resarch is needed to decide the best measure.'

I believe that we will always have to rely on subjective evaluation.  Any
attempt to quantify runs into paradoxes like: The Fool's Mate in Shatranj
is only fifty percent longer, compared to FIDE chess.

1. Pawn d2 - d3 
1. Knight g8 - f6 
2. Knight b1 - d2 
2. Knight f6 - d5 
3. Pawn b2 - b3 
3. Knight d5 - c3

It is highly preferable to quantify this factor if at all possible (in
accordance with L. Lynn Smith's proposed criteria, in my opinion). 
Otherwise, the sought game-specific evaluations are likely to degrade into
a chaos of widely-disparate subjective opinions and value judgments,
impossible to reach a general concensus on and hence, not readily useful
without reservation (as a table of values).

Whatever important game-specific factors are not accounted for by the
main, proposed initial sensitivity criteria can be supplemented by another
factor, designed for that purpose, as a positive or negative adjustment to
the total.

Some extensive, but short computer simulation experiments can be performed,
say, with the first ten moves. We can see: Percentage of finished games,
percentage of forced captures, percentage of threatened goals, and
percentage of clear disbalances between teams, defining a disbalance as a
difference of some quantity of points, assigning points to pieces and/or
other determinant factors according to the game in a rational way. But I´m
not clear which are the best measures to see how sensitive is the game to
openings

When there are subjectivities, statistics are good. Why not a poll?. Voters
can list the ten more sensitive games to opening, and the ten less
sensitive, in order. Ten poinst to first place, nine to the second, and so
on...

I believe that some might be mistaking the simple quantification of a game
through prove-able values as a determination of either its play-ability or
enjoyment.  These values will only establish particular parameters which a
game might be categorized.

For example: some players will prefer a game which takes a large number of
moves to develop before any significant threat occurs, and others will
enjoy a game which starts with exchanges.

The preference showed these values should be left to the potential player
to determine.  And this argument can be the source of many challenged
games.

[Note: I use the term 'argument' in its positive form.]

Larry, you are right, preferences may be different according to the player,
but it is a good indicative for a novice having an idea about what is he
going to see in the game. If you have never played a sharp game from its
beginnings, as Marsellais, Ximeracak or Pocket Mutation, by example, you
can play it without the needed care, being innocent about what is coming
up quickly, and the results may be decided in a few moves. If you have not
had previous contact with a game, it should be good some knowledge about
the sensitivity of the game to the initial moves, and the 'sharpness' of
the game, another subjective and difficult to stablish measure. Chaturanga
and FIDE-Chess are not very different, but Chess seems to be 'sharper'
than Chaturanga, and much more sensitive to the initial moves.