red interval of the size ≥ 1/dN/Ke.
Consider now stage (c), where both players place their last points. The
following situations are possible after R places his point in that round: (i)
there are two or more red intervals, or (ii) there is one red interval. Assume
that case (i) holds. Then G places a point in the largest red interval. If the
size of the newly created red interval was ≥ dN/K e, then the green point
was placed within this interval, as after the stage (b) there is no red interval
of the size ≥ 1/dN/Ke. Thus after the last round there is no red interval
of the size ≥ 1/dN/K e. Since, also, by Lemma 1, there is the same amount
of green and red intervals and, moreover, each of green intervals is a key
interval (as by strategy Y*, G never created a green interval apart from the
first stage), so G must be winning (recall that the size of a key interval is
≥ 1/dN/Ke).9
Now assume that case (ii) holds. Consider the situation before R’s move.
It must be that there is no red interval. Consider the group of circles for
which dN/Ke key positions are assigned. It cannot be that R has more
points than G on these circles (as otherwise, by Lemma 1 there would be
a red interval there). Similarly R cannot have more points on the group of
the circles with bN/Kc key positions assigned. Thus on each such group of
circles, classified by the number of key positions, G and R have the same
number of points. Suppose that after R’s move, a red interval is created in
the group of circles for which bN/K c key positions are assigned (notice that
such red interval may have the size ≥ 1/dN/Ke). Then, by Lemma 2, there
must be a bichromatic key interval in that group of circles (i.e. a bichromatic
key interval of the size 1/bN/Kc). Thus G can win by placing a point in
that key interval and creating a green interval of the size bigger than the
9Notice that G’s advantage may be arbitrarily small, as non-key intervals created by
R may be arbitrarily close in size to that of key intervals.
20
More intriguing information
1. EU enlargement and environmental policy2. The name is absent
3. The name is absent
4. The name is absent
5. ASSESSMENT OF MARKET RISK IN HOG PRODUCTION USING VALUE-AT-RISK AND EXTREME VALUE THEORY
6. An alternative way to model merit good arguments
7. The name is absent
8. Estimating the Impact of Medication on Diabetics' Diet and Lifestyle Choices
9. Place of Work and Place of Residence: Informal Hiring Networks and Labor Market Outcomes
10. The name is absent