摘要或動機

We investigate the machinery producing successive Simpson’s paradoxical reverse. Taking advantage of algebraic and geometric techniques, we obtain the following results. Take playing baseball for example. In our study, we find that Simpson’s paradox only occurs when the hitter’s hits over 3 times in one game. Set n equal to the times I will hit in one game. If my batting average in each game is at least(n ?1)/2 times higher than the others’; then I am sure that my total batting average would not be invert by the others. In order to find how many the lattice points in the triangle, we use Pick’s formula. But sometimes, the Pick’s formula is not appropriate to triangles whose vertex are not all lattice points. So we develop New Pick’s formula to estimate the number of lattice points in such kind of triangles. Besides, we also find an iterative algorithm to produce successive “Simpson reverse” phenomenon by using C⁺⁺ language, and we can therefore produce as many “Simpson’s set of four sequences” terms as we like(not beyond the computers’ upper limit).Moreover, if both sequences of ratios converge, then they must have the same limit.我們探討了一般人乍看之下顯得頗弔詭的辛普森詭論。我們配合GSP 作圖，用解析幾何、設立直角座標系和C++ 程式的運算，找出在特殊情況下或一般情況下所產生的辛普森數列組和特殊的性質，並且以棒球場上的打擊率為例子來做印證。通常一場棒球賽中，每個人平均上場3 次～4 次，經過我們的討論，發現要發生逆轉的機會只有在上場達到4 次或以上時才會發生。?了求出在直角座標系中可以滿足的格子點個數，我們用了Pick公式，但?了更準確的估計，我們引進了虛點的概念，重新推導出了新Pick 公式。另外，我們還發現，假設兩個人上場比賽，若打了2 場，且每場最多上場打擊K 次，其中的一個人的打擊率只要是另一個人的(k-1)/2倍以上就保證不會被逆轉。我們又找到了連續產生辛普森逆轉的演算法，利用C⁺⁺ 寫出程式，經由演算法和遞迴式，製造出項數可任意多(只要電腦能夠承受)的辛普森數列組，且我們發現若兩個比值數列接收斂，則極限趨近於同一個數值。