On the Coniiniilfij of Chance li 



at any time. This, or any other supposed law, may. after all, 

 be but a run of luck in au infinite enough distribution of 

 chances. 



To say, however, that a certain possibility has to be faced 

 is not to assert a probability, A complete theory requires a 

 discussion of even trifling possibilities, but the main business 

 of science, as of life, is with the large probabilities. Though 

 it be true that, no matter how large a gambler's fortune or 

 how small his stakes, there will inevitably, at fair play, in 

 time, come such a run of bad luck as to sweep him off iris feet; 

 still in the brief period of a life, the probability may be so 

 small as to be negligible. 



Small though the gambler's risk may be, we should deem 

 it enormous compared, say, to the probability that the law of 

 gravitation was a mere matter of chance. But we need not, 

 therefore, say that chance does not, or has not, played a part, 

 even in the law of gravitation. 



Suppose our gambler does not play a fair game. Let the 

 odds be slightly in his favor. Euin is no longer inevitable 

 even in infinite time. The tendency is for his gains to more 

 than cover his losses. The tendency may be imagined as 

 strong as one pleases; at last we have almost a certainty, a 

 law. Did the gambler start with playing fair, but then gradu- 

 ally perfect himself in methods of unfairness, we should have 

 a growing tendency, au evolution of law. 



Take a more mechanical illustration, due to Mr. C. S. 

 Pierce. 



Spin on a table 1024 coins each ten times. The probable 

 results are given by our previous table for 1024 throws of 

 ten coins, only for throws substitute coins, thus: 



141 



