Reply to Braun and Schmidt.
Reply to Braun and Schmidt.
复制标题
回复布劳恩和施密特。
DOI：
10.1016/j.cub.2015.05.008
复制
发表时间：
2015
期刊：
影响因子：
--
通讯作者：
Misirlisoy E
中科院分区：
文献类型：
--
作者：
Misirlisoy E
我们欢迎并赞赏 Braun 和 Schmidt [1] 对我们最近的信件 [2] 关于点球大战期间守门员行为中赌徒的谬误的评论。 Braun 和 Schmidt [1] 断言二项式比我们的重采样方法更适合分析数据。他们的分析似乎假设潜水方向呈二项分布，概率为 0.5。事实上，守门员的行为存在一些偏见。首先，我们表明踢球者会稍微偏向守门员右侧踢球，而守门员也会稍微偏向向右扑救。这种偏差并未超出我们数据中的传统显着性阈值，但在对动作序列进行建模时仍然不能忽略。从任何分布中重复独立采样会产生取决于偏差的顺序模式。因此，用于检验顺序依赖性假设的二项式分布应与实际样本具有相同的偏差。
We welcome and appreciate the comment by Braun and Schmidt [1] on our recent correspondence [2] about the gambler's fallacy in goalkeeper behaviour during penalty shoot-outs. Braun and Schmidt [1] assert that the binomial is more appropriate for analysing the data than our resampling approach. Their analyses appear to assume that dive direction is binomially distributed with a probability of 0.5. In fact, there are several biases in goalkeeper behaviour. First, we showed that kickers have a slight bias to kick to the goalkeeper's right, and that goalkeepers also have a slight bias to dive rightwards. This bias did not cross the conventional significance threshold in our data, but it still cannot be ignored when modellingsequencesof actions. Repeated independent sampling from any distribution produces a sequential pattern that depends on the bias. The binomial distribution used to test hypotheses about sequential dependence should, therefore, have the same biases as the actual sample.