his weight for advantage tables were very inefficent..
If you test his method over a large sample it falls down badly..