Volcanology Gone Wrong

 The above Tweet appeared on my timeline last night, followed some time later by the below Tweet for the same person:

I'll assume that the Tweeter is talking about this season (Twitter can be a little constraining with its 140 character limit) but regardless, it's a great example of someone misusing data out of context, and losing money as a result.

It seems to me that if Unders was a value bet before the game, then making a decision to trade out based on an irrelevant statistic is a mistake. Do you base your decision on whether or not to wear a raincoat on what the weather was like one year ago? Probably not, and similarly what has happened in Doncaster's previous matches in the 15 minutes after half-time may be a curiosity, but it isn't useful data. There's no edge in trading out at half-time - you're giving value away. The statistic is up there in the same degree of uselessness as the practice of using previous matches between clubs as a predictive indicator. 

If the Tweeter determined the Unders was value pre-game, and if his long-term record shows that he does indeed have an edge on these selections, then laying off the bet in-play is making the number one trading error - trading out of winning trades too soon.

If finding value is that easy, let it run. Trading out for no logical reason means you are reducing profits, not to mention there's the hidden cost of the time spent sitting in front of a screen, although some people seem to have more of that than others.

The match finished 0:0.