Comment This is the tail - it means more (Score 3, Informative) 172
We don't have enough history to gauge what actually has happened over time, so we have to estimate.
We approximate by finding big rocks or chemistry on earth, looking at craters on the moon, or this.
In all these cases we are using the small but frequent to infer the distribution of big but hugely problematic events. Our best answer the question about the likelihood of a killer impact is grossly changed if this tail is changed.
Think about it like floods. We ask how likely a 10,000 year flood is going to happen next year. We have ~100 years of rainfall data. We fit it to a distribution that is appropriate and then use those fit parameters to make a best guess. If our rain gauge was only measuring half the rain, we might under-estimate the actual risk by a factor of 10x or 20x.
There is good correlation between "killer impacts" and location of the sun in the galaxy (yes it moves around). We are starting to enter a higher risk region (transition to edge of arm) and perhaps the fundamental distribution is changing. In that case the history of craters on the moon or other might not be meaningful indicator of the near future.
Considering this I think good tracking is not a bad idea and should be thought out well and properly considered.