In our case, we must be very careful with data. A dramatic error in the number of doctors in Ethiopia would surely cause significant shortcomings in the reliability of any projections for Ethiopia. This means that any multi-national Nash equilibrium involving Ethiopia will also be distorted or perhaps missed entirely. Worse, we could easily come to completely wrong conclusions about whether Ethiopian schools are doing a good job (increased the count of doctors by several hundred from 2005 to get to 3749) or a miserable job (maybe the count went from 1849 in 2006 to 1806 in 2007, so schools did not provide enough replacements to deal with retirements, deaths and emigration).
Prior Click here
Next Click here
Next Nabro Click here
Next Maputo Click here
Next Gabon Click here
Next Iran Click here
Worst of all, suppose we are in doubt about the number of doctors working reported for Vietnam. We might wish to compare Vietnam and Ethiopia as both have about 90 million people and somewhat similar economies. An inaccurate figure for Ethiopia can easily cause us to believe an errant tally for Vietnam. It would be at least ironic were we to statistically prefer an imprecise figure for Ethiopia over what might be a truthful number for Vietnam. Demographers will observe that finding lots of comparisons for Ethiopia it tough - the next largest country is the Philippines with over 100 million people and the beyond that is Mexico. Likewise, the next smaller country is Egypt with 82 million followed by Germany and Turkey. Confusion over Eritrean data could influence decisions about a very diverse collection including Jordan, Laos, Paraguay, Papua New Guinea, El Salvador, Nicaragua, Kyrgyzstan, Denmark, Slovakia, and Finland. In the case of the simultaneous miracles in Brazil, Israel and Portugal the trailing lower case ell was read as a one and merged with the following numbers. The problem was obscured because Nepal and Senegal have not reported recently.
Before actually committing a new tally for 2009 for Eritrea we would do an autocorrelation with previous Eritrean tallies AND check the correlations with other 2009 Eritrean data. That's somewhat of a burden for the CPU, but manageable. Redoing a cluster analysis to verify that the new 2009 number  does not perturb any long-standing similarities between Eritrea and others is also plausible. Until we confront updates for 180 countries. Who goes first? This cries out for parallel processing using GPUs.