Lexander Morinex
Caldari
LDD Investments
   |
Posted - 2008.12.16 06:00:00 -
[1]
 Originally by: Treelox
Which means that the "good doctor" has either no concept how things work in game, has lousy long range forcast abilities, is not listened too, is not influence on these changes that might of already been in action prior to his arrival, or some combination of all of the above.
I am Statistician, not an Economist, but I do feel I can comment a little about the "good doctor". He seems to be committing a very common error amongst data analysts.
These large databases are prone to 'fishing expeditions' and other tricky data mining problems. There is a lot of random noise in a big database, and it is typical to generate lots of different metrics, meaningful or not. But in the end, the data you generate is only as good as the subject matter expertise that drives it.
In my current dissertation work we deal with low-level gene microarrays. I have spent countless hours pouring over chart after chart, often to find out that what seems so important to me is completely obvious to the biologist. It is a collaborative process, and the 'good doctor' seems to not spend enough time studying what players find to be important.
I am not convinced this kind of exploit would be discovered with the current information. With the sheer amount of white noise in the system, an exploit of this type can get missed. Even if one or two results showed up as significant, there is the deeper problem that the Type I (false positive) error rate is just too high for practical use. Unless the information can be filtered with expertise about the game, the statistical analysis is going to be useless.
The deeper concern I have is not that this exploit was 'missed', so much as the fact that the economic situation as a whole is not properly evaluated. The numbers coming out of the economic reports say very little about the player experience and a lot about a bunch of numbers. In statistics there is a distinction made between statistical significance (a result unlikely to happen by chance) and practical significance. Too many of the results shown in the economic reports have very little practical significance.
Traditional economic measures are no substitute for a deeper understanding of an online economy. Just because the metric meant something in the real world does not make it useful in the game. The 'good doctor' is producing a lot of metrics that might be useful to real life bankers but are essentially useless to players. It is rather sad to see that much intellectual brainpower essentially wasted chasing the wrong problem.
- Lexander Morinex
|