Hi,

We have experienced some really crazy data behaviour over the past three weeks or so.  Some of the shares, specially the Australian ones, were showing data that was just off.  We were monitoring this closely and did testing on our end, but it was the raw data coming from our source.  And then, after much ado, it just went away like that… weird stuff.

The data behaviour included graphs not showing, some of the graphs had the exact same data for different time scales, and the like.

But it is all good now.  And I don’t know how or why that bug appeared or disappeared, I wish I did though.

There is, however one share (AAPL) that is gonna look a bit funny, but that was a real scenario we were dealing with not some data fluke.  Apple, as you all must know, reduced the price of their shares to about $90 USD from the upper $500s.  Now we had a conundrum here.  Do we want to change the past?  Could we?  In a way -and that is clearly what all other applications have done- you need to change the past so your statistics make sense, specially if you are dealing with the stolidness of moving averages.  But in our case with our Viracocha graph we don’t really need to, since we are dealing with the variation, so spare the few days after the change when variation will be great, it will then adjust to the new reality.  So I don’t really know if we should change the other two graphs -the average price- and the -MACD-.  The latter is a pure moving average and it is clearly misaligned, and that would call for a readjustment of the past I suppose but even that one will be fine after 30 days or so; and besides we have always considered that one to be just for reference and double checking.  And with the average price, well that one is reflecting in our app what really happened to the share price, a bit of historical land-marking if you will.

So I am a bit undecided on AAPL.  Like I said Viracocha graph is un-sensitive to this monumental change in a way, so that one is working fine.