Predicting earnings by the end of the month

I have recently discovered a tool of Yuri Arcurs that can predict your monthly earnings based on the sales you had so far. This tool uses historical data of Arcurs. I started to use the tool and compare it to the linear prediction I believe most of us use. Its formula is simple and straightforward:
LinearPredicion=TotalDays*SalesSoFar/DaysSofar

I have started to take notes on both predictions from 3rd of September. Now that we have reached the middle of the month here are some statistical details:

	Arcurs's prediction	Linear prediction
Median	$71.31	$68.51
Average	$75.21	$74.62
Minimum	$57.71	$54.83
Maximum	$111.44	$123.8
Standard deviation	$14.69	$19.76

As it can be seen from the table above the usual difference between Arcurs' tool and the linear prediction is within the boundaries of standard deviation. For me it means that the tool cannot be used more precisely for my data than the simple linear prediction algorithm.

Below you can also see the chart showing the relations between the current earnings and the two predictions.
Finally you could observe that the linear prediction always grows within a given day. Its because the formula I used for the calculation is on daily basis. If I would have used smaller granularity in timing the differences between the two predictions would be probably smaller.

At the end of the month I plan to report when did these two predictions reached the best prediction and how close did they matched the real earnings figure.