1) The calculations for predictions must be shown
2) Predictions need to be rounded to the same level of accuracy as the raw data.
3) Your prediction must be put into context. Note: it is just an approximate value.
The validity of a prediction is dependent on two ideas.
1) How well does the trendline represent the centred moving mean. Look carefully at the end of the CMM. The most recent data is always going to be the most useful. If the CMM starts falling away or increasing rapidly at the end of the graph this may indicate change in the future. Note: It will rarely be a perfect fit.
2) How well the average seasonal effect represents the points that it represents.
Note: A prediction is always going to be an approximation and is also never going to represent unforseeable events.