- This topic is empty.
- May 13, 2021 at 12:30 pm #27902Jake MoodyParticipant
In the first quarter I was basing my research on data supplied by website before the recent changes and picked the suburb “Christies Beach, SA” to purchase an investment property. The old data was showing Christies Beach as 16.8% capital growth with a high confidence (one of the best in the state) and since the changes recently made it now shows as a low confidence and many other suburbs as being much better (up to 25% growth with a high confidence). Are you able to explain this? I am feeling that if I had waited and seen the recent data, I would have picked a different suburb.
Also, how often will these statistics be changing drastically? Or with the new layout will they only change gradually from now on?
looking forward to your reply,
- May 13, 2021 at 12:46 pm #27907KenKeymaster
Thanks for your email.
We have recently transitioned to a new model and reporting format for suburb ranking tables. The new model uses a revised method to compute error rate which in turn defines confidence. The ranking table presents the Capital Growth data as a per anum average estimated growth instead of a 2 year compound growth used previously.
We are currently working on transitioning the council area pages from the old to the new reporting format. I believe you are referring to the data currently presented on the council area page for this suburb in the old reporting format.
I’ve gone ahead and exported the suburb data in the new format for Onkaparinga from HtAG Store. Christies Beach as well as 3 other suburbs that are showing higher 2 year forecast values than before (as per your comment) are listed. As you can see although confidence changed in Christies Beach from last data release, the mid-point per anum growth is comparable to that of other suburbs i.e. 9% for Christies Beach and 11%,12%,8% for the other 3 suburbs respectively.
As these projections are based on a forecast algorithm which caries a degree of uncertainty, I suggest that the CG (Capital Growth) values are treated as an estimate of the general trend in the rate of growth and not as an actual percentage value. I’d also suggest that other fundamental metrics are taken into account first and suburbs shortlisted based on fundamentals with the CG projection then applied to asses a potential return on investment for a shortlisted set of suburbs.
Here is a screenshot of other metrics reported for the same suburbs.
Christies Beach passes the low demand / high supply cut off criteria defined in our data dictionary and is in general comparable to other suburbs on the list. However, note a significant difference in the Buy SI and Rent SI. With this in mind, despite the low confidence of the forecast, it is more likely that the CG for Christies Beach will be towards the high end of the range (21%) surpassing CG reported for other suburbs.
As to your question re changes, we are now on HtAG ML Model v2 so the confidence metric and the ranges will not change drastically going forward. Note that changes of up to plus or minus 3% in all reportable metrics are normal between data releases. Changes in low and very low confidence suburbs can be higher than 3%. This is due to the underlying data sparsity at the suburb level.
Although changes are common between data releases, we find that fluctuations in the reported data even out over time. This is another reason why the values should be treated as ballpark ranges and not actual figures.
I hope this clarifies your question. Please don’t hesitate to contact us again for further info.
- You must be logged in to reply to this topic.