A quick initial look (via ‘Dispatch Error’) at the Demand and Frequency drop on Wednesday afternoon 5th August 2026

Linton’s going to invest some time today to delve into some detail of the ‘next day public’ data sets now available from the AEMO (and our own higher-speed frequency readings) to try to piece together an understanding of what happened yesterday – where:

• Linton first noted ‘Major change in NSW ‘Market Demand’ on Wed 5th August 2026 – ‘just’ a data glitch, or more?’
• and then I noted ‘Significant drop in Mainland Frequency (Wed 5th Aug 2026) around the time of what appeared to be ‘just’ a Data Glitch’.

… but curiosity nagged me, so I quickly threw this together to get a head start on understanding some of what happened yesterday.

Utilising a pre-prepared query in the ‘Trends Engine’ within ez2view and shifting the dates (and adding on NSW ‘Market Demand’) we see the following:

Remember:

1)  To click on an image to open a larger-resolution copy in another tab; and

2) That those with a licence to the software can click here to open their own copy of this query in another tab.

Some brief notes:

1)  Those not so familiar with NEMDE or the dispatch process would benefit from the refreshers:

(a)  We published on 5th March 2025 in ‘(A first step into the complexities of) the Dispatch Interval in the Real World’; and then

(b)  In March 2026 we ‘Belatedly noting AEMO’s publication of the ‘NEM Dispatch Timing Fact Sheet’’

… because the exact sequence of how NEMDE works will be important to understanding here.

2)  So I’m wondering if the following was not the sequence of events

 [***** NOTE – HERE BE SPECULATION! *****]

(a)  The NSW ‘Market Demand’ process appears to have gone awry …

i.  Specifically, the Target Demand was stated (imagined?) to be dropping 2,328MW in NSW for the period ending 16:25

ii.  This Target would have been estimated at ~16:20

iii.  Why it might have been imagining a drop is unknown at this point

(b)  What would then have followed, logically, is that:

i.  Dispatch Targets for all units (i.e. Scheduled and Semi-Scheduled units) were resolved to be much lower (collectively) than they had been 5 minutes beforehand

ii.  Which (if the demand drop had been imagined) would have significantly dragged the frequency down

iii.  A mismatch of 2,328MW (or thereabouts) is enormous in the scheme of NEM demand (i.e. roughly 10% of what NEM-wide ‘Market Demand’ was at the time)

iv.  Which means that the frequency drop would have been massive

v.  Noting that we’ve not yet ascertained exactly when the low-point of frequency was, and how low it was (stay tuned to Linton’s following articles)

(c)  But then we see that at the end of the dispatch interval the coal and battery units massively over-performed to 16:25

i.  Again noting this is speculation, but it seems likely that this was due to Primary Frequency Response.

ii.  Remember that (as shown in other Case Studies before) they are the two categories that are heavy hitters

iii.  This would have provided a major support to a rapidly dropping frequency.

iv.  And useful to note the coincidence that Dave Scott (one of the architects of PFR) recently posted here on WattClarity a 4-part series about MPFR and FPP:

> Firstly ‘A recap on the objectives of the reforms for Mandatory Frequency Response, Frequency Performance Payments, and Regulation FCAS cost recovery‘

> Then Part 2 in ‘How MPFR, AGC and FPP-Reg-FCAS Recovery work together’

> Then Part 3 in ‘FPP-Reg-FCAS Recovery a year on: results and behaviours’

> Finally ‘FPP-Reg-FCAS Recovery: to what extent has it succeeded, and could it be improved?’.

v.  it’s too soon to conclude definitively (but certainly seems the most likely scenario) that ‘Primary Frequency Response (and in particular coal and battery units) save the day*!’

* frequent readers will recall that I hate that term ‘save the day’, so read it in those terms.

I could not also help noticing both:

1)  a string of ‘Manifestly Incorrect Inputs’ Market Notices first published from 16:20 (NEM time) on Wednesday 5th August 2026

… which may, or may not, be related – but are certainly coincident.

2)  but also nothing further in the Market Notices that might help to explain

We can see both in this NEMwatch snapshot at 08:00 Wednesday morning 6th August 2026:

Over to Linton now…