This is an article I’ve been intending to write and publish for many weeks (if not months) … but circumstances today prompted me to publish today:

1)  Today was the Brisbane iteration of the ‘Endgame Q3 seminar ‘Expanding the modeller’s toolkit’’ … out of which two things happened I’d like to flag here in this article:

(a)  In particular note Point #2 I’ve just written into these ‘Two of the points that resonated (with respect to the recent SA Wind Lull) in today’s Endgame Seminar’

(b)  Coupled with this was the conversation I had with the attendee sitting next to me, prompted by Oliver’s comment – but noting that I’d been planning this article for a long time.

2)  Also a factor is that next week a few of us will be at the 2026 ACES grand finale in 2026 winter conference season

… and I thought these two charts might prompt some discussion in some pockets of attendees there .

The following two trends:

1)  Provide daily-level data over 2.5 years since January 2024:

2)  But the real focus is the 24 months since the 1st July 2025 implementation of aspects of the ERI Rule Change, and particularly the state-of-charge for individual Large BESS (which were progressively transitioning to BDU via the IESS Rule Change).

3)  And both trends produced via the ‘Trends Engine’ within ez2view.

Trended Daily Stats for Aggregate Battery Operations

In this first trend we set the scene with a number of daily stats:

1)  Regional spot prices (daily average)

2)  Aggregate Large BESS capacity and discharge/charge (daily maximum)

3)  Aggregate Large BESS state of charge (i.e. via aggregate stored energy):

It’s the trend of the 3 x pink bars we want to focus on:

Those with a licence to the software can download their own copy of this query here.

Pay particular attention to the daily ratio!

With the focus on the three pink bars (i.e. which show daily min/ave/max state of charge in MWh … remember that a MWh is a volume and a MW is a rate!), we strip the other 9 series off the trend and add one more to replace:

Those with a licence to the software can download their own copy of this query here

What the trend line shows is the ratio (in percentage) of daily minimum state of charge (in MWh) to daily maximum state of charge (in MWh), that is aggregated across all batteries in the NEM.  So NEM analysts and junkies will understand that:

1)  This is NEM-wide, so takes account of whatever restrictions:

(a)  might have happened with respect to discharge (or charge) due to constraints  (and particularly inter-regional flow limitations) during the day

(b)  and also to do with poor pricing dynamics (which Dan asked about, and Franz followed on in exploration).

2)  And also that the maximums and minimums will typically be boosted/inflated by the ‘energy held in reserve for special purposes’ by batteries contacted to provide SIPS services.

(a)  remember how we showed that the Victoria Big Battery operated at much lower discharge than a novice might have thought they should on 13th February 2024:

i.  because of that particular reason.

ii.  even though load shedding was happening!

(b)  so the key point being the aggregate state of charge overstates how much energy is available because of these SIPS reservations.

3)  Furthermore:

(a)  what’s important to understand is that the percentage is not exactly the ‘percent full’ at the lowest point of the day.

(b)  In fact, if read in that way, the percentage overstates ‘minimum percent full’ across the entire daily fleet (i.e. because the aggregate MWh daily maximum will probably be some amount below 100%).

Pay attention…  does this match the assumptions in the model?

So, if the percentage is overstating ‘minimum daily amount held in reserve for calamity events’, and yet still regularly drops below 20% (across the NEM!), taking a riff from Oliver’s second point here … what does that say about the assumptions made in modelling:

1)  In terms of commercial outcomes; and

2)  In terms of Reliability, and Security of Supply?

For instance … why were we collectively so surprised when we found the batteries in South Australia had run dry on the morning of Monday 22nd June 2026 … when the 2-year long pattern of NEM-wide operations highlighted that this was actually very likely to happen during abnormal events?

If anyone (during the upcoming ACES event) is actually discussing/pondering these challenges, sing out – because I’d be interested to be involved!?