Facing the intricacies head-on for price forecasting in the NEM

In financial markets, buyers and sellers interchange financial assets -via online tools- and no physical delivery is required at the end of transactions. However, in electricity markets, sellers are required to deliver electricity physically. Not only that, but electricity markets are also different from financial markets in the way they operate in a daily setting. In day-ahead markets, for each day, market participants buy and sell electricity for the following day and prices and volumes are determined each hourly or half hourly on a daily basis. After day-ahead market closure, auctions in balancing market starts to eliminate real time imbalances in supply and demand in distribution. Further to that, intraday market, bridges the gap between day ahead and balancing market, is a continuous auction that buyers and sellers can trade before delivery. As new technologies and renewable energy opportunities emerge in electricity markets, managing market price risk in electricity day ahead markets have become more challenging since hedging opportunities can fall short in these markets. Therefore, forecasting electricity prices is the foremost important tool for market price risk management in electricity day ahead markets. Moreover, enhancing adoption of smart grids, and demand fluctuations have made it more crucial to forecast price in energy markets.

In the NEM, day-ahead price forecasting and intraday price forecasting correspond to 30 minutes pre-dispatch and 5 minutes pre-dispatch forecasts which are available in AEMO website[1]. However, these forecasts are announced after bidding period closes. For example, 30 minutes pre-dispatch forecast (day-ahead) for day t is made available after bidding period closes for day t at 12.30 in day t-1. However, using publicly available datasets in AEMO website, 30 minutes pre-dispatch and 5 minutes pre-dispatch prices could be predicted accurately 2 hours before the bidding periods are closed, so that market participants could benefit from taking better positions in favorable trade opportunities in their pre-dispatch bids. Last year, Presify has developed a remarkable NEM Price Forecasting module that delivers both 30 minutes pre-dispatch price forecasts and 5 minutes pre-dispatch price forecasts for each region in the NEM. Each forecast is regularly being updated in line with their frequency (30 minutes and 5 minutes respectively).

[1] https://aemo.com.au/energy-systems/electricity/national-electricity-market-nem/data-nem/market-data-nemweb

Presify NEM Price Forecasting module provides accurate forecasts and error metrics could be tracked through widgets on the dashboard. Presify pre-dispatch price forecasts, which are available starting from 20th of November 2022, are 31.25% more accurate, in mean absolute error, than publicly available pre-dispatch prices. Moreover, during this period around 55% of the time, Presify pre-dispatch forecasts are more accurate. Considering this with mean absolute error metric, this shows that Presify NEM Price Forecasting module consistently provides lower error forecasts.

Despite having the accurate point forecasts, these forecasts might not be able to ensure market price risk management totally due to high variations in electricity day ahead markets. Therefore, market participants seek for prediction intervals or probabilistic forecasts rather than point forecasts. Presify NEM Price Forecasting module also could provide probabilistic forecasts as well.

Presify NEM Price Forecasting module also provides complete integration with AEMO datasets.?