Why predicting this is difficult — and why BESS indices lead astray
An energy storage system, both in the arbitrage model and in the model with system services, looks on paper like a money-making machine. You charge during price troughs, discharge during peaks – and collect the difference. But in practice, this equation has many more variables than just energy prices. Predicting actual arbitrage revenues is a difficult task, inherently prone to error – yet too many investment models today are based on simplified indices that show unrealistically high profits because they assume perfect market prediction.
At VESS, we believe that investments in energy infrastructure should be based not on optimistic assumptions, but on reliable analyses. Our mission is to show the most probable reality, not the most lucrative one. Because a well-informed investor is a prepared investor – and only such an investor will be able to maintain a project in a volatile and increasingly competitive market.
Why is predicting energy prices so difficult?
It might seem that electricity prices can be predicted with great accuracy – after all, we have historical data, weather forecasts, information about renewable energy production or demand. But even the best predictive algorithms regularly make mistakes. Why?
- Energy markets are stochastic – they are governed by chance. Even if we predict that a day will be sunny, we don’t know exactly how much energy each PV farm in the country will produce. Nor which thermal unit will suddenly fail.
- Prices are extremely sensitive to short-term disruptions – small changes in production or demand can cause price jumps of hundreds of zlotys per MWh.
- Forecasts are always imperfect – and prediction errors are inevitable. The problem is that even a small error in the forecast can mean that the storage “shoots” with the charge/discharge cycle at the wrong hour.
- The market is changing – what worked yesterday won’t necessarily work tomorrow. The growing number of storage systems is changing the way prices form.
BESS indices and the illusion of perfection
Indices such as BDARI (BESS Day-Ahead Revenue Index) are based on the assumption that storage perfectly hits the price trough and peak every day. In other words – the storage knows exactly when to charge and when to discharge every morning.
This means that indices show the maximum possible revenue, assuming 100% decision effectiveness (perfect price prediction), which in reality occurs very rarely, and certainly never if we’re talking about a 10-15 year investment and accurate predictions on each of 5 thousand days.
Equally important, indices only show the current situation! In a nascent market, such as where energy storage systems are just emerging, basing your investment, which will only start operating in a year or even several years, on current data is definitely misleading. Of course, this doesn’t mean we’ll lose on the investment, but we’ll earn drastically less than we assumed.
Are indices therefore useless?
They are super useful, but not for investments. An operationally active storage must earn and must earn as much as possible. Indices allow us to determine how much a storage could have earned in a given month if we had perfect forecasts. This comparison allows the storage owner to see how well the optimizer is working. It will never be 100%, but there is a significant financial difference between 80% and 85%. However, there is one of two requirements that must be met for such comparisons to make sense: 1. The storage must be operational, 2. The market in the next 5-10 years must be similar to the current state. In the Polish market, we don’t have a multitude of storage systems, which means even more that it won’t be a stable market with the assumed investment outlays in storage in Poland. The conclusion is simple: it won’t be as good as it is now.
VESS: Analysis of reality, not optimism
At VESS, we create our own financial models and analyses of BESS projects, based on variant analysis and large-scale simulations. We don’t analyze dozens or hundreds of scenarios – we analyze hundreds of thousands of variants:
- meteorological scenarios
- price models
- trading strategies
- investment assumptions
- degree of competition and grid load
- degree of advancement of other renewable energy investments
Our goal is not only to show the most probable scenario, but above all to assess the risk of it not coming true. We show the investor:
- what is the chance that the result will be better or worse than the baseline scenario
- what the distribution of revenues looks like over time and what conditions it
- what operational decisions can reduce risk and which increase it
And what is crucial, we update all our calculations every month, because in such a dynamic market, no prediction is permanent. Thanks to this, the mentioned confidence interval of risk begins to narrow, and the investor is not surprised by the final result of the investment, because it was available on an ongoing basis.
Summary
There’s nothing wrong with using indices as a benchmark. But treating them as the basis for investment decisions is like planning a household budget based on the maximum lottery win. At VESS, we show investors what will happen when the market doesn’t play as we would like. And how to build an operating model that will be resilient, flexible, and profitable despite this.
Because an energy storage is not a financial instrument – it’s an operational asset. And its value lies not in the promise of profit, but in the ability to survive market reality.
Sponsored article

Maciej Konieczka
CEO of VESS, Expert in Energy Analytics and Storage. An experienced expert in the energy sector, specializing in the use of data analytics, artificial intelligence, and machine learning (ML) to optimize energy management. He is the founder and CEO of VESS (Virtual Energy Storage Systems) – an innovative technology company that develops advanced algorithms for managing energy storage systems, maximizing their profitability and operational efficiency. Before founding VESS, he served as the Director of Data and Analytics at Veolia Energy Contracting Poland, where he was responsible for data management, analytics, and modeling optimization strategies for energy markets and energy storage systems. Previously, he also worked at PGE S.A. Capital Group, PKN Orlen, and TRMEW Obrót (now Respect Energy), where he developed data-driven strategies and implemented Data Science solutions in energy trading, capacity market analyses, and technical forecasting for power plants. He is a visiting lecturer at Kozminski University and a PhD candidate at Warsaw University of Technology, where he conducts research on advanced data analytics and process modeling in the energy sector. He holds a DAMA CDMP certification in data management.