Two things that are both true. Near-term power models know the queue and the node, and fade after a few years. Long-run models assume investment will close every gap, and in my experience, they miss next winter. A Zone J battery needs both. Glad 6.1 runs don’t force the choice.

The expensive part of screening a site was never running the model. It comes back weeks later, and you have to rebuild the layers and filters you already had. PowerVAL has saved map views for that. Not glamorous. From what I’ve seen, those features make it a worthwhile tool.

The hardest part of nodal markets is that the hub forecast and the nodal forecast can both be right and still point in opposite directions for the same asset. Robert LaFaso is on a panel on exactly this at Energy Trading Week on Oct. 27. I’d like to hear how others underwrite it.

Early last year, NYISO and ISO-NE both expected winter peaks by the 2030s. ISO-NE’s May forecast now says roughly dual-peaking by 2035. The weather didn’t change, the heat pump credits did. My read is that a single-path Northeast load forecast is now mostly a policy bet.

The question I'd most want answered at Energy Trading Week is whether anyone treats embedded congestion as a structural input yet. My read is that it's the underpriced variable. Load growth sets the level, constraints set your node, and the node is what the trade runs on.

If offshore wind in New York and New England keeps slipping, what fills the gap, and who pays? I don't have a confident answer. I most want to hear what Michael Fisher, Robert LaFaso, and Kolby Koeck at @AscendAnalytics say on Thursday. Link in the reply.

The requested load isn't the forecast load. My colleague Paul Eory told @platts that @ERCOT_ISO's load requests are way beyond what's reasonably feasible. From what I've seen, the hard part now is deciding how much of that queue to believe. I don't think there is a clean answer.

Nodal precision is essential in year 2 and mostly false precision in year 20. I've built forecasts at both ends. The hard call is where on the horizon production cost modeling should hand off to long-run equilibrium. I'm not convinced it's the same year in every market.

Basis is where a lot of projects quietly die. The hub forecast pencils, the node doesn't, and too often nobody sees it until the full valuation is done. My bias after 24 years in this industry is to screen basis on a map before anyone opens a model.

The part of Northeast forecasting I find hardest is when policy risk and project risk move together. When offshore wind slips, targets, costs, and what gets built instead all shift at once. @AscendAnalytics digs into NYISO and @isonewengland on Oct 15, 11 am MT.