# Without Next Data — October 26, 2026

Sensitivity view: all 42 Next Data surveys are omitted, and the model is recalibrated using the remaining polls. Direct Polls remains included. This is an alternative calculation for October 26, not a finding that excluding a pollster improves accuracy.

This download contains the current optional scenario only. The published r44 forecast and its frozen history are separate and unchanged. Both R and D use the same 100,000 joint seat simulations; R uses weight_R, while D uses weight_D. Use probabilities, not unweighted D counts. Missing values remain unknown. Source observations remain visible even when omitted from this scenario; model_inputs and input_membership identify the observations actually used.