Curve Shift

I've been working to resolve issues between PNL and curve shift history. Here is the status.

? The value-at-risk engine pulls curve shift from the source databases. Specifically for power, using the DPR application (ad hoc table), curve shift for Jan 10th comes in at $246K, not the $415MM. Because data isn't stored longer than one week, efforts are centered on restoring required info.
? IT is "scrubbing" code and source tables to ensure proper curve shift is being used.
o Coordination between both gas (ERMS) and power (Enpower) teams are critical. $415MM curve shift day appears to be composed of both gas and power. (Although we are focused on Jan 9, 10th, issue is relevant throughout calendar year term)
? As of this moment, the current curve shift data validates the VAR engine, yet appears to a have couple of incorrect days. These incorrect days do not severely impacting VAR backtesting, however they represent a vastly different pnl.

The end game goal is to have curve shift correctly represented in the value-at-risk backtest data.