It doesn't represent the risk to gerrymandering at all. There are no error bars in voting patterns based on 4 year old data.

Part of the reason this next cycle is looking so bad for Republicans is _because_ of the gerrymandering they tried backfiring on them because they cut the lines too close to try and gain more seats.