The Senate probability actually ticks down by a fraction due to GOP chances dropping in NH, KY, and CO to counter increasing numbers in KS, AK, and AR.
The most common result in my current simulations is GOP leading 51-49. If I give KS to the GOP (either Roberts or Orman caucusing with the GOP), the GOP Senate probability jumps to 58%. The reason that the outlook is still hovering around 50% is because there are 6 seats in which the GOP has a 50-80% chance of victory (favored, but not strongly) and 1 seat which the Independent has a 62% chance of victory. If any of these seats get shifted to 100%, the probability becomes upper 50's for that party.
As a note, no forcast is really significant to anymore that 5% intervals. In all reality, 10% is a better bar to set for significance. This is why television broadcasts list precipitation chances in round numbers. I have never seen a news broadcast say "Today we have a 49% chance of rain." I would caution anyone to think of anything 41%-59% as anything other than basically a toss-up. While, you could say one party or another has a slight edge after 55%, there is still a 45% chance that the other party wins the Senate. This is why I give both the overall probability (and state by state), and a "snapshot" of the current tally. The snapshot is taking the most likely scenario (regardless how large of a lead) in each race. The tally adds up each one of these races taken individually.
Next week, PPP will give is polls out of Kansas and Idaho. Yes, you read that right. Idaho. I guess we will see if they find anything to the new-found money that has been spent on the Gubernatorial race there. I really don't expect them to find anything of note in Idaho. In a race that is barely polled, there is always the chance for a shock, but I just don't see it.