It has been four whole weeks since we took a look at the presidential race at the state level, and much has happened in the meantime. Both parties held their national conventions, multiple new scandals have emerged for President Donald Trump, and Joe Biden and Kamala Harris have begun hitting the campaign trail in person.

The race has tightened somewhat in the polls, but not in any dramatic "post-convention bounce" fashion. There are three things at work here, really, and it's impossible to separate them out for each individual poll. The first is that the conventions happened and a whole bunch of people watched them. The ratings for the Democrats were noticeably higher than for the Republicans, for whatever that's worth. The second thing in play is the calendar, as we're now beyond Labor Day, which is when voters traditionally start paying more attention to the race.

But the third thing is more subtle, because as we approach the actual election, the pollsters all switch over from counting "all voters" or "registered voters" to a narrower focus on "likely voters." Who exactly is a "likely voter" is determined by each polling organization, but they use different metrics. Some use past voting behavior as a key (counting, for instance, only voters who report having voted in the past two elections), while others use voter enthusiasm as a yardstick (asking how likely it is that the voter will actually bother to vote, and then only counting those who report they are totally committed to doing so, for example).

Obviously, this means that some pollsters will get it wrong. If, for example, one candidate motivates a whole bunch of voters who have never voted before, then they won't show up in the polls -- but their votes will show up on Election Day. Or a whole bunch of moderately-interested voters may decide to make the effort and vote, again creating a surge in votes that was not predicted by the pollsters. So there are plenty of caveats as we move into the "likely voter" phase of polling.

Presidential races normally tighten in September, and this year is no different. The race has gotten somewhat closer, although the polling has been remarkably stable for the entire contest. No matter what is in the political news, the polls have stayed roughly the same for months now. Joe Biden still leads by 7 or 8 points nationally, and he still leads or is tied in all the battleground states. Donald Trump hasn't flipped any of Biden's states, although he has reduced the margins in a few of the battlegrounds (turning a Biden advantage into a statistical tie). However, he's still so far behind that he's got a long way to go to even have a chance at winning.

Let's take a look at the chart of all the Electoral Votes (EV). As always, we should mention that all of our polling data comes from the incomparable Electoral-Vote.com site. Our first chart shows what percentage each candidate would get of the Electoral College if the election were held today and all the polling proved to be correct. Blue is for Biden, red is for Trump. If blue is above 50 percent, Biden will win. If red is below 50 percent, Trump will win.

[Click on any of theses graphs to see larger-scale versions.]

As you can see, Biden is still beating Trump by roughly 2-to-1, even if the race has gotten tighter. The last one of these columns was written on August 17, which was one day before a rather dramatic shift in the chart, as the next day Texas moved from Biden's column to Trump's. This gave Trump his best standing ever, with 203 Electoral Votes, which adds up to almost 38 percent. Since that point, however, there have been numerous shifts in the battleground states which have resulted in Biden doing better and Trump losing ground. Texas itself shifted from Biden to Trump, then back to Biden, then was tied for a week, and is now barely in Trump's column.

A total of eleven states showed some sort of movement in the past four weeks. Four of these wobbled around a bit and then wound up right back where they started from (as more polling came in). Two of these showed stronger polling for Biden before falling back again (Wisconsin and Nevada). One weakened for Biden before regaining strength (Pennsylvania). And North Carolina bounced back and forth numerous times, from Trump's column to Biden's, then to a tie, then back to Biden, back to a tie, before moving back to Trump.

Biden did improve in three states over the past month, as multiple polls showed he is widening his lead in Minnesota. Ohio, much like North Carolina, is wobbling back and forth between Biden and Trump, but in the last poll Biden showed a small lead. And in Montana, Trump's lead recently shrank a bit.

Trump improved in four states, but only in Texas did this mean the state moved into his column. In the other three (Florida, Arizona, and New Hampshire), Biden's lead has shrunk but still remains.

And one final footnote, we got our first poll for the cycle out of Oregon, which surprised no one by showing a solid lead for Biden.

Throughout all of this, Biden's EV count began the period down a whopping 38 EV (the loss of Texas) to a total of 335. Trump managed for the first time in the race to top 200 EV, since after adding Texas he was up to 203 EV. Trump then lost Ohio (tied) and on the same day saw Texas and North Carolina shift to Biden, which left him back down to 132 EV, while Biden climbed back up to 388 EV. Texas moved back to tied again, and eventually wound up in Trump's column. When all the dust had settled (from all the polls conducted after the conventions), Biden finished the period with 353 EV to Trump's 185 EV. This was an improvement for Trump over the entire period, as he began with 165 EV, to Biden's 373 EV. Percentage-wise, Biden now has 66 percent to Trump's 34 percent of the Electoral College votes, or (to put it slightly differently) exactly a 2-to-1 advantage.

Let's take a closer look at the relative strength of the state-level polling for both of the candidates, beginning with Biden. As always, "Strong" means being up in the polling by 10 points or better, "Weak" covers the ground from 5 to 10 points up, and "Barely" is a lead of under 5 points.

Right after writing the last of these columns, Texas flipped to Trump. You can see this in the big drop in Biden's "Barely" chart line on 8/18. A week later, Texas flipped back to Biden, stayed there a week, and then moved to being tied before returning to Trump. This is the second big bump (both upwards and then down again) in the Barely chart line.

This is an interesting phenomenon, but a healthy dose of caution is required to interpret it. The mere fact that we're talking about Texas as a possible swing state should really be jaw-droppingly astounding. Texas? Blue? Wow. Even the possibility is rather amazing. However, Joe Biden is not likely to actually win the state. If Biden wins Texas, he will have won so many other battleground states that we'll be measuring the size of his landslide against Ronald Reagan's, most likely.

But this could lead to a false sense of confidence, because Texas is the second-largest state by population in the country. As you can see in the chart, when Texas moves, it makes a big difference. Here's where the note of caution is necessary, since even if Biden is up in the Texas polling by one or two points, nobody should count on him actually winning the state until every vote has been counted -- because it skews your perception of how well Biden has to do in all the other states. So Biden's chart line right now -- without Texas -- is a lot more realistic than the sky-high totals he was charting with Texas (marginally) in his column.

Of course, the other chartlines are more important to see the real state of the race. Biden's Strong line has been remarkably steady and remarkably high throughout the entire contest, moving from a low of 213 EV to a high of 258 EV. He is currently at the low end of this range, but that's not all that bad a place to be in, really. At this point four years ago, Hillary Clinton was between 160 and 170 EV in the Strong category. Eight years ago, Barack Obama was struggling to break 200 EV at this point. And in 2008, Obama was down to 130-140 EV right now. So 213 EV is a pretty good spot for Biden to be in, all things considered.

Of more concern is the drop Biden has seen over the past four weeks in the Weak category. Biden began the period with 86 EV in Weak, went up to 90 EV, dropped down to 36 EV, and now finds himself at 56 EV. For the first time in the entire race, Biden's "Strong Plus Weak" line has dipped just below the 270 EV he needs to win the race, after a weak poll in Arizona today puts him at only 269 EV. Again, though, this is not really a cause for panic when you compare where Biden is with where Clinton and Obama were at this point. Here's the "Strong Plus Weak" Democratic comparison chart for the last four presidential races:

This puts things in perspective, as you can see how much better Biden is doing than Clinton or Obama at the same point in the race (Biden's green line drops off to zero right where we are on the calendar, where the data for this year currently ends). Biden had been just above 300 EV right before the conventions, and has now dropped to 269 EV. But please remember, this is just the Strong Plus Weak measurement, which doesn't take into account any of the battleground states.

The two most concerning states right now for Team Biden would have to be Arizona and New Hampshire. Now, New Hampshire is tiny (only 4 EV total), but Arizona is a bit bigger at 11 EV. Arizona started the period in the Weak Biden category, moved up to Strong Biden, but then fell back to Weak Biden and just today all the way to Barely Biden. That's a worrisome trend, if it proves to be true. Nevada is also only in Barely Biden territory, so perhaps it is time to schedule a swing through the Southwest states for the campaign?

For the entire period, Joe Biden is down 4 EV in the Strong category, finishing at 213 EV. In Weak, he dropped even further, from 86 EV down to 56 EV. He did improve in Barely, up from 70 EV to close with 84 EV. In Strong Plus Weak, Biden moved from 303 EV down to 269 EV, and overall Biden lost 20 EV for the period, moving from a grand total of 373 EV down to 353 EV. However, this is still a whopping 83 EV more than he needs to win, to put it into perspective.

So let's move on to see how Donald Trump has been doing. First, his breakdown chart:

When the period started, Texas moved all the way from Barely Biden to Weak Trump, so you can see this clearly on the graph (that big spike upward in the middle range, on 8/18). A week later, Texas softened up for Trump, which caused this spike to fall back down again.

Donald Trump actually lost a bit of ground in his Strong category, as Montana softened up to only Weak Trump. He started with 81 EV in the Strong category and finished with only 78 EV. Trump saw one good poll in Texas and one good poll in Ohio which moved both states briefly into Weak Trump, but later polls moved the states right back out of the column. This bounced the Weak numbers from 39 EV at the start up to 77 EV (with the addition of Texas), but in the end Trump only gained the 3 EV here that he lost from Strong (Montana), finishing up with only 42 EV in Weak column.

Trump's Barely numbers showed a lot more movement, as several of the battleground states wobbled both into and out of the Barely Trump category over the course of the past four weeks. Trump started with 45 EV in Barely, saw this drop down to only 12 EV, but then finished stronger (after Texas moved back to Weak Trump) with 65 EV in Barely.

Overall, Trump's total share of EV rose from 165 EV to 203 EV -- the first time he's been above 200 EV in the entire race -- but then fell sharply back to 132 EV after the loss of Texas. From there, Trump improved steadily to finish with 185 EV overall -- a net gain of 20 EV for the period.

But let's put this in some historical perspective by examining Trump against the last three presidential Republican races, in the Strong Plus Weak category:

Here, too, Trump hit a high point, breaking 150 EV for the first time (with Texas as Weak, he was at 158 EV for a week). By doing so, he briefly beat his own numbers from 2016, which is notable because this is the first and only time so far that Trump has bested any of the previous three Republican candidates' numbers. In 2016, Trump was consistently above 150 EV, while this time he's only topped that milestone once. Trump hasn't seen any convention bounce here, which helped propel both Mitt Romney and John McCain over 200 EV. So by historical standards, Trump is lagging badly at this point in the race.

Even if all the battleground states currently in Trump's column voted for him, he's still a whopping 85 EV away from the goal of 270 EV. So while the race has tightened in several states, Trump is still struggling to have much of any path to victory at all.

My Picks

Which brings us to the "gut feeling" part of the program, where I attempt to sort the states into the categories of "Safe," "Probable," and "Leaning" for both candidates, while leaving a few in the "Too Close To Call" tossup category. As usual, checking out Electoral-Vote.com's individual state graphs is crucial to see how much polling is being conducted as well as overall trends (rather than just taking the most recent poll as gospel). Also, there are full lists of states at the bottom of this column, so you can see where I've placed them all (together with their EV totals).

Likely States -- Biden

Safe Biden (18 states, 213 EV)
No change at all. Every Safe state for Biden remains Safe. Oregon finally conducted their first poll of the election, but it held no surprises as Biden shows a big lead.

Probable Biden (4 states, 56 EV)
There was some movement here, as New Hampshire moved down to only Lean Biden, but Minnesota moved up to take its place. Michigan continues to look pretty strong for Biden, so I'm leaving it as Probable Biden once again.

The other two states wobbled a bit, but in the end I decided to leave both Pennsylvania and Wisconsin here for now. Pennsylvania had one poll showing a much closer race than all the others, but subsequent polling has shown that this was likely an outlier and not an indication of a trend. This will bear close watching over the next few weeks, though.

Wisconsin likewise seemed to get a little closer, as Trump has made a big push for the state (both Trump and Biden visited in the past few weeks). But again, the polling still shows Biden with a fairly healthy lead here, so for now I'm keeping it in the Probable Biden column.

Likely States -- Trump

Safe Trump (12 states, 72 EV)
No change here at all. Mostly this is due to a lack of polling from these states, but even if more polls had been posted they likely would still have shown Trump with large leads.

Probable Trump (7 states, 48 EV)
No change here at all, either. Last time we moved Kansas and South Carolina here from Safe Trump, and eventually they might move back up, but I'm going to have to see some more polling before that happens. Montana weakened for Trump, but I can't see moving it down to Lean Trump quite yet, so it also stays here for now. The other states in the Probable Trump category are: Alaska, Missouri, Tennessee, and Utah.

Tossup States

Lean Biden (3 states, 21 EV)
There was a lot of shuffling in the Lean Biden category, in all directions. New Hampshire moved down to Lean Biden from Probable Biden, while Minnesota moved back up to Probable. Florida moved back to the Too Close To Call category as the polling there got a lot closer, which essentially wiped out Biden's Florida lead (for now, at any rate). Trump has never actually led in Florida, but he's getting closer so for the time being it has to be considered a tossup.

Nevada remains quite close, but there is always a dearth of good polling out of the state. Biden shows a small lead in the polls which have been released, and the state is getting bluer each year, it seems, so for now it has to be seen as Leaning Biden.

Arizona was looking so strong that I thought I'd be moving it up to Probable Biden, but a poll posted today showed a much closer race than all the others. This may be an outlier, since other polling showed a double-digit lead for Biden only a few weeks ago, so we'll have to wait and see what happens in the next few weeks. But for now, it stays as only Lean Biden.

Lean Trump (1 state, 6 EV)
No change here. Arkansas stays as the only Lean Trump state.

Too Close To Call (6 states, 122 EV)
There was one big addition to the Too Close To Call category, as Florida now moves back into tossup status. None of the other five states (Georgia, Iowa, North Carolina, Ohio, Texas) showed any sort of trending lead for either candidate. In almost all of them, the polling has bounced the states from Biden's column to Trump's and back again in a slow-moving game of ping-pong. Because of the two 800-pound gorillas in this category (Texas, with 38 EV, and Florida with 29 EV), the total for the Too Close To Call category is a whopping 122 EV.

Final Tally

As usual, the race is tightening a bit after Labor Day and the conventions. Whether this is a result of actual shifts among voters or merely just more people paying attention is hard to determine as of yet. Also, the shift by the pollsters to counting only likely voters is certainly a factor as well.

Neither Joe Biden nor Donald Trump got much of a convention bounce to speak of. Perhaps this was a result of the conventions being right next to each other in the calendar, the unprecedented virtual nature of the conventions, or just the voters already having made up their minds. Again, it's tough to tell, really.

Trump improved marginally over the past four weeks, but not dramatically so. The race has so far been incredibly stable -- much more so than previous races have been. Joe Biden still holds a huge lead over Donald Trump, no matter which way you measure it. In just the Strong states, Biden is at 213 EV to Trump's 78 EV. In Strong Plus Weak, Biden is at 269 EV while Trump again has less than half this number, at only 120 EV. Biden would only need to win one state from his Barely category to win the race, while Trump would need every state in his Barely category as well as all the states in Biden's Barely category just to tie Biden in the Electoral College. The paths to victory for Biden are many, while almost no path to victory exists for Trump. Overall, Biden leads Trump by 353 EV to 185 EV.

If the polling is correct, Biden has to win only one state from the following list to win the Electoral College: Arizona, Florida, Georgia, Iowa, Nevada, North Carolina, New Hampshire, Ohio, and Texas. Any of those states going for Biden means he wins the race. Winning more than one state will merely expand Biden's margin of victory. So you have to ask yourself which is more likely: Biden will win at least one from that list, or Trump will win all of them?

So even though the race is marginally tighter, Joe Biden is still in excellent shape. He's doing better than Hillary Clinton or Barack Obama did at this stage of the race, while Trump is doing worse than his own 2016 run and John McCain and Mitt Romney.

Overconfidence is always a dangerous thing, of course, as was proven beyond a shadow of a doubt in 2016. But the 2020 race is looking extremely good for Joe Biden, from where I sit.

[Full Data:]
(State electoral votes are in parenthesis following each state's name. Washington D.C. is counted as a state, for a total of 51.)

Joe Biden Likely Easy Wins -- 22 States -- 269 Electoral Votes:

Safe States -- 18 States -- 213 Electoral Votes
California (55), Colorado (9), Connecticut (7), Delaware (3), Hawaii (4), Illinois (20), Maine (4), Maryland (10), Massachusetts (11), New Jersey (14), New Mexico (5), New York (29), Oregon (7), Rhode Island (4), Vermont (3), Virginia (13), Washington D.C. (3), Washington (12)

Probable States -- 4 States -- 56 Electoral Votes
Michigan (16), Minnesota (10), Pennsylvania (20), Wisconsin (10)

Donald Trump Likely Easy Wins -- 19 States -- 120 Electoral Votes:

Safe States -- 12 States -- 72 Electoral Votes
Alabama (9), Idaho (4), Indiana (11), Kentucky (8), Louisiana (8), Mississippi (6), Nebraska (5), North Dakota (3), Oklahoma (7), South Dakota (3), West Virginia (5), Wyoming (3)

Probable States -- 7 States -- 48 Electoral Votes
Alaska (3), Kansas (6), Missouri (10), Montana (3), South Carolina (9), Tennessee (11), Utah (6)

Tossup States -- 10 States -- 149 Electoral Votes:

Tossup States Leaning Biden -- 3 States -- 21 Electoral Votes
Arizona (11), Nevada (6), New Hampshire (4)

Tossup States Leaning Trump -- 1 State -- 6 Electoral Votes
Arkansas (6)

Too Close To Call -- 6 States -- 122 Electoral Votes
Florida (29), Georgia (16), Iowa (6), North Carolina (15), Ohio (18), Texas (38)

Polling data gaps:

Polled, but no recent polling data -- 8 States
(States which have not been polled since the beginning of August, with the dates of their last poll in parenthesis.)

Alaska (7/24), Arkansas (6/10), Connecticut (5/4), North Dakota (3/5), Maryland (5/23), Mississippi (2/28), Tennessee (5/22), Washington (7/27)

No polling data at all, yet -- 11 States
(States which have not been polled so far this year.)

Hawaii, Idaho, Illinois, Louisiana, Nebraska, Rhode Island, South Dakota, Vermont, Washington D.C., West Virginia, Wyoming

-- Chris Weigant

Follow Chris on Twitter: @ChrisWeigant