Welcome to Upset Watch
If you're new to Pickwatch, or have never read Upset Watch, here's what we do:
With over 35% of NFL games ending in an underdog outright win, it's important to know when upsets are most likely to occur, and which teams to pick. Last season, we had our third straight winning season, with a combined winning total of over +$8,000 in that time.
Starting every Wednesday, we select a number of underdogs that have a statistically higher chance of winning or covering than others. Often, these are teams that nobody in the media is picking.
To do this, we use historical trends and a proprietary weighted ranking of each team's current 'true' performance, to assess how well they are playing, and generate score predictions that help us make straight up, ATS and points total picks.
Contents (click to jump to section - Pro subscription required for picks):
2023 Upcoming Changes to Upset Watch
Upset Watch has been a part of our offering on Pickwatch for a long time. So long, in fact, that it seems odd that it has never been given the technical development it warrants.
With Pickwatch renewing our focus on the basics this season, we also wanted to finally give Upset Watch the love it deserves too. We know that 90% of Pro members (if not more) join almost entirely because of Upset Watch, and it's a no-brainer to make some changes on that basis.
That said, as ever, I was keen that it can't just be change for change's sake. It has to improve things. Luckily, there were some very easy wins for us, and they'll be launching over the next week or so.
New layout
Upset Watch has, pretty obviously, outgrown being a simple blog post. There are formatting and scale issues that I run up against every week, and from your perspective, I don't think it's easy to read one giant blog post to find the right part that matters most to you.
That's why from next week, Upset Watch will become it's own section of the site. You'll be able to quickly navigate using a side menu, so if you just want the picks, you can easily do so with a single click. Similarly, if you are just checking a single game, instead of scrolling through looking for it in a wall of text, you can just click the link on the left hand side. Here's a quick layout example we are working from:
In addition, each game will have clearly defined sections for each team, and a clear picks section at the bottom of the page that gives you the straight up (money line), ATS, Points Total, and best player prop picks for that game.
The new layout will be with us next week, and we'll be making changes based on your feedback once it launches.
New Data
Apart from making Upset Watch 10x easier to use, I was also keen that some of the data I use be available for you to explore every week.
That means that for now, starting with 10 core stats per team, we will make our weighted stats available for you to see as a full table. That means you can see which teams rank highest in 'true' performance in key areas such as pass rushing or defensive efficiency.
Long tiome readers will know that this ability to quantify a performance is the key to our success. We approach every game with the mindset that either team can win in theory, and we can use our weighted stats to find teams that have been playing better than their records suggest. Now, you can too.
Oh, a note here, some of you LOVE the header images that compare units on our pages. I do too, but putting them together was laborious and slow. From next week, those images will no longer be done manually, instead they will be calculated automatically and you can see them as early as Tuesday, once we move the week on.
Confidence Picks
I've more or less been doing confidence picks for some tijme, but from this point, I'll be adding them into their own dedicated section. I hate it when people select one confidence pick and say 'that's my lock of the week' - it doesn't take into account that half of the readers will already have used that team.
Instead, I'll be giving a full ranking list, and also highlighting where I think teams are good value as an unfancied confidence pick, or favorites to avoid.
More pick types
As mentioned, I am going to be including my points total and player prop picks this season. Last season, we added points total over/under picks to Pickwatch, and lots of you love them. We have a full scale user contest with a $1,000 top prize, and weekly prizes of $100 for winners, plus Pro and Essential membership up for grabs.
Some of you may also know that I had to stop making my Points total picks last season, because I was leading the contest by a considerable margin. This was unintended, I was simply testing my algorithmic score predictions, and this year, I'm going to do the same in Upset Watch.
Our other new contest comes in the form of our Player Prop contest. Get the most profit each week and you can win the same prizes as our other contests, including $1,000 for the season winner. To help you along, I'll be using our algorithms to highlight some prop picks for you - albeit I encourage you to try making your own too, or following the better users on here.
The future
My first priority is making life easier for you. I think all of these changes do exactly that, and retain all of the old functionality, while expanding on it to whatever degree you are comfortable. Apart from the page layout changing slightly (for the better) you will see no difference in how I write or what I'm writing.
Up next? I'm keen to offer a 'schedule' view for each team, listing their weighted performance. This is very much what I use to quantify teams and understand whether they are getting better or worse, and where their strengths lie.
To do that, I'm also keen to give you the flexibility to include and exclude data at your discretion. I can give you a starting point, but if - for example, you wanted to see a team's performance filtered by certain conditions (Examples: who the Head Coach was, what the weather was like, whether the opponent had over X % passing attempts on average going into the game), we can do that.
This kind of flexible, easy to read, useful table has long been my dream, and we're close to being able to do it now. The key will be whether we can make it user friendly, and not only appeal to people like myself who can spend hours a week researching silly stats. Instead, I want it to be something like our main tables (on a much bigger scale) where you control what you see each week, and how it works.
Anyway, that is for the future, and depends heavily on how the next few months go for Pickwatch, but my final thing is to say thanks to every one of you Pro users. We have tens of thousands of people paying to use Pickwatch, and without you, as I mentioned last week, it'd be pretty tough, if not impossible, for us to do what we do.
Why underdogs and not favorites?
Because underdog winners yield bigger profits from the greater odds you face, you don't need to win every pick you make. Our motto is that with Upset Watch, you make more profit from less picks, with our average selection being +150. That means that with a win % of just 41%, you'd make a profit.
Since we began tracking our picks in the current format, our profits have been consistently good each season, as we add more datapoints and improve our methodology.
If you'd placed $100 on every pick we tipped over the last 3 years, you'd have won:
2020 profit: +$1,800
2021 profit: +$3,200
2022 profit: +$2,900
That's why people love us. The nature of what we do makes it hard. You won't win every pick, but the ones you do win will mean you return a bigger profit, and if you use consistent methods, they will do so more often.
Our weighted stats in the header for each game are different to the 'raw' stats you can find on other websites.
Weighted stats indicate the performance of a team against the averages of their opponents. For example, if a team scores 30 points against a team that gives up 29 points on average, their weighted performance will be +1. If they score 30 points against a team that gives up 20 points on average, their performance is +10, because they scored 10 points more than that opponent normally allows.
Over the course of a season, we can see the balance of the weighted performance as a 'true' measure of a team's real caliber, and also a team's recent performances, and whether they indicate that an upset is on the cards soon.
A team that ranks highly in any weighted category generally has a good performance in this statistic regardless of how strong their opponents are.
About the stats
I use around 40 stats to collate Upset Watch, but above each game, you'll see an overview of both teams across the major areas of the game - points, passing yards, running yards, Pass protection, and the corresponding defensive stats.
By placing the units side-by-side with the unit they are up against (eg: Pass protection vs Pass rush) you can not only see where a team is strongest, but also whether the opposing unit is likely to be strong or weak in that area.
Lastly, we also have an algorithm that uses weighted stats to generate predictions of yards and points for each team. These predictions have proven incredibly accurate over the last few years, and they're improving all the time as we add more data into them.
Last season was another great one for us. In total, we had only one week where we lost more than 11% of what was staked (excluding the Super Bowl as that is all or nothing), and of course, by that time, we were already very much in a position of strength.
As always, we tend to get better as the season goes on. In the early phase of the season there are opportunities though, as the odds and forecasts for each team tend to be based on last season. Of course, we factor in the previous year when making our picks in week 1, but we analyze each team to see improvements and regression. As always, we keep an open mind, as many of our best picks come from teams that people assume will be bad based on nothing more than their previous season record.
I get asked this question a lot, about the best way to use the information in Upset Watch.
I've transferred this section of Upset Watch to a separate article, as I attempt to make the weekly column easier to read, so if you want to see my tips for how to maximize profit alongside many thousands of other subscribers, check this link out.
To read the picks, you'll need a Pickwatch Pro subscription. Get 20% off today with the code VETERAN
