 On Wednesday I went to Twitter and I tweeted out my top 10 running backs and wide receivers based on what my projections are saying. Now as I said in the tweet I emerged from my mom's basement with finished projections so I decided to share what those projections said. And as is usually the case, I got crushed. One person was mad at me that both Vikings receivers weren't in the top 10. Now I can understand that to a degree I guess and the projections currently have them as top 15 wide outs. So it's not like they're far off. Another account quote tweeted the list and said that I was part of the national media who seemed to underrate T.Y. Hilton. T.Y. Hilton for what it's worth is the wide receiver 11. But here's where it got weird. There was also a group of Packers fans who were relentlessly hitting my mentions about Aaron Jones. And it was because Aaron Jones liked my tweet so people saw it. And somehow of all the players in the NFL people were calling me an Aaron Jones hater. What a world we live in. One thing I noted when I tweeted what the projections spit out was that those projections are not my rankings. And that confused people a bit. I actually got an email from Eric W that I think brought up a very valid question. It said, you posted some projections and said these weren't rankings. For draft and analysis purposes, are we better off using the projections alone without any human bias? People talk about how we overestimate our ability to predict things, especially in fantasy. Are we better off just simply taking the projections data and basing decisions and rankings off that or are you adjusting ranks based on upside or other soft factors that I would argue might not actually be beneficial? So let me explain why I don't view projections and rankings as the same thing. Making projections can be done in a lot of different ways. The way I create them is through what I guess I would call a top-down approach. Based on a good number of inputs, I get team information. So the number of plays a team will run, the number of passes, the number of rush attempts, those types of data points. Once those are locked in, all of the player projections feed into those team projections. That way you're not just creating projections without any sort of context. It's all fitting together within what that team is actually going to do. This is why rushing share and target share then becomes important. You can take the rushing and passing volume and then split up rushing attempts and targets based on market share. To get those rushing and target share numbers, there will be some form of subjectivity. Like, we know certain stats are sticky year over year and we can find trends based on historical data. It's not like we're just throwing random target share numbers out there that have no backbone. But there's still some guesswork being done. And that's why it's a projection. It's not called a certainty. For me personally, the process of building out projections each season isn't necessarily to get a list of players to draft from. Because my projections are going to be wrong, just like my rankings are going to be wrong. I hope they're more right than what others are working off of, but it would be foolish for me to think that there's zero variance with the projections that I build. Not only that, but projections are a mean outcome. They're what would be expected to happen. At the end of Eric's email, he says, or are you adjusting ranks based on, quote, upside or other soft factors that I would argue might not actually be beneficial? So with something like upside, which gets thrown around a lot, I actually think it does matter. It matters quite a bit. What upside is basically telling us is a player's realistic maximum ceiling. And when we're projecting something, we're only getting a player's realistic average. And when we're looking at how many points a player is going to score across a season, that floor and ceiling equation matters. And let's not confuse this with how a player is scoring those points in a given week. Like you most definitely can argue that Tyreek Hill's up and down scoring last season didn't actually matter given what he did during his good weeks. But when entering a season, the floor and ceiling equation is different. You're talking about how many points these players are going to end up scoring, not how they're scoring them. And with some players, there's more certainty. We're certain that Deandre Hopkins is going to see a large target share and be productive. We're certain that Ezekiel Elliott and Saquon Barkley will have massive workloads. There's less certainty for Levy on Bell and Antonio Brown. They're on completely new teams. That means their range of outcomes shifts because we know less about their situation. It could mean that they have more upside due to ambiguity, but it could also mean that they have a much lower floor. So my projections will look at the average outcome. They'll say, this is what the expectation is for Antonio Brown this season. But I may look at that and say, I know that the wide receiver position is deep through round four. So if I spent an early round pick on Antonio Brown, I can strive more for ceiling because there are solid backup plans. I actually think this is even more true when you look at lower tiered players. Like the difference between how I rank my top 20 at each position isn't going to be drastically different than the way that you do. What's different is how we approach the later rounds. And I'm not talking about draft strategy. I'm talking about who we pick because in the middle and the late rounds, you really should be striving for upside. But how I view a player's path to upside may be drastically different than the way that you do. As it pertains to projections though, a projection isn't going to use its imagination like your own brain will. A projection last season wasn't going to tell you to draft James Conner, but we know that there were other circumstances that may have led you to getting James Conner. We know there was at least some chance that Levy on Bell wouldn't play for at least a game or two. So you may have bumped Conner up in your rankings while your projection remained the same because you're looking more at upside and you're telling yourself a story to that upside. And that story isn't captured in a projection. That's why rankings and projections don't have to match. And that's why the two shouldn't match.