From Warriors to Knicks, How We Predict the NBA 2018-19

The NBA is back in action this week, which means it's time to boost our NBA prognosis and see what the future of the league could hold. You may be surprised to learn that the 2018 Golden State Warriors and Houston Rockets look good on paper, while the New York Knicks and Sacramento Kings do not.

OK, so it's not really surprising. But in the vast space that separates favorites from favorite players, there are still scenarios and interesting players to watch. And we'll suggest a new way to track these developments throughout the season, after slightly modifying this year's forecast. Scroll the screen down if you want to know all the details, but in a nutshell, our interactive page now contains two projection options: Pure Elo and CARMELO. The former is well known to readers of FiveThirtyEight – it's the simplest version that appears on our franchise history pages – but the latter is more complex. He uses team depth charts and constantly updates players' odds to track the skill level of each team. accounting for trades, injuries and unhappy teammates).

Let's take a look at our pre-season CARMELO projections for each conference, starting with the dominant West:

Favorites: Obviously, the warriors are the big favorites to win the West for a fifth consecutive season, with a 61% chance of returning to the NBA Finals. Of course, the rockets Last year, the United States pushed Golden State to the brink of the playoffs. They added the superstar at one point, Carmelo Anthony, but the Warriors still have a considerable advantage over their talents, even after taking into account the time remaining to win DeMarcus Cousins ​​pickup. Houston will title the next group of Western challengers, however – a group that also includes the surprising Jazz and the stable Thunder (which could be better after giving up Anthony and trading for Dennis Schroder). But who are we laughing at? This is the title of the Warriors to win. Again.

Moving? After improving to 46 victories last year from 40 the previous season, the nuggets could still be on the rise in 2019 – our model predicts 50 wins in Denver. Similarly, Lakers should be the biggest league win this year after hiring an Ohio boy this summer. LeBron James' new team is a strange hodgepodge of veteran spares that may not be confused, but it still has a 84% chance to bring the franchise back to playoffs for the first time since 2013. And despite the loss of Cousins, the pelicans could still play 50 wins after catching Julius Randle of the Lakers (especially if Anthony Davis continues to play as a most useful player).

Decendre? The Western middle class has some candidates to be the least improved team of the conference: first, the Blazers seems to be about to regress after playing over his heads last season, after being overshadowed in the playoffs and having done little to improve during the summer. Also Spurs will test Gregg Popovich's legendary coaching talent with breathtaking lineup that is compounded by recent injuries, such as the ACL rift in goalkeeper Dejounte Murray. Every year, Popovich seems to challenge the unfortunate, but there are limits. And the Mowers could step back in a full season without DeAndre Jordan and Blake Griffin. In addition, we must provisionally place the Wolves of the woods in this category, given the drama involving Jimmy Butler. We still plan to win 45 games after embedding some It's likely that Butler will not speak for them, but that number would drop to 40 wins if Butler actually leaves.

Lottery watch: With Mike Conley back in training and some new faces on board, the Grizzlies could be one of the most improved teams in the league – but that will probably not be enough to put them in the playoff picture. (We give them a post-season probability of 12%.) The other players from the West are even further away: Mavericks, Suns and kings each has less than 4 percent chance of participating in the playoffs. That might change slightly if Dallas rookie Luka Doncic outperforms CARMELO's bullish projection, but it's likely the three teams will battle for the first pick in the draft, with the Kings having an advantage in the race to the bottom.

Favorites: A season after struggling for the Eastern record – which was later knocked down by the playoff Cavs – the Raptors are back at the top of the projected conference hierarchy. Who knows if Kawhi Leonard will stay in Toronto beyond 2019, but for now, he's doing the Raps (which, by the way, will not have LeBron in the conference playoffs for torment them further), favorites to qualify for the NBA Final for the first time. Of course, the Celtics Gordon Hayward and Kyrie Irving, who are coming back from an injury, will have something to say about it. They have in any case tended to outperform this kind of projections under coach Brad Stevens. And the Sixers hope to put a dead season behind them, relying on last year 's escape with more stellar games from Ben Simmons and Joel Embiid. Our model does not expect much from the former first choice, Markelle Fultz, and whatever it provides could give Philly an added advantage.

Moving? the Bucks are an eternal choice to make "The Leap" behind the MVP candidate Giannis Antetokounmpo, although they have not yet prepared everything. Our model always has faith; he calls for a better performance in Milwaukee – and an outside chance of being qualified for the finals of the NBA. Similarly, magicians Always try the potential before finding ways to disappoint, but the health of John Wall and, yes, the passage of Marcin Gortat to Dwight Howard (which always brings a defensive value) makes our model optimistic about the chances of Washington. Finally, the hornets are looking to rebound after two consecutive nightmarish campaigns of 36 wins, and they have the talent to challenge at least a .500 record and a playoff spot in 2019.

Decendre? No team will be stronger in 2019 than the Cavaliers, which could belong to the lottery conversation below after losing LeBron James again via a free agent. In spite of what Tristan Thompson thinks, do not expect a fifth consecutive participation in the final for the Cavs, or at a close level. the Pacers and Heat, two good teams surprised last season, could also see a fall: Miami usually remains in a composition that is hardly better than the average in the league, while Indiana's acquisition of favorite FiveThirtyEight, Tyreke Evans, could not enough to compensate for the regression due to a team that probably outperformed last year. Finally, pistons There is no plan to improve the situation, and this is bad news for a team that missed the playoffs last year despite its all-season rugby performance with Griffin. We give Detroit a little more than the playoff money, but little chance of damage once on the spot.

Lottery watch: The bottom of the east is a disaster area. A couple of bad teams from 2018 (the netting and Magic) are slowly moving towards respectability – if not the fight for the playoffs – but the falcons, bulls and Knicks still one of the worst in the league. New York in particular has the worst rating of CARMELO in the NBA, after taking into account that Kristaps Porzingis could play very little this season (or not at all). After potentially missing out on the playoffs for a sixth consecutive season, the Knicks' best hope might be to return to the icy envelopes of the provisional lottery.

Methodology

For those looking for more details on our new model, welcome! We have not changed much from these pre-season forecasts: they are always based on our CARMELO projections for each player, combined with depth charts for each team, and we still take into account the experience of playoffs in our simulations for the playoffs. The real change will occur at the beginning of the season. We will start updating players' talent ratings during the campaign.

These ratings will be based on the 2018-19 Box Plus / Minus and Real Plus-Minus figures as published by Basketball-Reference.com and ESPN, respectively. Starting with a weight given to our pre-season offensive and defensive plus and minus CARMELO scores, we combine the performance of the current season with player play time for 2018-19 to generate new assessments balancing our pre-season screenings and seasonal performance.

These numbers will feed our depth charts to feed each team's CARMELO rating, a better estimate of the amount of talent on the list at any given time. In the beginning, these will use estimates of pre-season play time, except in the case of major injuries and trades; We are still tinkering with how to handle small transactions and changes in the expected minutes during the season. (Stay tuned for updates on later versions of this new model!) We then use these ratings in a similar setting to that of Elo to simulate the rest of the season and the playoffs.

All this means that we are removing from our old friend the "CARM-Elo" ranking, which was always a delicate compromise between basic Elo and a more complicated system based on depth maps. You may be wondering why we are making this change; The reason lies in the way our seasonal screenings treated the super-dominant teams like the Warriors, which effectively made the regular season meaningless. Although our pre-season forecasts generally give good results in various prognostic competitions, our in-season update process was flawed for a number of reasons: we counted injuries and transactions only indirectly (with the help of party results), and these results could also be very misleading. , especially in the stretch where the central teams rested the players – or just to reserve energy for the playoffs. So our new CARMELO system will be updated in almost the opposite way: it will record the performance of the team indirectly (as the number of players increases or decreases), but will be based on who is on each team – and how well our algorithm thinks they are.

As for our pure Elo predictions, they are based on a two-track Elo system, in which each team simultaneously maintains a "regular season Elo" that moves faster and a "Elo of playoffs" running more slowly. the effect of wins and losses at the end of the season, and we found that it was a little more predictive of the success of the playoffs in the past.

You can find both systems on our NBA 2018-19 predictions page, with CARMELO by default. (But you can switch between the two.) Because it solves many of the flaws of the old method, we believe that our new CARMELO system will improve predictions and give you a better insight into what is happening in the NBA. If you've reached this stage, thank you for reading and we look forward to another season of basketball with you.