By Dean Oliver, 2024-10-09 20:00:00
由生成式人工智能翻译，译文内容可能不准确或不完整，以原文为准。
数据分析的一大功能就是预测未来。这就像一种没有被拍成电影的超能力。
ESPN 在 NBA 领域的数据分析超能力就是篮球实力指数 (BPI)。这种超能力并非完美，但它预测未来的能力确实比那些在联盟通行证上不停切换比赛的普通球迷要强。它会分析球队的历史表现、阵容变化以及赛程安排，并根据球员的临场发挥来调整对球队表现的预期。
例如，BPI 认为孟菲斯灰熊队上赛季表现不佳，是因为贾·莫兰特(Ja Morant) 只打了 9 场比赛，而伤病缠身的球队给了很多本赛季不会得到那么多上场时间的球员大量出场时间。本赛季灰熊队的预计首发阵容（莫兰特、德斯蒙德·贝恩(Desmond Bane)、小贾伦·杰克逊(Jaren Jackson Jr.)、马库斯·斯马特(Marcus Smart) 和扎克·埃迪(Zach Edey)）在上赛季总共只打了 137 场 NBA 比赛，平均每人 27 场，而这恰好是灰熊队上赛季的胜场数。经验丰富的老将们回归了，而埃迪在夏季联赛中看起来像是一位年度最佳新秀的候选人，因此 BPI 认为他们在竞争激烈的西部联盟中有 69% 的概率进入季后赛。
BPI 难以预测的是伤病，我们所有人也是如此。过去的伤病是未来伤病的最佳预测指标，因此灰熊队可能会再次遭遇伤病困扰，BPI 对此持谨慎态度。这也是包括灰熊队在内的顶级球队往往会回归到 41 胜的原因之一；BPI 认为本赛季没有任何一支球队有可能赢得 60 场比赛的胜利。
对于上赛季排名垫底的球队，BPI 还假设他们会认真比赛而不是摆烂……呃，进入发展模式。因此，BPI 对布鲁克林篮网队、波特兰开拓者队和华盛顿奇才队等球队的预测高于大多数预测系统，因为它认为这些球队会让本·西蒙斯(Ben Simmons) 或博扬·博格达诺维奇(Bojan Bogdanovic) 这样的球员而不是年轻的诺亚·克洛尼(Noah Clowney) 上场。或者，就奇才队而言，让马尔科姆·布罗格登(Malcolm Brogdon) 而不是二年级前锋比拉尔·库利巴利(Bilal Coulibaly) 上场；或者让杰拉米·格兰特(Jerami Grant) 而不是斯科特·亨德森(Scoot Henderson) 在开拓者队上场。
但无论如何，BPI 会随着赛季的进行而调整。
介绍完这些，让我们来回答 BPI 可以解决的一些热门话题。
谁将赢得总冠军？
上赛季，BPI 预测密尔沃基雄鹿队拥有最佳夺冠概率，达到 21%。它看好扬尼斯·阿德托昆博(Giannis Antetokounmpo) 和克里斯·米德尔顿(Khris Middleton) 的锋线组合以及达米安·利拉德(Damian Lillard) 的后场实力。但它无法完全预料到教练组上演的肥皂剧（它其实不怎么看肥皂剧）。特别是阿德托昆博，他质疑老帅阿德里安·格里芬(Adrian Griffin)，直到后者被解雇，而新帅也不比老帅强。这对他们没有帮助，BPI 不会忘记这一点。
今年，BPI 预测凯尔特人队有 26% 的概率卫冕 NBA 总冠军，这是联盟最高的概率。它非常看好他们，并希望那里不会出现肥皂剧。波士顿上赛季每个位置上都人才济济，那些在前一年对教练乔·马祖拉(Joe Mazzulla) 的决定说三道四的人在球队一路高歌猛进的过程中也不得不保持沉默。BPI 并没有明确看到马祖拉脑子里想的东西，但它看到了足够多的东西，并认为他们将为波士顿带来第 19 座总冠军奖杯。
打进 NBA 总决赛的最佳机会
根据 ESPN BPI 预测*
| 球队 | 夺冠概率 | 打进总决赛概率 |
|---|---|---|
| 1. 波士顿凯尔特人队 | 25.9% | 38.6% |
| 2. 俄克拉荷马城雷霆队 | 12.5% | 24.0% |
| 3. 费城 76 人队 | 12.5% | 22.0% |
| 4. 达拉斯独行侠队 | 10.3% | 20.5% |
| 5. 克利夫兰骑士队 | 5.7% | 11.7% |
| 6. 明尼苏达森林狼队 | 4.8% | 11.1% |
| 7. 丹佛掘金队 | 4.6% | 11.0% |
| 8. 密尔沃基雄鹿队 | 4.2% | 9.3% |
| 9. 迈阿密热火队 | 3.8% | 8.1% |
| 10. 孟菲斯灰熊队 | 2.6% | 6.9% |
| 11. 菲尼克斯太阳队 | 1.6% | 4.6% |
| 12. 萨克拉门托国王队 | 1.6% | 4.1% |
| 13. 洛杉矶湖人队 | 1.5% | 4.0% |
| 14. 新奥尔良鹈鹕队 | 1.4% | 3.8% |
| 15. 洛杉矶快船队 | 1.3% | 3.7% |
| 16. 休斯顿火箭队 | 1.2% | 3.3% |
不过，凯尔特人队也有一些强劲的竞争对手。费城 76 人队已经变得更好了，但全明星中锋乔尔·恩比德(Joel Embiid) 能否保持健康（上赛季仅出场 39 场比赛，出场时间为 1,309 分钟）始终是一个问题。如果恩比德无法保持健康，那么泰瑞斯·马克西(Tyrese Maxey) 和保罗·乔治(Paul George) 的二人组也能提供相当大的安全保障。（我们这些构建 BPI 来预测未来的人仍然对其预测恩比德健康状况的能力持怀疑态度。我们将在后续部分详细讨论。）
在西部联盟，BPI 非常看好俄克拉荷马城雷霆队及其才华横溢的年轻核心阵容：谢伊·吉尔杰斯-亚历山大(Shai Gilgeous-Alexander)、杰伦·威廉姆斯(Jalen Williams) 和切特·霍姆格伦(Chet Holmgren)，他们上赛季表现出色，今年应该会更好。此外，他们还从纽约尼克斯队签下了中锋以赛亚·哈尔滕施泰因(Isaiah Hartenstein)，后者在尼克斯队教练汤姆·锡伯杜(Tom Thibodeau) 的简单教导下表现出色。ESPN BET 认为雷霆队的夺冠概率仅次于凯尔特人队，BPI 也同意这一点。
在模拟赛季开始之前，BPI 已经确定排名前 18 的球队中有 12 支来自西部，这意味着凯文·杜兰特(Kevin Durant) 领衔的菲尼克斯太阳队和勒布朗·詹姆斯(LeBron James) 领衔的洛杉矶湖人队进入季后赛的概率都只有五成左右。目前，森林狼队、掘金队、独行侠队和灰熊队的 BPI 季后赛预测排名都高于他们。
杰森·塔图姆(Jayson Tatum)、杰伦·布朗(Jaylen Brown) 和波士顿凯尔特人队是卫冕 NBA 总冠军的热门球队。布莱恩·巴比诺/NBAE via Getty Images
独行侠队和他们的新三巨头能重返总决赛吗？
拥有卢卡·东契奇(Luka Doncic)、凯里·欧文(Kyrie Irving) 和克莱·汤普森(Klay Thompson) 的独行侠队是六支拥有三名入选过最佳阵容球员的球队之一。
拥有三位（或更多）球星确实有帮助。回顾过去二十年，拥有四名或更多最佳阵容球员的球队平均每个赛季能赢下 50 场比赛。拥有三名最佳阵容球员的球队平均能赢下 47 场比赛。拥有两名最佳阵容球员的球队平均能赢下 46 场比赛。拥有一名最佳阵容球员的球队（例如上赛季拥有吉尔杰斯-亚历山大的俄克拉荷马城雷霆队，他们在西部联盟中赢下了 57 场比赛）平均能赢下 39 场比赛。没有最佳阵容球员的球队平均只能赢下 34 场比赛。
这些平均数据大体上支持了这一观点，但其中存在着重要的变数。在像独行侠队这样拥有三名最佳阵容球员的球队中，有 57% 的球队没有突破季后赛首轮（包括上赛季的太阳队）。
但独行侠队不一样，对吧？东契奇还年轻，欧文仍然是我们见过的终结能力最强的后卫，而汤普森将会获得并命中很多空位投篮的机会。幸运的是，BPI 看到了这些球员的实力和发展趋势。本赛季，BPI 预测独行侠队拥有第四高的夺冠概率，达到 10%。这个百分比看起来可能不高，但为了便于理解，金州勇士队在 2022 年夺冠时拥有三名最佳阵容球员，而他们那个赛季的夺冠概率只有 2%。
需要说明的是，76 人队今年有四名入选过最佳阵容的球员（除了恩比德和乔治，替补球员安德烈·德拉蒙德(Andre Drummond) 和凯尔·洛瑞(Kyle Lowry) 是你可能忘记的两位）。除了独行侠队和太阳队，其他拥有三名最佳阵容球员的球队还有凯尔特人队、明尼苏达森林狼队和萨克拉门托国王队。
我们如何看待阵容重组的尼克斯队和森林狼队？
如果你还没听说的话，卡尔-安东尼·唐斯(Karl-Anthony Towns) 现在效力于纽约尼克斯队。朱利叶斯·兰德尔(Julius Randle) 和唐特·迪温琴佐(Donte DiVincenzo) 现在效力于明尼苏达森林狼队。
在这笔交易之前，BPI 认为尼克斯队的未来一片光明，很有可能获得季后赛前六的席位。
然后 BPI 看到这笔交易，并认为它对森林狼队有利。这两支球队的季后赛概率都有所上升，但森林狼队的季后赛概率从 65% 左右上升到了 80%，而尼克斯队在东部的季后赛概率只上升了一点点，达到了 65%。森林狼队的夺冠概率从 4% 上升到了 5%，而尼克斯队的夺冠概率从 3% 下降到了 1%。BPI 并不认为尼克斯队的前景有多么光明。
原因是 BPI 认为唐斯并不比兰德尔或迪温琴佐强多少。森林狼队得到了两名优秀的球员，他们每个赛季能提供 4,000 分钟的出场时间，而尼克斯队只得到了一名优秀的球员，他在过去五个赛季中平均每个赛季只能出场 1,700 分钟。是的，我们知道，BPI 下次出现在麦迪逊广场花园时可能会被嘘下场。
预测 NBA 最佳（和最差）进攻和防守球队
在现代 NBA 这个充满技术型射手的时代，防守需要大量的跑动和沟通。如果沟通不畅，那就需要更多的跑动。现在的球场空间拉得很大，以至于油漆区经常是空的。每个人都在投篮。即使是 33% 命中率的三分球，其价值仍然高于大多数中距离投篮。这意味着三分球的随机性无处不在。有些晚上，他们会投进。
有些晚上，凯尔特人队会像他们在季后赛第二场对阵骑士队的比赛中那样，投出 35 记三分球，命中 8 球，然后输掉 24 分。然后，末日预言者们会批评马祖拉几天。BPI 承认三分球的这种随机性，并预测进攻和防守之间的差距会缩小。要么是这样，要么是它在试图弄清楚哪些球队的投篮在本赛季能够持续命中。
如果是前者，那么 BPI 排名前五的进攻球队将是凯尔特人队、雷霆队、76 人队、独行侠队和步行者队。凯尔特人队、步行者队和雷霆队上赛季都在这个行列，而 76 人队则凭借乔治的加盟进入了第一梯队。独行侠队则凭借其三名最佳阵容球员和几名令人生畏的大个子球员入选。
进攻预测
根据 ESPN BPI 预测*
| 球队排名 | 进攻 BPI |
|---|---|
| 最佳球队 | |
| 1. 波士顿凯尔特人队 | 3.0 |
| 2. 俄克拉荷马城雷霆队 | 2.4 |
| 3. 费城 76 人队 | 2.3 |
| 4. 达拉斯独行侠队 | 2.1 |
| 5. 印第安纳步行者队 | 2.0 |
| 最差球队 | |
| 26. 奥兰多魔术队 | -2.4 |
| 27. 波特兰开拓者队 | -2.6 |
| 28. 底特律活塞队 | -2.7 |
| 29. 圣安东尼奥马刺队 | -3.0 |
| 30. 华盛顿奇才队 | -4.2 |
在 BPI 排名垫底的五支进攻球队中，活塞队、开拓者队和马刺队都榜上有名。奇才队上赛季险些上榜，但他们交易走了队内三分球命中率最高的球员泰厄斯·琼斯(Tyus Jones) 和三分球命中率第三高的球员德尼·阿夫迪亚(Deni Avdija)，换来了可以抵挡所有打铁声的篮筐。
在防守方面，BPI 预计森林狼队和雷霆队将重返前五。灰熊队拥有小贾伦·杰克逊和马库斯·斯马特，而且他们对健康的乐观态度也让他们进入了前五。76 人队也要感谢乔治的加盟。骑士队上赛季没有进入前五，但他们在前一个赛季排名第一，所以本赛季的排名也在情理之中。
在 BPI 排名垫底的五支防守球队中，没有鲁迪·戈贝尔(Rudy Gobert) 的犹他爵士队再次上榜，而奇才队和老鹰队预计也将继续排在这个梯队。步行者队进入了垫底五名，部分原因是教练里克·卡莱尔(Rick Carlisle) 似乎已经接受了这支球队喜欢快节奏的事实……以及随之而来的后果。他们当然可以做得更好。
然后是活塞队。新教练 J.B. 比克斯塔夫(J.B. Bickerstaff) 帮助骑士队建立了良好的防守体系，我认为年轻球员们在防守端会做出积极的回应，但 BPI 认为我错了。实际上，BPI 并没有考虑教练的因素，但它仍然认为我错了。
防守预测
根据 ESPN BPI 预测*
| 球队排名 | 防守 BPI |
|---|---|
| 最佳球队 | |
| 1. 明尼苏达森林狼队 | 2.2 |
| 2. 俄克拉荷马城雷霆队 | 1.9 |
| 3. 孟菲斯灰熊队 | 1.8 |
| 4. 费城 76 人队 | 1.6 |
| 5. 克利夫兰骑士队 | 1.5 |
| 最差球队 | |
| 26. 印第安纳步行者队 | -1.4 |
| 27. 亚特兰大老鹰队 | -1.7 |
| 28. 华盛顿奇才队 | -1.8 |
| 29. 犹他爵士队 | -1.9 |
| 30. 底特律活塞队 | -2.1 |
BPI 对可能影响库珀·弗莱格争夺战的球队的看法
所有人的目光都将集中在那支获得状元签并有机会选中杜克大学球星库珀·弗莱格(Cooper Flagg)（目前 ESPN 百强榜排名第一的球员）的球队身上，但还将有三支球队参与到另一场对乐透签位的争夺中。
圣安东尼奥马刺队拥有 2025 年选秀大会的两个首轮签——来自亚特兰大老鹰队的无保护首轮签和来自芝加哥公牛队的受前六顺位保护的首轮签。尽管看好维克多·文班亚马(Victor Wembanyama)，但 BPI 并不认为马刺队本赛季会取得好成绩（预计胜场数只有 31 场）。但马刺球迷可以感到安慰：你可以关注另外两支球队，看看 2025 年休赛期会有多热闹。
不过，老鹰队有很大的动力去争取好成绩。球队管理层可不想看到一个高顺位选秀权溜走。他们更愿意看到一个低顺位选秀权溜走。球队管理层有这种动力并不意味着球员们也认同。这就是一支有凝聚力的球队需要从老板到管理层再到教练和球员都齐心协力的地方。在一支特雷·杨(Trae Young) 的性格并没有让他赢得其他球员喜爱的球队里，老鹰队可能很难找到这种程度的凝聚力。
我希望我能说动力是 BPI 明确考虑的一个因素，因为它会影响到每个赛季五场或更多比赛的胜负。它在一定程度上反映了这一点，因为在过去几年里，杨的积极和消极影响都体现在了他的数据中。目前，BPI 认为老鹰队可能会以 37 胜的成绩结束常规赛，进入季后赛的概率约为 34%。它甚至预测老鹰队欠马刺队的选秀权会落在第 11 顺位，老鹰队应该把这作为一个合理的预期，而马刺队也会对此感到满意。
公牛队的情况则更加微妙。如果公牛队在全明星赛前后保持着他们通常 35-39 胜的节奏（这也是 BPI 目前的预测），那么他们可能会在赛季末进入全面发展模式（或者你想怎么称呼它）。无论 BPI 有多聪明，它都无法读懂公牛队管理层的想法，尤其是在距离做出这样的决定还有几个月的时候。目前，BPI 认为他们有 20% 的概率成为排名垫底的六支球队之一，但乒乓球可能会随机地将他们的选秀顺位推迟到第七位，这对马刺队来说至少是件好事。
数据中的痛点：BPI 如何分析伤病
BPI 实际上会根据球队的不同阵容版本进行评分，它会使用所有球员都健康的完整阵容版本。它还会对一支球队可能的季后赛阵容进行评分，这个阵容接近完整阵容，但会考虑一些可能缺阵的球员。然后还有一个比赛级别的版本，例如，新奥尔良鹈鹕队在锡安·威廉姆森(Zion Williamson) 可能缺阵的比赛中会派哪些球员上场。这个版本对灰熊队和 76 人队来说尤其重要，实际上对任何一支命运都取决于一两名球员的球队来说都很重要。
在每场比赛中，BPI 都会试图区分完整阵容和实际参赛球员对比赛结果的影响程度。灰熊队上赛季因伤缺阵的比赛超过 400 场，相当于每场比赛都有五名球员缺阵。76 人队的恩比德在赛季末遭遇半月板受伤后，已经准备好参加附加赛和随后的季后赛，BPI 预计到了这一点并将其考虑在内。
不幸的是，它仍然没有预料到威廉姆森会在鹈鹕队的附加赛中受伤。
预测恩比德、威廉姆森或洛杉矶快船队球星科怀·伦纳德(Kawhi Leonard) 这样重要但容易受伤的球员的常见伤病并不是 BPI 特别擅长的领域，但这很重要。球员们可能会受伤并错过重要的比赛，但 BPI 错过了这些信息。
鹈鹕队球星锡安·威廉姆森和 76 人队中锋乔尔·恩比德都希望在本赛季保持健康。蒂姆·恩瓦丘库武/Getty Images
黑马？别忘了骑士队
当我查看 BPI 的预测结果，并试图了解它没有看到的联盟动态时，我能想到的最大惊喜就是克利夫兰骑士队。他们的首发阵容中有三位全明星球员（多诺万·米切尔(Donovan Mitchell)、达柳斯·加兰(Darius Garland) 和贾莱特·阿伦(Jarrett Allen)），还有可能迎来突破的球星埃文·莫布里(Evan Mobley) 和马克斯·斯特鲁斯(Max Strus)。他们在东部半决赛中与凯尔特人队交手，米切尔缺席了两场比赛，阿伦则缺席了整个系列赛，但他们仍然给对手带来了很大的压力。
骑士队本休赛期最引人注目的变化是在 J.B. 比克斯塔夫执教了四个赛季后，聘请肯尼·阿特金森(Kenny Atkinson) 为主教练。有些教练的变动可能会带来糟糕的结果，有些则会带来美好的结果。密尔沃基雄鹿队上赛季的教练变动就带来了糟糕的结果，但我认为骑士队在阿特金森的带领下经历的早期阵痛最终会带来美好的结果。根据 BPI 的预测结果，我会选择他们作为最有可能爆冷进入 NBA 总决赛的球队。（而且 BPI 对他们夺冠概率的预测高于大多数博彩公司。）
技术细节
BPI 经过多年的构建和改进，拥有我见过的最复杂的球队实力评分算法。这个经过改进的新版本参考了过去几个赛季的数据。我喜欢它根据球员的出场情况对球队进行不同级别的评分，以及它复制了近年来进攻效率不断提高的总体趋势。它根据球员的出场情况对球队获胜概率的预测也相当复杂。
但这并不意味着它没有缺陷，也不意味着它会比拉斯维加斯的博彩公司更厉害。我当然可以说，BPI 中有些东西会在一定程度上进行修改。它没有考虑到所有信息，而且它对球员价值的估计是基于 Basketball-Reference.com 上使用的正负值等数据统计，而这些数据统计本身就存在缺陷。我们计划逐步用更好的球员指标来取代它。
点击查看原文：NBA Basketball Power Index: Projections, biggest questions for 2024-25
NBA Basketball Power Index: Projections, biggest questions for 2024-25
One thing that analytics can do is see the future. It’s like a superpower without the movies made about it.
ESPN’s analytics superpower for the NBA is the Basketball Power Index (BPI). The superpower isn’t perfect, but it is better at seeing the future than the average guy doing clicker finger reps on League Pass. It sees how good teams were before, how they have changed, and who they have to play. It adjusts to the teams that overperformed underperformed, based on the players in the game.
For example, the BPI understands the Memphis Grizzlies underperformed last year with Ja Morant playing just nine games and the injury-riddled team giving plenty of minutes to guys who won’t get them this season. This season’s projected starters (Morant, Desmond Bane, Jaren Jackson Jr., Marcus Smart and Zach Edey) played a total of 137 NBA games last season, an average of 27 per player, which is coincidentally how many games Memphis won last season. The experienced guys are back, and Edey looked like a Rookie of the Year candidate in Summer League, so the BPI sees them with a 69% chance of making the playoffs in the very crowded Western Conference.
What the BPI struggles to foresee – as we all do – is injuries. Past injuries are the best projection of future injuries, so it could happen again in Memphis, and the BPI is a little leery of it. That’s one reason why top teams – not just Memphis – tend to regress toward 41 wins; the BPI doesn’t see any team likely to win 60 games this season.
Among the teams that finished at the bottom last season, the BPI also assumes that they actually compete and not tank … err, go into development mode. As a result, the BPI forecasts teams like the Brooklyn Nets, Portland Trail Blazers and Washington Wizards higher than most projection systems, thinking they’re actually going to play guys like Ben Simmons or Bojan Bogdanovic over youngster Noah Clowney. Or, in the Wizards’ case, Malcolm Brogdon over second-year forward Bilal Coulibaly; or Jerami Grant over Scoot Henderson in Portland.
But in either case, the BPI will adapt as the season progresses.
With that introduction, let’s answer the top party topics that the BPI can address.
Who’s gonna win it all?
Last season, the BPI picked the Milwaukee Bucks to have the best title odds, with a 21% chance to win the championship. It liked the talent,ed combination of forwards Giannis Antetokounmpo and Khris Middleton with guard Damian Lillard. But it couldn’t fully account for the coaching soap opera that played out (it doesn’t really watch soap operas). Antetokounmpo, in particular, questioned the old boss, Adrian Griffin, until he was let go, and then the new boss was no better than the old boss. That didn’t help them, and the BPI doesn’t forget.
This year, the BPI is giving the Celtics a league-high 26% chance to repeat as NBA champions. It really likes them and is hoping for no soap operas there. Boston had talent at every position last year, and the people who got on coach Joe Mazzulla’s case the year before for his decisions had to stay silent through a dominant run. The BPI doesn’t explicitly see the details running through Mazzulla’s brain, but it sees enough, and thinks they will lead to a 19th banner raised in Boston.
Best chance to make the NBA Finals
*according to ESPN BPI
| Team | Win Title | Make Finals |
|---|---|---|
| 1. Boston Celtics | 25.9% | 38.6% |
| 2. Oklahoma City Thunder | 12.5% | 24.0% |
| 3. Philadelphia 76ers | 12.5% | 22.0% |
| 4. Dallas Mavericks | 10.3% | 20.5% |
| 5. Cleveland Cavaliers | 5.7% | 11.7% |
| 6. Minnesota Timberwolves | 4.8% | 11.1% |
| 7. Denver Nuggets | 4.6% | 11.0% |
| 8. Milwaukee Bucks | 4.2% | 9.3% |
| 9. Miami Heat | 3.8% | 8.1% |
| 10. Memphis Grizzlies | 2.6% | 6.9% |
| 11. Phoenix Suns | 1.6% | 4.6% |
| 12. Sacramento Kings | 1.6% | 4.1% |
| 13. Los Angeles Lakers | 1.5% | 4.0% |
| 14. New Orleans Pelican | 1.4% | 3.8% |
| 15. LA Clippers | 1.3% | 3.7% |
| 16. Houston Rockets | 1.2% | 3.3% |
There are legitimate competitors to the Celtics, though. The Philadelphia 76ers have gotten better, but whether All-Star big man Joel Embiid can hold up (39 games for just 1,309 minutes played last season) is always a question. The duo of Tyrese Maxey and Paul George also provides a considerable safety net if he doesn’t. (Those of us who built the BPI to see the future are still a bit skeptical of its ability to predict the impact of Embiid’s health. More on that in a subsequent section.)
In the Western Conference, the BPI loves the Oklahoma City Thunder and their talented, young core of Shai Gilgeous-Alexander, Jalen Williams and Chet Holmgren, who were good last season and should be even better this year. Plus, they picked up center Isaiah Hartenstein from New York, who did his job very well under the simple teachings of Knicks coach Tom Thibodeau. ESPN BET has the Thunder second to the Celtics in championship odds, and the BPI agrees.
Before the BPI simulates the season, it has already determined that 12 of the top 18 teams by talent are out West, which means teams like Kevin Durant’s Phoenix Suns and LeBron James’ Los Angeles Lakers remarkably face coin-flip chances as to whether they make the playoffs. Right now, Minnesota, Denver, Dallas and Memphis all have better BPI projections of making the playoffs.
Jayson Tatum, Jaylen Brown, and the Boston Celtics are favorites to repeat as NBA champions. Brian Babineau/NBAE via Getty Images
Can the Mavericks and their new Big 3 get back to the Finals?
The Mavericks – with Luka Doncic, Kyrie Irving and Klay Thompson – are one of six teams with three players who’ve made an All-NBA team in their careers.
Having three stars (or more) does help. Looking back over the past two decades, teams that entered a season with four or more All-NBA players averaged about 50 wins a season. Teams with three averaged 47 wins. Teams with two averaged 46 wins. Teams with one – such as last season’s Oklahoma City squad with Gilgeous-Alexander that had a West-leading 57 wins – averaged 39 wins. Teams that didn’t have an All-NBA player averaged only 34 wins.
Those are averages that generally support the idea, but there is important variability. Of teams like Dallas, with three All-NBA players, 57% of them didn’t get past the first round of the playoffs (including the Suns last season).
But that’s not Dallas, right? Luka is young, Kyrie is still the best finishing guard we’ve ever seen, and Klay will see and make a lot of open looks. Fortunately, the BPI sees what these guys are and where they’re trending. This season has the Mavs with the fourth-best chances at a title with 10% of the hypothetical futures. That percentage may not seem like much, but for perspective, Golden State won the title in 2022 with three All-NBA players when their chance that season was only about 2%.
For the record, Philly has four players this year who have made All-NBA Teams (along with Embiid and George, reserves Andre Drummond and Kyle Lowry are the ones you forgot). Along with Dallas and Phoenix, the other teams with an All-NBA trio are the Celtics, Minnesota Timberwolves and Sacramento Kings.
What do we make of the revamped Knicks and Timberwolves?
If you somehow missed it, Karl-Anthony Towns is now with the New York Knicks. Julius Randle and Donte DiVincenzo are now Minnesota Timberwolves.
Before the trade, the BPI saw a pretty good future for the Knicks with a good chance of a top-six seed in the playoffs.
Then the BPI saw the trade. And liked it … for Minnesota, at least. Both teams saw their playoff chances increase, but the Wolves’ playoff chances went up from approximately 65% to 80% in the West, while the Knicks edged up a hair to 65% in the East. The Wolves’ title odds went from 4% to 5%, whereas the Knicks went down from 3% to 1%. The Knicks’ high-end future actually doesn’t seem so high to the BPI.
The reason for that is that the BPI doesn’t see Towns as dramatically better than Randle or DiVincenzo. The Wolves got two good players who provide a good 4,000 minutes per season, whereas the Knicks got one good player who has averaged about 1,700 minutes over the past five seasons. Yes, we know, the BPI is probably going to get booed out of Madison Square Garden the next time it shows up.
Projecting the NBA’s best (and worst) offenses and defenses
In the modern NBA world of skilled shooters, defense is a lot of running and communication. More running if there’s poor communication. Now the floor is often so spread out that the paint is wide open. Everyone is shooting. The value of even a 33% 3-point shot is still higher than most midrange shots. What this means is that the randomness of the 3-point shot is pervasive. Some nights, they’re gonna fall.
And some nights, the Celtics are going to shoot 35 of them, make eight, and lose by 24 to the Cavaliers, like they did in Game 2 of their playoff series. Then doomsayers will then criticize Mazzulla for a couple of days. the BPI acknowledges this randomness in 3s and projects that offenses and defenses are going to be a little closer together. Either that, or it’s throwing up its hands trying to figure out which teams’ shots are going to consistently fall this season.
If it’s the former, the BPI’s top five offensive teams project to be the Celtics, Thunder, 76ers, Mavericks and Pacers. The Celtics, Pacers and Thunder were all there last season, but Philly enters the top echelon courtesy of George. And Dallas is there with its three All-NBA guys and a couple of rim-scaring big men.
Offensive Projections
*according to ESPN BPI
| Team Rank | Off. BPI |
|---|---|
| BEST TEAMS | |
| 1. Boston Celtics | 3.0 |
| 2. Oklahoma City Thunder | 2.4 |
| 3.Philadelphia 76ers | 2.3 |
| 4. Dallas Mavericks | 2.1 |
| 5. Indiana Pacers | 2.0 |
| WORST TEAMS | |
| 26. Orlando Magic | -2.4 |
| 27. Portland Trail Blazers | -2.6 |
| 28. Detroit Pistons | -2.7 |
| 29. San Antonio Spurs | -3.0 |
| 30. Washington Wizards | -4.2 |
For the BPI’s bottom five offensive teams, Detroit, Portland and San Antonio all repeat. Washington just missed it last season, but it traded away Tyus Jones, its top 3-point shooter, and Deni Avdija, its third-best 3-point shooter, in return for reinforced backboards to withstand all the clanking.
On defense, the BPI projects Minnesota and Oklahoma City to return to the top five. Memphis has Jaren Jackson Jr. and Marcus Smart, and the optimism of health that puts it in the top five. Philly can thank George again for its inclusion here, too. Cleveland didn’t make the top five last season, but it was No. 1 the season before, so this season’s appearance makes sense.
For the BPI’s worst defensive teams, the Rudy Gobert-less Utah Jazz make another appearance, while the Wizards and Hawks similarly project to repeat in this tier. Indiana enters the bottom five, partially because coach Rick Carlisle has seemingly accepted that this group loves going fast and … to heck with the consequences. They could be better, for sure.
And then there’s Detroit. New coach J.B. Bickerstaff helped construct a good defense in Cleveland, and I think the young guys will respond initially on that end, but the BPI thinks I’m off. Actually, the BPI doesn’t acknowledge the coaches, but it still thinks I’m off.
Defensive Projections
*according to ESPN BPI
| Team Rank | Off. BPI |
|---|---|
| BEST TEAMS | |
| 1. Boston Celtics | 3.0 |
| 2. Oklahoma City Thunder | 2.4 |
| 3.Philadelphia 76ers | 2.3 |
| 4. Dallas Mavericks | 2.1 |
| 5. Indiana Pacers | 2.0 |
| WORST TEAMS | |
| 26. Orlando Magic | -2.4 |
| 27. Portland Trail Blazers | -2.6 |
| 28. Detroit Pistons | -2.7 |
| 29. San Antonio Spurs | -3.0 |
| 30. Washington Wizards | -4.2 |
What the BPI sees for teams who could impact the Cooper Flagg watch
All eyes will be on which team lands the coveted No. 1 pick for the chance to potentially take Duke standout Cooper Flagg (the current top-ranked prospect on ESPN’s Big Board), but there’ll be another race involving three teams with major lottery implications.
The San Antonio Spurs are owed two first-round picks for the 2025 draft – an unprotected pick from the Atlanta Hawks and a top-six-protected selection from the Chicago Bulls. The BPI doesn’t project the Spurs to do well this season (expected win total of just 31), despite loving Victor Wembanyama. But Spurs fans can take solace: You can watch those two other teams to see how fun the 2025 offseason will be.
Atlanta has a huge incentive to be good, though. The brass doesn’t want to see a high draft pick fly away. The Hawks would rather see a low draft pick fly away. The fact that the organization has that incentive doesn’t mean that the players share it. That’s where a cohesive organization needs everyone playing for each other, from ownership to management to coaches to players. On a team where Trae Young’s personality hasn’t exactly endeared him to others, that level of cohesion might be difficult for the Hawks to find.
I wish I could say motivation was explicitly a factor for the BPI because it matters on the order of five or more wins per season. It is somewhat reflected in there because Young has had his share of both positive and negative influence show up in his stats the past several years. For now, the BPI sees Atlanta likely finishing with 37 wins and about a 34% chance of making the playoffs. It even projects the draft pick the Hawks owe the Spurs landing at No. 11, which Atlanta should set as a reasonable expectation, and something that the Spurs will be happy enough with.
Chicago is in a more precarious situation. If the Bulls are on pace for their usual 35-39 wins around the All-Star break (which is what the BPI projects now), they could go into full-on development mode (or whatever you want to call it) to finish the season. No matter how smart the BPI is, it can’t read the minds of the Bulls organization, especially months before such a decision is likely to be made. For right now, it sees them with about a 20% chance of being among the bottom six teams, but the pingpong balls could randomly push their pick back to No. 7, which would be fun for the Spurs, at least.
Pains in the stats: How the BPI analyzes injuries
The BPI actually separates out what version of a team it rates, using the full-strength version in which no one is injured. It also has a rating for a team’s likely playoff roster, which is close to full strength but factors some players who may be out. And then there is a game-level version such as, for example, who the New Orleans Pelicans are putting out on the floor on the nights Zion Williamson might be out. This version is especially important for the Grizzlies and 76ers, too – really any team whose fortunes rely strongly upon the talents of one or two players.
With every game, the BPI tries to separate how much of the performance is associated with the full-strength squad vs. who actually played. Memphis had over 400 games missed because of injury last season, which is roughly five players per game. Philadelphia had Embiid ready to return for the play-in and the subsequent playoffs after his late-season meniscus injury, which the BPI anticipates and accounts for.
Unfortunately, it still didn’t anticipate Williamson getting injured in the Pelicans’ play-in game.
Forecasting the usual injuries to important but injury-prone players such as Embiid, Williamson or LA Clippers star Kawhi Leonard is not something that the BPI does particularly well, but it matters. Players will likely get hurt and miss important games, but the BPI has missed that chatter.
Pelicans star Zion Williamson and 76ers big man Joel Embiid are looking to remain healthy this season. Tim Nwachukwu/Getty Images
Surprise contender? Don’t sleep on the Cavaliers
As I look at the BPI’s projections and get a feel for what’s happening around the league that it doesn’t see, the biggest surprise I can see coming is the Cleveland Cavaliers. Their starting five includes three All-Stars (Donovan Mitchell, Darius Garland and Jarrett Allen), along with potential breakout stars Evan Mobley and Max Strus. They pushed the Celtics in the conference semifinals without Mitchell for two games and Allen the whole series.
The Cavs’ most notable change this offseason was hiring Kenny Atkinson as the coach after four full seasons under Bickerstaff. Some coaching changes can go sour and some sweet. Milwaukee’s change last season went sour, but I think that early Cavs growing pains with Atkinson will ultimately end up sweet. And I’d pick them, from the BPI’s output, as the most reasonable outside-the-box NBA Finals contender. (And the BPI likes their championship odds more than most gambling houses.)
Technospiel
The BPI has been constructed and modified over several years and has the most sophisticated algorithm I’ve seen for a power rating. This new and improved version looked at past seasons to learn from them. I like how it rates teams at varying levels of player availability and that it replicates the general trend of increasing offense through the years. Its projection of a team’s chance to win a game in light of player availability is also quite intricate.
But that doesn’t mean that it’s without flaws or that it’s going to beat Vegas. I can certainly say that there are things in the BPI that will get revised to some degree. It doesn’t incorporate every piece of information and it estimates player value based on a box score statistic like plus/minus used on Basketball-Reference.com, which has its flaws. We are planning on gradually replacing that with better versions of player metrics.
By Dean Oliver, via ESPN