下面是一个没有 CASE 声明：
在你拿到你的 wins_per_user 只需数一数 user_ids ，和分组依据 wincount 再一次：

SELECT
  wincount AS WinCount, count(user_id) as UserCount
FROM
  (SELECT 
   count(*) as wincount, user_id
  FROM a 
    where status='win'
  GROUP BY user_id) as wins_per_user
GROUP BY wincount

结果

WinCount UserCount
1        1
2        3

您需要将两级select查询用于 Group By . 子选择查询也称为派生表。
在内部select查询中，获取每个用户的总wins数。
现在，在外部选择中， Group By 在总的胜利，并计算用户为他们。
请尝试以下操作：

SELECT 
  dt.total_wins AS wins, 
  COUNT(*) AS players_count 
FROM 
(
  SELECT
    user_id, 
    COUNT(*) AS total_wins 
  FROM a 
  WHERE status = 'win' 
  GROUP BY user_id
) AS dt 
GROUP BY dt.total_wins

架构（mysql v5.7）

CREATE TABLE a (status VARCHAR(255), user_id int);

INSERT INTO a (status, user_id)
VALUES ('win', 1),
('win', 2),
('win', 3),
('win', 2),
('win', 4),
('win', 4),
('win', 1),
('not win', 1);

查询结果

| wins | players_count |
| ---- | ------------- |
| 1    | 1             |
| 2    | 3             |

db fiddle视图