Pulling info from the same table in different orders and quantities using MySQL

Question

First time post. As an example, say I have a table with 12 entries;

+--------+------------+
| name   | date       |
+--------+------------+
| P1     | 2011-12-01 |
| P2     | 2012-01-01 |
| P3     | 2012-02-01 |
| P4     | 2012-02-01 |
| P5     | 2012-02-01 |
| P6     | 2012-05-01 |
| P7     | 2012-04-01 |
| P8     | 2012-05-01 |
| P9     | 2012-05-01 |
| P10    | 2012-06-01 |
| P11    | 2012-06-01 |
| P12    | 2012-03-01 |
+--------+------------+

I have grouped the information within the table by the date, kept a count of all of the id's (not shown) and created a running total. Which I have implemented as so;

(SELECT date, c as count, (@s := @s + c) as run_total 
FROM
(SELECT D1.date, Count(D1.person_id) as c, @s := 0
FROM table D1
GROUP BY month(D1.date)
ORDER BY D1.date) as T1);

Output;

+------------+-------+-----------+
| date       | count | run_total |
+------------+-------+-----------+
| 2011-12-20 |     1 |         1 |
| 2012-01-31 |     1 |         2 |
| 2012-02-20 |     3 |         5 |
| 2012-03-05 |     1 |         6 |
| 2012-04-20 |     1 |         7 |
| 2012-05-20 |     3 |        10 |
| 2012-06-01 |     2 |        12 |
+------------+-------+-----------+

Good so far. This is where I loose it. How can I extract the most recent 4 entries and get something that looks like this??? where the dates remain in ascending order and more importantly the run_total stays the same;

-----What I am trying to achieve-----

----------------------------------
| 2012-03-05 |     1 |         6 |
| 2012-04-20 |     1 |         7 |
| 2012-05-20 |     3 |        10 |
| 2012-06-01 |     2 |        12 |
+------------+-------+-----------+

-------------------------------------

In trying to work this out I created this;

(SELECT date, c as count, (@s := @s + c) as run_total
FROM
(SELECT D1.date, Count(D1.person_id) as c, @s := 0
FROM table D1
GROUP BY month(D1.date)
ORDER BY D1.date desc limit 4) as T1)
ORDER BY date;

But I did not get the desired outcome;

+------------+-------+-----------+
| date       | count | run_total |
+------------+-------+-----------+
| 2012-03-05 |     1 |         1 |
| 2012-04-20 |     1 |         2 |
| 2012-05-20 |     3 |         5 |
| 2012-06-01 |     2 |         7 |
+------------+-------+-----------+

The dates are in order but the run_totals are all messed up. 1 is supposed to be 6, 2 -> 7, 5 -> 10, and 7 -> 12 as shown earlier.

Again, what I am trying to achieve is;

----------------------------------
| 2012-03-05 |     1 |         6 |
| 2012-04-20 |     1 |         7 |
| 2012-05-20 |     3 |        10 |
| 2012-06-01 |     2 |        12 |
+------------+-------+-----------+

Answer 1

SELECT * FROM
(SELECT *
FROM (SELECT date, count, (@s := @s + c) as run_total 
FROM
(SELECT D1.date, Count(D1.person_id) as c, @s := 0
FROM table D1
GROUP BY month(D1.date)
ORDER BY D1.date) as T1) tmp
ORDER BY `date` DESC
LIMIT 4) tmp2
ORDER BY `date` ASC

It takes the 4 most recent records of your query, then reorders them by date.
The totals remain unchanged.

Answer 2

(SELECT date, c as count, (@s := @s + c) as run_total 
FROM
(SELECT D1.date, Count(D1.person_id) as c, @s := 0
FROM table D1
GROUP BY month(D1.date)
ORDER BY D1.date ASC) as T1) LIMIT 4;