Batch:
DROP TABLE IF EXISTS tmpRandomLRP;

CREATE TEMPORARY TABLE tmpRandomLRP AS
SELECT DISTINCT
    path,
    subsong,
    [%jsp3_loved%], 
    COALESCE(NULLIF([%jsp3_last_played%], ''), '1970-01-01 00:00:00') AS [%jsp3_last_played%]
FROM (
    SELECT
        first_value(path) OVER w_album path,
        first_value(subsong) OVER w_album subsong,
        [%jsp3_loved%], 
        COALESCE(NULLIF([%jsp3_last_played%], ''), '1970-01-01 00:00:00') AS [%jsp3_last_played%]
    FROM MediaLibrary
    WINDOW w_album AS (PARTITION BY album ORDER BY random())
) AS subquery
ORDER BY random();

Query:
WITH CTE AS(
 SELECT path, subsong, [%jsp3_loved%], COALESCE(NULLIF([%jsp3_last_played%], ''), '1970-01-01 00:00:00') AS [%jsp3_last_played%]
 FROM tmpRandomLRP
)
SELECT path,
       subsong
FROM CTE
WHERE ([%jsp3_loved%]) AND ([%jsp3_last_played%] < datetime('now', '-300 minutes'))
ORDER BY [%jsp3_last_played%]
LIMIT 2

Ah, guess it's not worth the trouble...?

I'm not sure how feasible it is as I'm still a beginner with SQL queries...

Temp tables in batch part just so that the results are randomised? and then in the query part apply my filter conditions?

I have some tracks which have the same title but from different albums, to rule out this confusion would it be better to have the Group as %path% - %subsong%

I think the only real solution is to make it possible to configure the padding around the artwork separately. I'll consider it...

Definitely not easily. It will become quite complex and probably you will need to create temporary tables in the batch part of the sql definition at least for performance reasons.

No. Besides that you probably meant %title%, it would only rule out, that you get both tracks together from the set of tracks, which you defined with your SQL (as long as %title% for both tracks is different).

Sorry, I failed to understand exactly for the Second Case... just so that we are on the same page I've made two pools one as Least Recently Played and the other as Lowest Play Count, I've shared my settings for Least Recently Played and I wanted to know if it's possible to combine the query you suggested to prevent tracks from the same album playing...

For the group should I change it to %track%? or what would be the best scenario here to set for the Group parameter?

You mentioned not to use CTE as well, but I found out that if I don't include that in "SELECT path, subsong FROM" my COALESCE doesn't take place and replace the NULL values inside the database.

COALESCE(NULLIF([%jsp3_last_played%], ''), '1970-01-01 00:00:00') < datetime('now', '-300 minutes')
AND [%jsp3_loved%]

Name: Least Recently Played
Group: %album%
Determine probability for picking groups based on group size: Untick
Number of groups to add: 1
Number of tracks to add: 0
Number of minutes to add: 0
Number of MBs to add: 0
Limit after sorting: Untick
Sort: Empty
Sort / limit each group: Empty
Source / Filter
Type: SQL (requires foo_sqlite)
Query
WITH: CTE AS(
 SELECT path, subsong, [%jsp3_loved%], COALESCE(NULLIF([%jsp3_last_played%], ''), '1970-01-01 00:00:00') AS [%jsp3_last_played%]
 FROM MediaLibrary
)

SELECT path, subsong FROM: CTE

WHERE: [%jsp3_last_played%] < datetime('now', '-300 minutes')
AND [%jsp3_loved%]

ORDER BY: [%jsp3_last_played%]

LIMIT: 2

Filter with the content of the target playlist: Tick
Filter with the content of the target playlist using a title format expression: Empty
Priority: 1