diff --git a/Cargo.lock b/Cargo.lock
index 82422e9..2cc5da4 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -2066,7 +2066,7 @@ dependencies = [
 
 [[package]]
 name = "opsqueue"
-version = "0.30.1"
+version = "0.30.4"
 dependencies = [
  "anyhow",
  "arc-swap",
@@ -2120,7 +2120,7 @@ dependencies = [
 
 [[package]]
 name = "opsqueue_python"
-version = "0.30.2"
+version = "0.30.4"
 dependencies = [
  "anyhow",
  "chrono",
diff --git a/libs/opsqueue_python/Cargo.toml b/libs/opsqueue_python/Cargo.toml
index c856d8a..51df4a2 100644
--- a/libs/opsqueue_python/Cargo.toml
+++ b/libs/opsqueue_python/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "opsqueue_python"
-version = "0.30.2"
+version = "0.30.4"
 edition = "2021"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
diff --git a/libs/opsqueue_python/opsqueue_python.nix b/libs/opsqueue_python/opsqueue_python.nix
index dd06c91..e355fda 100644
--- a/libs/opsqueue_python/opsqueue_python.nix
+++ b/libs/opsqueue_python/opsqueue_python.nix
@@ -30,6 +30,7 @@ buildPythonPackage rec {
       ".toml"
       ".lock"
       ".db"
+      ".md"
     ];
   };
 
diff --git a/opsqueue/Cargo.toml b/opsqueue/Cargo.toml
index f210031..724d373 100644
--- a/opsqueue/Cargo.toml
+++ b/opsqueue/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "opsqueue"
-version = "0.30.1"
+version = "0.30.4"
 edition = "2021"
 description = "lightweight batch processing queue for heavy loads"
 repository = "https://github.com/channable/opsqueue"
diff --git a/opsqueue/migrations/20250416081453_track_chunk_size.down.sql b/opsqueue/migrations/20250416081453_track_chunk_size.down.sql
new file mode 100644
index 0000000..c095e57
--- /dev/null
+++ b/opsqueue/migrations/20250416081453_track_chunk_size.down.sql
@@ -0,0 +1,3 @@
+ALTER TABLE submissions DROP COLUMN chunk_size;
+ALTER TABLE submissions_completed DROP COLUMN chunk_size;
+ALTER TABLE submissions_failed DROP COLUMN chunk_size;
diff --git a/opsqueue/migrations/20250416081453_track_chunk_size.sql b/opsqueue/migrations/20250416081453_track_chunk_size.up.sql
similarity index 100%
rename from opsqueue/migrations/20250416081453_track_chunk_size.sql
rename to opsqueue/migrations/20250416081453_track_chunk_size.up.sql
diff --git a/opsqueue/migrations/20250803174028_better_random_order_formula.down.sql b/opsqueue/migrations/20250803174028_better_random_order_formula.down.sql
new file mode 100644
index 0000000..d1913ab
--- /dev/null
+++ b/opsqueue/migrations/20250803174028_better_random_order_formula.down.sql
@@ -0,0 +1,38 @@
+-- 1. Drop all indexes and finally the column itself
+DROP INDEX random_chunks_order;
+DROP INDEX random_chunks_metadata_order;
+DROP INDEX random_chunks_metadata_order2;
+ALTER TABLE chunks DROP COLUMN random_order;
+ALTER TABLE chunks_metadata DROP COLUMN random_order;
+
+-- 2. Recreate the column with its **old, bad** definition
+ALTER TABLE chunks ADD COLUMN random_order INTEGER NOT NULL GENERATED ALWAYS AS (
+    (((submission_id + chunk_index) % 65536) * 40503) % 65536
+    ) VIRTUAL;
+ALTER TABLE chunks_metadata ADD COLUMN random_order INTEGER NOT NULL GENERATED ALWAYS AS (
+    (((submission_id + chunk_index) % 65536) * 40503) % 65536
+) VIRTUAL;
+
+
+-- 3. Recreate all dropped indexes
+CREATE INDEX random_chunks_order ON chunks (
+      random_order
+    , submission_id
+    , chunk_index
+);
+
+CREATE INDEX random_chunks_metadata_order ON chunks_metadata (
+      metadata_key
+    , metadata_value
+    , random_order
+    , submission_id
+    , chunk_index
+);
+
+CREATE INDEX random_chunks_metadata_order2 ON chunks_metadata (
+      metadata_key
+    , random_order
+    , metadata_value
+    , submission_id
+    , chunk_index
+);
diff --git a/opsqueue/migrations/20250803174028_better_random_order_formula.up.sql b/opsqueue/migrations/20250803174028_better_random_order_formula.up.sql
new file mode 100644
index 0000000..98cf1de
--- /dev/null
+++ b/opsqueue/migrations/20250803174028_better_random_order_formula.up.sql
@@ -0,0 +1,43 @@
+-- 1. Drop all indexes and finally the column itself
+DROP INDEX random_chunks_order;
+DROP INDEX random_chunks_metadata_order;
+DROP INDEX random_chunks_metadata_order2;
+ALTER TABLE chunks DROP COLUMN random_order;
+ALTER TABLE chunks_metadata DROP COLUMN random_order;
+
+-- 2. Recreate the column with its new, proper definition
+--
+-- Compared to the OG definition, we ensure that the top 42 bits of `submission_id`
+-- which contain the timestamp part of the snowflake,
+-- always participate in the `random_order`,
+-- since the lower 22 bits are likely to be `0` except when under peak load.
+ALTER TABLE chunks ADD COLUMN random_order INTEGER NOT NULL GENERATED ALWAYS AS (
+    (((submission_id + (submission_id >> 22) + chunk_index) % 65536) * 40503) % 65536
+    ) VIRTUAL;
+ALTER TABLE chunks_metadata ADD COLUMN random_order INTEGER NOT NULL GENERATED ALWAYS AS (
+    (((submission_id + (submission_id >> 22) + chunk_index) % 65536) * 40503) % 65536
+) VIRTUAL;
+
+
+-- 3. Recreate all dropped indexes
+CREATE INDEX random_chunks_order ON chunks (
+      random_order
+    , submission_id
+    , chunk_index
+);
+
+CREATE INDEX random_chunks_metadata_order ON chunks_metadata (
+      metadata_key
+    , metadata_value
+    , random_order
+    , submission_id
+    , chunk_index
+);
+
+CREATE INDEX random_chunks_metadata_order2 ON chunks_metadata (
+      metadata_key
+    , random_order
+    , metadata_value
+    , submission_id
+    , chunk_index
+);
diff --git a/opsqueue/opsqueue_example_database_schema.db b/opsqueue/opsqueue_example_database_schema.db
index bae0b8c..bced270 100644
Binary files a/opsqueue/opsqueue_example_database_schema.db and b/opsqueue/opsqueue_example_database_schema.db differ
diff --git a/opsqueue/src/consumer/strategy.rs b/opsqueue/src/consumer/strategy.rs
index c975b5e..ff45e5a 100644
--- a/opsqueue/src/consumer/strategy.rs
+++ b/opsqueue/src/consumer/strategy.rs
@@ -99,7 +99,12 @@ impl Strategy {
         match self {
             Oldest => qb.push("\nORDER BY submission_id ASC"),
             Newest => qb.push("\nORDER BY submission_id DESC"),
-            Random => qb.push("\nORDER BY random_order ASC"),
+            Random => {
+                // It is **very** important that we do not apply extra sorting here.
+                // For the implementation of the 'cutting the deck' technique
+                // we rely on the order being 'whatever comes out of the UNION ALL'
+                qb
+            }
             PreferDistinct { .. } => {
                 // **no** change in sort order. PreferDistinct passes the sort order on to the inner strategies that it unions.
                 qb
@@ -114,6 +119,9 @@ pub type ChunkStream<'a> = BoxStream<'a, Result<Chunk, sqlx::Error>>;
 #[cfg(test)]
 #[cfg(feature = "server-logic")]
 pub mod test {
+    use crate::common::chunk::ChunkSize;
+    use crate::common::StrategicMetadataMap;
+
     use super::*;
     use itertools::Itertools;
     use sqlx::Row;
@@ -180,11 +188,11 @@ pub mod test {
 
         assert_streaming_query(qb, &explained);
         insta::assert_snapshot!(explained, @r"
-        1, 0, MERGE (UNION ALL)
-        3, 1, LEFT
-        7, 3, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
-        26, 1, RIGHT
-        30, 26, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
+        1, 0, COMPOUND QUERY
+        2, 1, LEFT-MOST SUBQUERY
+        5, 2, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
+        22, 1, UNION ALL
+        25, 22, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
         ");
     }
 
@@ -211,12 +219,12 @@ pub mod test {
         12, 8, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
         18, 8, LIST SUBQUERY 2
         20, 18, SCAN json_each VIRTUAL TABLE INDEX 0:
-        60, 1, UNION ALL
-        63, 60, SCAN chunks
-        66, 60, CORRELATED SCALAR SUBQUERY 6
-        70, 66, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        76, 66, LIST SUBQUERY 2
-        78, 76, SCAN json_each VIRTUAL TABLE INDEX 0:
+        63, 1, UNION ALL
+        66, 63, SCAN chunks
+        69, 63, CORRELATED SCALAR SUBQUERY 6
+        73, 69, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        79, 69, LIST SUBQUERY 2
+        81, 79, SCAN json_each VIRTUAL TABLE INDEX 0:
         ");
     }
 
@@ -244,12 +252,12 @@ pub mod test {
         12, 8, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
         18, 8, LIST SUBQUERY 2
         20, 18, SCAN json_each VIRTUAL TABLE INDEX 0:
-        60, 1, UNION ALL
-        63, 60, SCAN chunks
-        66, 60, CORRELATED SCALAR SUBQUERY 6
-        70, 66, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        76, 66, LIST SUBQUERY 2
-        78, 76, SCAN json_each VIRTUAL TABLE INDEX 0:
+        63, 1, UNION ALL
+        66, 63, SCAN chunks
+        69, 63, CORRELATED SCALAR SUBQUERY 6
+        73, 69, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        79, 69, LIST SUBQUERY 2
+        81, 79, SCAN json_each VIRTUAL TABLE INDEX 0:
         ");
     }
 
@@ -271,29 +279,33 @@ pub mod test {
         insta::assert_snapshot!(explained, @r"
         1, 0, COMPOUND QUERY
         2, 1, LEFT-MOST SUBQUERY
-        4, 2, CO-ROUTINE inner_company_id
-        5, 4, MERGE (UNION ALL)
-        7, 5, LEFT
-        11, 7, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
-        30, 5, RIGHT
-        34, 30, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
-        77, 2, SCAN inner_company_id
-        81, 2, CORRELATED SCALAR SUBQUERY 5
-        85, 81, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        91, 81, LIST SUBQUERY 3
-        93, 91, SCAN json_each VIRTUAL TABLE INDEX 0:
-        127, 1, UNION ALL
-        129, 127, CO-ROUTINE inner_company_id
-        130, 129, MERGE (UNION ALL)
-        132, 130, LEFT
-        136, 132, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
-        155, 130, RIGHT
-        159, 155, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
-        202, 127, SCAN inner_company_id
-        206, 127, CORRELATED SCALAR SUBQUERY 7
-        210, 206, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        216, 206, LIST SUBQUERY 3
-        218, 216, SCAN json_each VIRTUAL TABLE INDEX 0:
+        3, 2, COMPOUND QUERY
+        4, 3, LEFT-MOST SUBQUERY
+        7, 4, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
+        16, 4, CORRELATED SCALAR SUBQUERY 5
+        20, 16, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        26, 16, LIST SUBQUERY 3
+        28, 26, SCAN json_each VIRTUAL TABLE INDEX 0:
+        63, 3, UNION ALL
+        66, 63, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
+        76, 63, CORRELATED SCALAR SUBQUERY 5
+        80, 76, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        86, 76, LIST SUBQUERY 3
+        88, 86, SCAN json_each VIRTUAL TABLE INDEX 0:
+        123, 3, UNION ALL
+        124, 123, COMPOUND QUERY
+        125, 124, LEFT-MOST SUBQUERY
+        128, 125, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
+        137, 125, CORRELATED SCALAR SUBQUERY 7
+        141, 137, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        147, 137, LIST SUBQUERY 3
+        149, 147, SCAN json_each VIRTUAL TABLE INDEX 0:
+        184, 124, UNION ALL
+        187, 184, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
+        197, 184, CORRELATED SCALAR SUBQUERY 7
+        201, 197, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        207, 197, LIST SUBQUERY 3
+        209, 207, SCAN json_each VIRTUAL TABLE INDEX 0:
         ");
     }
 
@@ -323,71 +335,142 @@ pub mod test {
         2, 1, LEFT-MOST SUBQUERY
         3, 2, COMPOUND QUERY
         4, 3, LEFT-MOST SUBQUERY
-        6, 4, CO-ROUTINE inner_priority
-        7, 6, MERGE (UNION ALL)
-        9, 7, LEFT
-        13, 9, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
-        32, 7, RIGHT
-        36, 32, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
-        79, 4, SCAN inner_priority
-        83, 4, CORRELATED SCALAR SUBQUERY 5
-        87, 83, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        93, 83, LIST SUBQUERY 3
-        95, 93, SCAN json_each VIRTUAL TABLE INDEX 0:
-        123, 4, CORRELATED SCALAR SUBQUERY 11
-        127, 123, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        133, 123, LIST SUBQUERY 9
-        135, 133, SCAN json_each VIRTUAL TABLE INDEX 0:
-        169, 3, UNION ALL
-        171, 169, CO-ROUTINE inner_priority
-        172, 171, MERGE (UNION ALL)
-        174, 172, LEFT
-        178, 174, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
-        197, 172, RIGHT
-        201, 197, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
-        244, 169, SCAN inner_priority
-        248, 169, CORRELATED SCALAR SUBQUERY 7
-        252, 248, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        258, 248, LIST SUBQUERY 3
-        260, 258, SCAN json_each VIRTUAL TABLE INDEX 0:
-        288, 169, CORRELATED SCALAR SUBQUERY 11
-        292, 288, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        298, 288, LIST SUBQUERY 9
-        300, 298, SCAN json_each VIRTUAL TABLE INDEX 0:
-        334, 3, UNION ALL
-        335, 334, COMPOUND QUERY
-        336, 335, LEFT-MOST SUBQUERY
-        338, 336, CO-ROUTINE inner_priority
-        339, 338, MERGE (UNION ALL)
-        341, 339, LEFT
-        345, 341, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
-        364, 339, RIGHT
-        368, 364, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
-        411, 336, SCAN inner_priority
-        415, 336, CORRELATED SCALAR SUBQUERY 5
-        419, 415, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        425, 415, LIST SUBQUERY 3
-        427, 425, SCAN json_each VIRTUAL TABLE INDEX 0:
-        455, 336, CORRELATED SCALAR SUBQUERY 13
-        459, 455, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        465, 455, LIST SUBQUERY 9
-        467, 465, SCAN json_each VIRTUAL TABLE INDEX 0:
-        501, 335, UNION ALL
-        503, 501, CO-ROUTINE inner_priority
-        504, 503, MERGE (UNION ALL)
-        506, 504, LEFT
-        510, 506, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
-        529, 504, RIGHT
-        533, 529, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
-        576, 501, SCAN inner_priority
-        580, 501, CORRELATED SCALAR SUBQUERY 7
-        584, 580, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        590, 580, LIST SUBQUERY 3
-        592, 590, SCAN json_each VIRTUAL TABLE INDEX 0:
-        620, 501, CORRELATED SCALAR SUBQUERY 13
-        624, 620, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
-        630, 620, LIST SUBQUERY 9
-        632, 630, SCAN json_each VIRTUAL TABLE INDEX 0:
+        5, 4, COMPOUND QUERY
+        6, 5, LEFT-MOST SUBQUERY
+        9, 6, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
+        18, 6, CORRELATED SCALAR SUBQUERY 5
+        22, 18, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        28, 18, LIST SUBQUERY 3
+        30, 28, SCAN json_each VIRTUAL TABLE INDEX 0:
+        57, 6, CORRELATED SCALAR SUBQUERY 11
+        61, 57, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        67, 57, LIST SUBQUERY 9
+        69, 67, SCAN json_each VIRTUAL TABLE INDEX 0:
+        104, 5, UNION ALL
+        107, 104, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
+        117, 104, CORRELATED SCALAR SUBQUERY 5
+        121, 117, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        127, 117, LIST SUBQUERY 3
+        129, 127, SCAN json_each VIRTUAL TABLE INDEX 0:
+        156, 104, CORRELATED SCALAR SUBQUERY 11
+        160, 156, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        166, 156, LIST SUBQUERY 9
+        168, 166, SCAN json_each VIRTUAL TABLE INDEX 0:
+        203, 5, UNION ALL
+        204, 203, COMPOUND QUERY
+        205, 204, LEFT-MOST SUBQUERY
+        208, 205, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
+        217, 205, CORRELATED SCALAR SUBQUERY 7
+        221, 217, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        227, 217, LIST SUBQUERY 3
+        229, 227, SCAN json_each VIRTUAL TABLE INDEX 0:
+        256, 205, CORRELATED SCALAR SUBQUERY 11
+        260, 256, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        266, 256, LIST SUBQUERY 9
+        268, 266, SCAN json_each VIRTUAL TABLE INDEX 0:
+        303, 204, UNION ALL
+        306, 303, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
+        316, 303, CORRELATED SCALAR SUBQUERY 7
+        320, 316, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        326, 316, LIST SUBQUERY 3
+        328, 326, SCAN json_each VIRTUAL TABLE INDEX 0:
+        355, 303, CORRELATED SCALAR SUBQUERY 11
+        359, 355, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        365, 355, LIST SUBQUERY 9
+        367, 365, SCAN json_each VIRTUAL TABLE INDEX 0:
+        402, 204, UNION ALL
+        403, 402, COMPOUND QUERY
+        404, 403, LEFT-MOST SUBQUERY
+        405, 404, COMPOUND QUERY
+        406, 405, LEFT-MOST SUBQUERY
+        409, 406, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
+        418, 406, CORRELATED SCALAR SUBQUERY 5
+        422, 418, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        428, 418, LIST SUBQUERY 3
+        430, 428, SCAN json_each VIRTUAL TABLE INDEX 0:
+        457, 406, CORRELATED SCALAR SUBQUERY 13
+        461, 457, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        467, 457, LIST SUBQUERY 9
+        469, 467, SCAN json_each VIRTUAL TABLE INDEX 0:
+        504, 405, UNION ALL
+        507, 504, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
+        517, 504, CORRELATED SCALAR SUBQUERY 5
+        521, 517, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        527, 517, LIST SUBQUERY 3
+        529, 527, SCAN json_each VIRTUAL TABLE INDEX 0:
+        556, 504, CORRELATED SCALAR SUBQUERY 13
+        560, 556, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        566, 556, LIST SUBQUERY 9
+        568, 566, SCAN json_each VIRTUAL TABLE INDEX 0:
+        603, 405, UNION ALL
+        604, 603, COMPOUND QUERY
+        605, 604, LEFT-MOST SUBQUERY
+        608, 605, SEARCH chunks USING INDEX random_chunks_order (random_order>?)
+        617, 605, CORRELATED SCALAR SUBQUERY 7
+        621, 617, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        627, 617, LIST SUBQUERY 3
+        629, 627, SCAN json_each VIRTUAL TABLE INDEX 0:
+        656, 605, CORRELATED SCALAR SUBQUERY 13
+        660, 656, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        666, 656, LIST SUBQUERY 9
+        668, 666, SCAN json_each VIRTUAL TABLE INDEX 0:
+        703, 604, UNION ALL
+        706, 703, SEARCH chunks USING INDEX random_chunks_order (random_order<?)
+        716, 703, CORRELATED SCALAR SUBQUERY 7
+        720, 716, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        726, 716, LIST SUBQUERY 3
+        728, 726, SCAN json_each VIRTUAL TABLE INDEX 0:
+        755, 703, CORRELATED SCALAR SUBQUERY 13
+        759, 755, SEARCH submissions_metadata USING COVERING INDEX lookup_submission_by_metadata (metadata_key=? AND metadata_value=? AND submission_id=?)
+        765, 755, LIST SUBQUERY 9
+        767, 765, SCAN json_each VIRTUAL TABLE INDEX 0:
         ");
     }
+
+    use crate::db::Connection;
+    use futures::stream::TryStreamExt as _;
+
+    #[sqlx::test]
+    /// Tests whether the 'cutting the deck' technique is working
+    ///
+    /// We do this by checking whether two selects in a huge amount of available chunks
+    /// give a different result.
+    /// (There is a super tiny chance of this test flaking).
+    pub async fn test_random_strategy_is_random(pool: sqlx::SqlitePool) {
+        let db_pools = crate::db::DBPools::from_test_pool(&pool);
+
+        let mut conn = db_pools.writer_conn().await.unwrap();
+        let input_chunks: Vec<_> = (0..10_000).map(|x| Some(format!("{x}").into())).collect();
+        crate::common::submission::db::insert_submission_from_chunks(
+            None,
+            input_chunks.clone(),
+            None,
+            StrategicMetadataMap::default(),
+            ChunkSize::default(),
+            &mut conn,
+        )
+        .await
+        .unwrap();
+
+        let mut conn = db_pools.reader_conn().await.unwrap();
+        let mut query_builder = QueryBuilder::default();
+        let vals1: Vec<Chunk> = Strategy::Random
+            .build_query(&mut query_builder, &Default::default())
+            .build_query_as()
+            .fetch(conn.get_inner())
+            .try_collect()
+            .await
+            .unwrap();
+
+        let mut query_builder = QueryBuilder::default();
+        let vals2: Vec<Chunk> = Strategy::Random
+            .build_query(&mut query_builder, &Default::default())
+            .build_query_as()
+            .fetch(conn.get_inner())
+            .try_collect()
+            .await
+            .unwrap();
+
+        assert!(vals1 != vals2)
+    }
 }