diff --git a/lib/datalog/db.sx b/lib/datalog/db.sx
index 9df4234b..93544c7e 100644
--- a/lib/datalog/db.sx
+++ b/lib/datalog/db.sx
@@ -48,17 +48,37 @@
       (and (has-key? edb rel-key)
            (has-key? (get edb rel-key) tk)))))
 
-;; Evaluation strategy. Default :semi-naive (the only strategy
-;; currently implemented). :magic is reserved for goal-directed
-;; magic-sets evaluation — calling it now logs a one-time "deferred"
-;; note and falls back to semi-naive.
+;; Evaluation strategy. Default :semi-naive (used by dl-saturate!).
+;; :naive selects dl-saturate-naive! (slower but easier to reason
+;; about). :magic is a marker — goal-directed magic-sets evaluation
+;; is invoked separately via `dl-magic-query`; setting :magic here
+;; is purely informational. Any other value is rejected so typos
+;; don't silently fall back to the default.
+(define
+  dl-strategy-values
+  (list :semi-naive :naive :magic))
+
 (define
   dl-set-strategy!
   (fn
     (db strategy)
-    (do
-      (dict-set! db :strategy strategy)
-      db)))
+    (cond
+      ((not (dl-keyword-member? strategy dl-strategy-values))
+        (error (str "dl-set-strategy!: unknown strategy " strategy
+                    " — must be one of " dl-strategy-values)))
+      (else
+        (do
+          (dict-set! db :strategy strategy)
+          db)))))
+
+(define
+  dl-keyword-member?
+  (fn
+    (k xs)
+    (cond
+      ((= (len xs) 0) false)
+      ((= k (first xs)) true)
+      (else (dl-keyword-member? k (rest xs))))))
 
 (define
   dl-get-strategy
diff --git a/lib/datalog/scoreboard.json b/lib/datalog/scoreboard.json
index e7621023..4dd9e4b1 100644
--- a/lib/datalog/scoreboard.json
+++ b/lib/datalog/scoreboard.json
@@ -1,13 +1,13 @@
 {
   "lang": "datalog",
-  "total_passed": 275,
+  "total_passed": 276,
   "total_failed": 0,
-  "total": 275,
+  "total": 276,
   "suites": [
     {"name":"tokenize","passed":31,"failed":0,"total":31},
     {"name":"parse","passed":23,"failed":0,"total":23},
     {"name":"unify","passed":29,"failed":0,"total":29},
-    {"name":"eval","passed":43,"failed":0,"total":43},
+    {"name":"eval","passed":44,"failed":0,"total":44},
     {"name":"builtins","passed":26,"failed":0,"total":26},
     {"name":"semi_naive","passed":8,"failed":0,"total":8},
     {"name":"negation","passed":12,"failed":0,"total":12},
@@ -16,5 +16,5 @@
     {"name":"magic","passed":37,"failed":0,"total":37},
     {"name":"demo","passed":21,"failed":0,"total":21}
   ],
-  "generated": "2026-05-11T09:34:17+00:00"
+  "generated": "2026-05-11T09:40:12+00:00"
 }
diff --git a/lib/datalog/scoreboard.md b/lib/datalog/scoreboard.md
index 4676d363..c06a2f24 100644
--- a/lib/datalog/scoreboard.md
+++ b/lib/datalog/scoreboard.md
@@ -1,13 +1,13 @@
 # datalog scoreboard
 
-**275 / 275 passing** (0 failure(s)).
+**276 / 276 passing** (0 failure(s)).
 
 | Suite | Passed | Total | Status |
 |-------|--------|-------|--------|
 | tokenize | 31 | 31 | ok |
 | parse | 23 | 23 | ok |
 | unify | 29 | 29 | ok |
-| eval | 43 | 43 | ok |
+| eval | 44 | 44 | ok |
 | builtins | 26 | 26 | ok |
 | semi_naive | 8 | 8 | ok |
 | negation | 12 | 12 | ok |
diff --git a/lib/datalog/tests/eval.sx b/lib/datalog/tests/eval.sx
index 937a49dc..2625f04e 100644
--- a/lib/datalog/tests/eval.sx
+++ b/lib/datalog/tests/eval.sx
@@ -359,6 +359,16 @@
           (do (dl-set-strategy! db :magic) (dl-get-strategy db)))
         :magic)
 
+      ;; Unknown strategy values are rejected so typos don't silently
+      ;; fall back to the default.
+      (dl-et-test!
+        "unknown strategy rejected"
+        (dl-et-throws?
+          (fn ()
+            (let ((db (dl-make-db)))
+              (dl-set-strategy! db :semi_naive))))
+        true)
+
       ;; dl-saturated?: no-work-left predicate.
       (dl-et-test! "saturated? after saturation"
         (let ((db (dl-program
diff --git a/plans/datalog-on-sx.md b/plans/datalog-on-sx.md
index 791c77d8..d1b84163 100644
--- a/plans/datalog-on-sx.md
+++ b/plans/datalog-on-sx.md
@@ -15,7 +15,7 @@ for rose-ash data (e.g. federation graph, content relationships).
 
 ## Status (rolling)
 
-`bash lib/datalog/conformance.sh` → **275/275 across 11 suites**
+`bash lib/datalog/conformance.sh` → **276/276 across 11 suites**
 (tokenize, parse, unify, eval, builtins, semi_naive, negation, aggregates,
 api, magic, demo). Source is ~3100 LOC, tests ~2900 LOC, public API
 documented in `lib/datalog/datalog.sx`.
@@ -320,6 +320,13 @@ large graphs.
 
 _Newest first._
 
+- 2026-05-11 — `dl-set-strategy!` accepted arbitrary keyword values
+  silently. Typos like `:semi_naive` or `:semiNaive` were stored
+  uninspected; the saturator then used the default and the user
+  never learned their setting was a typo. Validator added: strategy
+  must be one of `:semi-naive`, `:naive`, `:magic`. 1 regression test;
+  276/276.
+
 - 2026-05-11 — Anonymous-variable renamer collided with user-written
   `_anon<N>` symbols. The renamer started counter at 0 and produced
   `_anon1, _anon2, ...` unconditionally; if the user wrote