Pace IGDB requests; rewrite the new help text

IGDB allows 4 requests a second, and an import searching three entries
at once (plus detail lookups) went over it. Every IGDB request now waits
at least 300 ms after the last one, and a 429 is retried after 1, 2 and
4 seconds before giving up.

The import dialog, ranking editor and their error messages are reworded
in plainer language, and talk about "spots" instead of "places".

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

internal/providers/igdb.go

@@ -99,11 +99,15 @@ func (g *igdb) query(ctx context.Context, body string) ([]Result, error) {
 			Name         string `json:"name"`
 		} `json:"platforms"`
 	}
-	for attempt := 0; ; attempt++ {
-		tok, err := g.token(ctx, attempt > 0)
+	refreshed, backoff := false, time.Second
+	for {
+		tok, err := g.token(ctx, refreshed)
 		if err != nil {
 			return nil, err
 		}
+		if err := g.r.igdbPace.wait(ctx); err != nil {
+			return nil, err
+		}
 		req, err := http.NewRequestWithContext(ctx, http.MethodPost, "https://api.igdb.com/v4/games", strings.NewReader(body))
 		if err != nil {
 			return nil, err
@@ -112,7 +116,19 @@ func (g *igdb) query(ctx context.Context, body string) ([]Result, error) {
 		req.Header.Set("Authorization", "Bearer "+tok)
 		req.Header.Set("Content-Type", "text/plain")
 		err = g.r.do(req, &games)
-		if isStatus(err, 401) && attempt == 0 {
+		if isStatus(err, 401) && !refreshed {
+			refreshed = true
+			continue
+		}
+		// Rate limited anyway (another app on the same key, say): back off
+		// and try again, up to about 7 seconds in all.
+		if isStatus(err, 429) && backoff <= 4*time.Second {
+			select {
+			case <-time.After(backoff):
+			case <-ctx.Done():
+				return nil, ctx.Err()
+			}
+			backoff *= 2
 			continue
 		}
 		if err != nil {