ChainBench · Flotapponnier · Sep 8, 2026 · Sep 8, 2026
@@ -61,13 +61,13 @@ abstract: |
   same event.
 
 methodology:
-  - "Aggregators measured: Mobula, Codex, GeckoTerminal, Serialized."
+  - "Aggregators measured: Mobula, Codex, GeckoTerminal."
   - "Chains: Base, BNB Chain, Solana, Robinhood Chain."
   - "Regions: us-east, eu-west, sgp. Cross-region median reported in the headline."
   - "Reference: archive nodes per chain, validated against block hashes."
   - "Metric: gauge `head_lag_seconds`, sampled every 15 seconds. Aggregated over the 24-hour window using `quantile_over_time`."
   - "Success rate: presence ratio. Share of expected sampling slots where a value was actually emitted (5,760 expected per provider per day at the 15s cadence). 100% means the aggregator's feed was reachable for the full window."
-  - "Cardinality: 4 aggregators × 4 chains × 3 regions = 48 active series."
+  - "Cardinality: 3 aggregators × 4 chains × 3 regions = 36 active series."
 
 findings:
   - "Cross-chain average puts {{best_name}} on top at {{best_p50}} (p50, 24h) across {{count}} providers. Per-chain leaders differ: Base {{best_name:chain:base}} at {{best_p50:chain:base}}, Solana {{best_name:chain:solana}} at {{best_p50:chain:solana}}, BNB {{best_name:chain:bnb}} at {{best_p50:chain:bnb}}, Robinhood {{best_name:chain:robinhood}} at {{best_p50:chain:robinhood}}. Chains differ in block cadence (Solana 400ms vs Base/BNB 2-3s), which shifts the floor independently of provider speed."
@@ -217,30 +217,6 @@ providers:
           p50: avg by (aggregator) (quantile_over_time(0.50, head_lag_seconds{aggregator="codex", region="sgp"}[24h])) * 1000
           series: avg_over_time(head_lag_seconds{aggregator="codex", region="sgp"}[1h]) * 1000 unless (changes(head_lag_seconds{aggregator="codex", region="sgp"}[15m]) == 0 and count_over_time(head_lag_seconds{aggregator="codex", region="sgp"}[15m]) > 5 and on(chain, region) sum by (chain, region) (changes(head_lag_seconds{region="sgp"}[15m])) > 5 and on(aggregator, region) avg_over_time(ws_connected{aggregator="codex", region="sgp", chain=""}[15m]) > 0.8)
 
-  - slug: serialized
-    name: Serialized
-    tag: WebSocket trades stream
-    formula: "Median seconds between an on-chain swap on the bench pools and the same trade arriving on Serialized's WebSocket trades stream (subscribed per pool by its token side), sampled every 15s over 24h."
-    queries:
-      p50: avg by (aggregator) (quantile_over_time(0.50, head_lag_seconds{aggregator="serialized"}[24h])) * 1000
-      p90: avg by (aggregator) (quantile_over_time(0.90, head_lag_seconds{aggregator="serialized"}[24h])) * 1000
-      p99: avg by (aggregator) (quantile_over_time(0.99, head_lag_seconds{aggregator="serialized"}[24h])) * 1000
-      mean: avg by (aggregator) (avg_over_time(head_lag_seconds{aggregator="serialized"}[24h])) * 1000
-      success: clamp_max(avg by (aggregator) (count_over_time(head_lag_seconds{aggregator="serialized"}[24h]) / 5760), 1)
-      sample_size: sum(count_over_time(head_lag_seconds{aggregator="serialized"}[24h]))
-      series: avg_over_time(head_lag_seconds{aggregator="serialized"}[1h]) * 1000 unless (changes(head_lag_seconds{aggregator="serialized"}[15m]) == 0 and count_over_time(head_lag_seconds{aggregator="serialized"}[15m]) > 5 and on(chain, region) sum by (chain, region) (changes(head_lag_seconds{}[15m])) > 5 and on(aggregator, region) avg_over_time(ws_connected{aggregator="serialized", chain=""}[15m]) > 0.8)
-      live_activity: sum(changes(head_lag_seconds{aggregator="serialized"}[15m]))
-      regions:
-        - region: us-east
-          p50: avg by (aggregator) (quantile_over_time(0.50, head_lag_seconds{aggregator="serialized", region="us-east"}[24h])) * 1000
-          series: avg_over_time(head_lag_seconds{aggregator="serialized", region="us-east"}[1h]) * 1000 unless (changes(head_lag_seconds{aggregator="serialized", region="us-east"}[15m]) == 0 and count_over_time(head_lag_seconds{aggregator="serialized", region="us-east"}[15m]) > 5 and on(chain, region) sum by (chain, region) (changes(head_lag_seconds{region="us-east"}[15m])) > 5 and on(aggregator, region) avg_over_time(ws_connected{aggregator="serialized", region="us-east", chain=""}[15m]) > 0.8)
-        - region: eu-west
-          p50: avg by (aggregator) (quantile_over_time(0.50, head_lag_seconds{aggregator="serialized", region="eu-west"}[24h])) * 1000
-          series: avg_over_time(head_lag_seconds{aggregator="serialized", region="eu-west"}[1h]) * 1000 unless (changes(head_lag_seconds{aggregator="serialized", region="eu-west"}[15m]) == 0 and count_over_time(head_lag_seconds{aggregator="serialized", region="eu-west"}[15m]) > 5 and on(chain, region) sum by (chain, region) (changes(head_lag_seconds{region="eu-west"}[15m])) > 5 and on(aggregator, region) avg_over_time(ws_connected{aggregator="serialized", region="eu-west", chain=""}[15m]) > 0.8)
-        - region: ap-southeast
-          p50: avg by (aggregator) (quantile_over_time(0.50, head_lag_seconds{aggregator="serialized", region="sgp"}[24h])) * 1000
-          series: avg_over_time(head_lag_seconds{aggregator="serialized", region="sgp"}[1h]) * 1000 unless (changes(head_lag_seconds{aggregator="serialized", region="sgp"}[15m]) == 0 and count_over_time(head_lag_seconds{aggregator="serialized", region="sgp"}[15m]) > 5 and on(chain, region) sum by (chain, region) (changes(head_lag_seconds{region="sgp"}[15m])) > 5 and on(aggregator, region) avg_over_time(ws_connected{aggregator="serialized", region="sgp", chain=""}[15m]) > 0.8)
-
   - slug: geckoterminal
     name: GeckoTerminal
     tag: REST feed

@@ -3,7 +3,7 @@
 > **Pre-onboarding evaluation.** Run before Serialized is wired into any live harness, so the
 > decision to include or exclude them on each bench is documented and reproducible.
 >
-> **Version:** v1.3, 2026-09-06 (§8 corrected in v1.1; §16.3 root cause corrected and §17 added in v1.2; §18 added in v1.3: Serialized wired into bench 001). Author: internal. Key used: tenant `OpenChainBench`,
+> **Version:** v1.4, 2026-09-08 (§8 corrected in v1.1; §16.3 root cause corrected and §17 added in v1.2; §18 added in v1.3: Serialized wired into bench 001; §18.1 added in v1.4: that wiring reverted after measurement). Author: internal. Key used: tenant `OpenChainBench`,
 > plan `starter`, keyId `d5511a080aaa`, issued 2026-09-04.
 
 ---
@@ -503,7 +503,7 @@ Layer 3, the real bench: snapshot every provider's verdict at mint, resolve on-c
 Snapshot `top10HoldersPct` and `bundlersHoldingsPct` as the primary signals per §17.5.
 
 
-## 18. Bench 001: Serialized is in (v1.3, 2026-09-06)
+## 18. Bench 001: Serialized wired, then pulled (v1.4, 2026-09-08)
 
 Earlier sections called 001 "blocked on a policy decision". That was the wrong framing, and it hid a
 practical question nobody had tested: does Serialized's stream cover the four bench pools at all?
@@ -533,3 +533,50 @@ Solana attempt looked like a failure.
 `head_lag_ref_seconds` against the node reference clock matched by `txHash`, which is the one that can
 rank providers. Every event carries `txHash`, `block`, `poolAddress` and a `preconfirmed` boolean per
 their docs, so Base flashblocks are visible rather than inferred.
+
+### 18.1 Why it was reverted
+
+Measured 2026-09-08 before promoting the wiring to `main`. One WS connection, the four bench pools,
+a 240 s capture, plus a targeted 150 s run that asked a public Base node whether the block existed
+at the moment each trade arrived. Single vantage (workstation, NTP offset +0.078 s), so absolute
+latencies are not comparable to harness numbers; the signs and the orders of magnitude are.
+
+**The ruler matches.** Their `at` is the block timestamp exactly, so `head_lag_seconds` would measure
+the same quantity as Mobula's `trade.Date` and Codex's `event.Timestamp`:
+
+| chain | n | median(`at` − block timestamp) |
+|---|---|---|
+| base | 105 | +0.000 s |
+| bnb | 1 | +0.000 s |
+| solana | 28 | +0.473 s (artefact: `getBlockTime` returns whole seconds) |
+
+**But they read a different thing.** On Base, every sampled trade arrived before its own block:
+
+| check | result |
+|---|---|
+| median(receipt − block timestamp), base | −0.911 s |
+| share of base events with negative lag | 105/105 (100%) |
+| trades received before the block existed on a public node | 11/11 |
+| median lead over block publication | 1.04 s |
+
+In all eleven cases the node head at receipt was exactly `target block − 1`: the block was not late in
+propagation, it did not exist yet. The spread (−0.21 s to −1.80 s) is the width of one Base slot.
+Serialized streams sequencer preconfirmations; Mobula, Codex and GeckoTerminal read sealed blocks.
+
+**Decision: excluded from bench 001.** Not because the measurement is wrong, but because the two
+emissions are different products. A preconfirmation carries no finality guarantee and can be
+reordered, so the ~1.5 s lead is a latency/finality trade-off, not pure speed. Publishing both in one
+ranking would present that trade-off as superiority. `RecordHeadLag` also drops negatives outright
+(`metrics.go`), so on the legacy series they would show as no data on Base while being the fastest —
+and any rare positive sample would stick on the gauge and become their published p50.
+
+Two findings worth raising with them:
+
+- `preconfirmed` was `false` on all eleven events that preceded block publication. A consumer cannot
+  tell the two regimes apart from the payload.
+- Robinhood delivered **0 events in 240 s** despite an acknowledged subscription, against 79 in 90 s
+  recorded on 2026-09-06 (§18 table). Re-verify before any future wiring.
+
+**Reopen if** the bench gains an emission-regime dimension (sealed vs preconfirmed), or a landing-rate
+companion series shows their preconfirmations reach sealed blocks 1:1. Serialized stays on benches
+004, 005, 008 and 090, where the comparison is like for like.
@@ -10,7 +10,6 @@ import (
 type Config struct {
 	CoinGeckoAPIKey      string
 	MobulaAPIKey         string
-	SerializedAPIKey     string
 	DefinedSessionCookie string
 	MonitorRegion        string // Deployment region: us-west, us-east, singapore, etc.
 	MobulaWSURL          string // Mobula fast-trade WebSocket endpoint (allows staging to use EU-specific cluster)
@@ -22,7 +21,6 @@ func loadEnv() (*Config, error) {
 	// First, try to load from environment variables (for production/Railway)
 	config.CoinGeckoAPIKey = strings.TrimSpace(os.Getenv("COINGECKO_API_KEY"))
 	config.MobulaAPIKey = strings.TrimSpace(os.Getenv("MOBULA_API_KEY"))
-	config.SerializedAPIKey = strings.TrimSpace(os.Getenv("SERIALIZED_API_KEY"))
 	config.DefinedSessionCookie = strings.TrimSpace(os.Getenv("DEFINED_SESSION_COOKIE"))
 	config.MonitorRegion = strings.TrimSpace(os.Getenv("MONITOR_REGION"))
 	config.MobulaWSURL = strings.TrimSpace(os.Getenv("MOBULA_WS_URL"))

@@ -811,9 +811,6 @@ func runHeadLagMonitor(config *Config, stopChan <-chan struct{}) {
 	wg.Add(1)
 	go runGeckoTerminalHeadLagMonitor(config, stopChan, &wg)
 
-	wg.Add(1)
-	go runSerializedHeadLagMonitor(config, stopChan, &wg)
-
 	// Wait for all to finish
 	wg.Wait()
 	fmt.Println("[HEAD-LAG] All monitors stopped")