TDC Ventures — Forecast Track Record
Not trading advice.
Aggregate stats
As of 3 November: twenty-eight forecasts resolved, two open. The two calls targeting 30 October resolved on 3 November: Pakistan shred 211 landed in range (Fastmarkets MB-STE-0887 $414.79, +$1.79), while Turkey shredded missed on both legs, each narrowly above the top of the range (Kallanish $419.00, +$9.00; Fastmarkets MB-STE-0095 $418.24, +$8.24, its first print on or after the target date). We called direction down on Turkey and both legs rose. The two calls targeting 6 November are open.
| Metric | Value |
|---|---|
| Fully resolved | 28 (38 scored sub-components, including our transaction-basis leg; 3 lapsed and excluded; 2 unscoreable and excluded) |
| Open, not yet fully resolved | 2 (Pakistan shred 211 CFR Qasim and Turkey shredded scrap CFR, each targeting 6 November) |
| Declared but unscored (no observation) | 2 (Pakistan shred 211, 11 September and 18 September, Kallanish legs — see below) |
| Hit rate — published-benchmark basis (excludes our own transaction-basis leg) | 26 of 37 = 70.3% |
| Hit rate including our transaction-basis leg (the number we quote as the headline) | 26 of 38 = 68.4% |
| Direction accuracy | 18 of 37 = 48.6% (18 of 38 = 47.4% including the transaction leg) |
| Mean absolute error vs central — price forecasts | $4.68/t excluding our own transaction-basis leg ($5.05/t including it). The spread forecast’s error is shown separately below — see “Why the spread call is not in the average.” |
| Median absolute error | $3.71/t excluding the transaction leg ($4.50/t including it) |
| MAE excluding outliers | $1.46/t (as of 12 October; not recomputed in this update) |
| September 2026 hit rate | 6 of 12 = 50.0% (price-basis; excludes the 1 September spread call, which is in range but reported separately below) |
| September 2026 mean absolute error | $6.43/t |
| August 2026 hit rate | 2 of 4 = 50% |
| August 2026 mean absolute error | $7.30/t |
| October 2026 hit rate (forecasts that resolved in the month) | 11 of 14 = 78.6% (price-basis; Turkey shredded 30 October both legs missed, Pakistan shred 211 30 October in range; Turkey shredded 2 October both legs, Pakistan shred 211 2 October, Turkey shredded 9 October Kallanish in range / Fastmarkets out, Pakistan shred 211 16 October, Turkey shredded 16 October both legs, Pakistan shred 211 23 October, Turkey shredded 23 October both legs) |
| October 2026 mean absolute error | $4.33/t |
| Best call | Turkey Shredded CFR (Kallanish), 2026-10-02 — $0.00/t (0.00%), on the central; the previous best was 2026-06-01 at $0.50/t |
| Worst call, published-benchmark basis | Turkey Shredded CFR (Kallanish), 2026-09-11 — $17.50/t (+4.42%) |
| Worst call, any basis | Pakistan Shred 211, transaction basis, 2026-08-06 — $18.00/t (+4.64%), now only $0.50/t ahead of the Turkey Sep-11 miss above. On the spread forecast our central was off by $119.50 (39.8%) — a different unit, described in full below |
Signed errors, all thirty-seven price-level calls scored on published benchmarks or our own transaction basis (the spread call is shown separately): +$0.00 · +$0.50 · +$0.64 · −$0.69 · +$0.75 · +$0.91 · +$1.07 · +$1.19 · +$1.21 · −$1.32 · +$1.35 · +$1.76 · −$1.79 · +$2.00 · −$2.28 · −$2.50 · +$2.91 · −$4.50 · +$4.56 · +$4.66 · +$5.00 · +$5.00 · −$5.09 · +$6.50 · +$6.64 · −$7.50 · +$7.66 · +$8.00 · +$8.84 · +$11.00 · −$12.09 · +$12.23 · +$17.50 · +$18.00 (30 October additions: +$1.79 Pakistan, +$8.24 and +$9.00 Turkey).
Twenty-six of the thirty-eight scored calls landed in range and fifteen of the thirty-seven price errors were $2.00/t or less. The 30 October Turkey call missed on both legs, each by a small margin above the range ceiling ($1.00 and $0.24): we called direction down, the 26 October Fastmarkets fall reversed the next day, and both publishers ended the week higher. Both 23 October calls landed in range with positive errors (+$5.00, +$0.91, +$1.35), as did both 16 October calls (+$8.00, +$2.91, +$1.07) after the 9 October downside break** — six consecutive sub-components have now printed above our central. The Fastmarkets Turkey leg has been flat at $409.91 for each of our last three resolutions, so a directional call on that leg has scored wrong every time since 9 October.
Why we quote one figure that excludes our own transaction-basis leg and one that includes it. Our 6 August Pakistan shred call declared two resolution bases: an index forecast scored against a published benchmark anyone can look up, and a transaction forecast scored against observed dealt levels. The index leg hit; the transaction leg produced what was then our worst miss (now our second-worst — see below). We quote the published-benchmark basis alongside the all-in figure because excluding an inconvenient leg from the headline would make it meaningless.
On our known bias. Through late July every error but one was positive — we were forecasting systematically too low, and we said so at the time. August broke that pattern in both directions at once: one call missed $18 high, another $7.50 low. Every resolution since (4 September, 11 September) has moved the pattern back toward “too low”: five consecutive sub-components have now missed high, none low. We are treating this as a live, worsening bias to correct for, not a settled one. The 9 October Turkey miss runs the other way — we called the plateau to break up and it broke down — so the bias is not a one-directional “too low”; the more consistent failure is calling the direction of a plateau break.
By market
| Market | Scored | In range | Notes |
|---|---|---|---|
| Turkey Shredded Scrap CFR | 22 | 13 | The 30 October call (direction down) missed on both legs, narrowly above the $403–418 range: Kallanish $419.00 (+$9.00), Fastmarkets $418.24 (+$8.24, first print on or after the target date); direction wrong on both. The 23 October call (direction down) landed in range on both legs (Kallanish $414.00, +$5.00; Fastmarkets $409.91, +$0.91); direction right on Kallanish, wrong on Fastmarkets (flat). The 16 October call (direction down) landed in range on both legs (Kallanish $415.00, +$8.00; Fastmarkets $409.91, +$2.91); direction right on Kallanish, wrong on Fastmarkets (flat). The 9 October call (direction up) was a partial: Kallanish in range ($417.50, −$4.50), Fastmarkets $5.09 below the floor ($409.91, −$12.09), direction wrong on both. Best call on the book (now $0.00/t, 2 October, Kallanish on its central) sits alongside its worst-in-market: the 2 October call landed in range on both legs (Kallanish $420.00, Fastmarkets $414.91); the 11 September call missed high on both legs, one of them (+$17.50) now our second-worst error anywhere; the 18 September call split, with Kallanish missing high (+$11.00) and Fastmarkets in range (+$4.66); the 25 September call landed in range on both legs (Kallanish +$5.00, Fastmarkets +$1.19). One earlier miss came from a publisher re-basing its whole European chain, not from a market move. |
| Pakistan Scrap Import / Shred 211 CFR Port Qasim | 11 | 9 | Two misses broke high (+$12.23, +$7.66) and one transaction-basis call missed high (+$18.00); the eight Fastmarkets index hits (+$1.21, +$4.56, −$1.32, −$2.28, −$0.69, +$1.07, +$1.35, +$1.79) all landed inside range, the first two above their central tendency, the next three below it, and the three most recent (16 October $414.07, 23 October $415.35, 30 October $414.79) above it. A companion publisher leg on both the 11 September and 18 September calls never printed again after a print roughly six weeks earlier and closed unscoreable both times — see below. |
| Turkey HMS 1&2 80:20 CFR | 2 | 2 | Both legs in range, errors +$2.00 and +$1.76. |
| Pakistan HMS 80:20 CFR Port Qasim | 1 | 1 | Proxy-derived spread call, +$0.75. |
| LME copper cash/3M spread | 1 | 1 | Our first spread call. In range, but the central was 39.8% high and the range was very wide — see below. |
| Dark markets (rails, PNS, windmill plates) | 3 | — | All three excluded from the headline. See “Calls we threw out” below. |
Recent resolutions
| Target date | Market | Basis | Forecast range | Actual | Hit | Error $/t |
|---|---|---|---|---|---|---|
| 2026-10-30 | Turkey shredded | Kallanish | $403–418 (c: $410) | $419.00 (30 Oct) | ✗ | +$9.00 |
| 2026-10-30 | Turkey shredded | Fastmarkets MB-STE-0095 | $403–418 (c: $410) | $418.24 (2 Nov, first print on or after target) | ✗ | +$8.24 |
| 2026-10-30 | Pakistan shred 211 | Fastmarkets MB-STE-0887 | $406–419 (c: $413) | $414.79 (30 Oct) | ✓ | +$1.79 |
| 2026-10-23 | Turkey shredded | Kallanish | $400–416 (c: $409) | $414.00 (23 Oct) | ✓ | +$5.00 |
| 2026-10-23 | Turkey shredded | Fastmarkets MB-STE-0095 | $400–416 (c: $409) | $409.91 (23 Oct) | ✓ | +$0.91 |
| 2026-10-23 | Pakistan shred 211 | Fastmarkets MB-STE-0887 | $408–420 (c: $414) | $415.35 (23 Oct) | ✓ | +$1.35 |
| 2026-10-16 | Turkey shredded | Kallanish | $398–416 (c: $407) | $415.00 (16 Oct) | ✓ | +$8.00 |
| 2026-10-16 | Turkey shredded | Fastmarkets MB-STE-0095 | $398–416 (c: $407) | $409.91 (16 Oct) | ✓ | +$2.91 |
| 2026-10-16 | Pakistan shred 211 | Fastmarkets MB-STE-0887 | $408–420 (c: $413) | $414.07 (16 Oct) | ✓ | +$1.07 |
| 2026-10-09 | Turkey shredded | Kallanish | $415–430 (c: $422) | $417.50 (9 Oct) | ✓ | −$4.50 |
| 2026-10-09 | Turkey shredded | Fastmarkets MB-STE-0095 | $415–430 (c: $422) | $409.91 (9 Oct) | ✗ | −$12.09 |
| 2026-10-02 | Pakistan shred 211 | Fastmarkets MB-STE-0887 | $413–425 (c: $418) | $417.31 (9 Oct) | ✓ | −$0.69 |
| 2026-10-02 | Turkey shredded | Kallanish | $412–428 (c: $420) | $420.00 (2 Oct) | ✓ | 0.00 |
| 2026-10-02 | Turkey shredded | Fastmarkets MB-STE-0095 | $412–428 (c: $420) | $414.91 (2 Oct) | ✓ | −$5.09 |
| 2026-09-18 | Pakistan shred 211 | Fastmarkets MB-STE-0887 | $415–428 (c: $420) | $418.68 (18 Sep) | ✓ | −$1.32 |
| 2026-09-18 | Pakistan shred 211 | Kallanish | $415–428 (c: $420) | no print since six weeks before target | — unscoreable | n/a |
| 2026-09-25 | Pakistan shred 211 | Fastmarkets MB-STE-0887 | $414–426 (c: $420) | $417.72 (25 Sep) | ✓ | −$2.28 |
| 2026-09-25 | Turkey shredded | Kallanish | $408–424 (c: $416) | $421.00 (25 Sep) | ✓ | +$5.00 |
| 2026-09-25 | Turkey shredded | Fastmarkets MB-STE-0095 | $408–424 (c: $416) | $417.19 (25 Sep) | ✓ | +$1.19 |
| 2026-09-11 | Pakistan shred 211 | Fastmarkets MB-STE-0887 | $408–423 (c: $415) | $419.56 (11 Sep) | ✓ | +$4.56 |
| 2026-09-11 | Pakistan shred 211 | Kallanish | $408–423 (c: $415) | no print since five weeks before target | — unscoreable | n/a |
| 2026-09-18 | Turkey shredded | Kallanish | $397–413 (c: $406) | $417.00 (18 Sep) | ✗ | +$11.00 |
| 2026-09-18 | Turkey shredded | Fastmarkets MB-STE-0095 | $397–413 (c: $406) | $410.66 (18 Sep) | ✓ | +$4.66 |
| 2026-09-11 | Turkey shredded | Kallanish | $391–401 (c: $396) | $413.50 (11 Sep) | ✗ | +$17.50 |
| 2026-09-11 | Turkey shredded | Fastmarkets MB-STE-0095 | $391–401 (c: $396) | $404.84 (11 Sep) | ✗ | +$8.84 |
| 2026-09-04 | Turkey shredded | Kallanish | $386–396 (c: $391) | $397.50 (4 Sep) | ✗ | +$6.50 |
| 2026-09-04 | Turkey shredded | Fastmarkets MB-STE-0095 | $386–396 (c: $391) | $397.64 (4 Sep) | ✗ | +$6.64 |
| 2026-09-04 | Pakistan scrap import | Fastmarkets MB-STE-0887 | $408–418 (c: $412.50) | $420.16 (4 Sep) | ✗ | +$7.66 |
| 2026-09-01 | LME copper cash/3M spread | LME official settlements | $120–650 backwardation (c: $300) | +$180.50 (1 Sep) | ✓ | −$119.50 (spread, not a price — see below) |
| 2026-08-06 | Pakistan shred 211 | Index (MB-STE-0887) | $406–420 (c: $413) | $414.21 (7 Aug) | ✓ | +$1.21 |
| 2026-08-06 | Pakistan shred 211 | Transaction basis | $383–396 (c: $388) | $406.00 (19 Aug) | ✗ | +$18.00 |
| 2026-08-06 | Turkey shredded | Kallanish | $390–400 (c: $395) | $392.50 (6 Aug) | ✓ | −$2.50 |
| 2026-08-06 | Turkey shredded | Fastmarkets MB-STE-0095 | $390–400 (c: $395) | $387.50 (6 Aug) | ✗ | −$7.50 |
| 2026-08-06 | Pakistan PNS | dark market — lapsed | $382–398 (c: $388) | $376 (9 Mar, last pre-target) | ✗ stale | −$12.00 |
| 2026-08-06 | Pakistan windmill plates | dark market — lapsed | $432–460 (c: $440) | $440 (30 Jun, last pre-target) | ✓ circular | $0.00 |
| 2026-07-28 | Pakistan R65 rails | dark market — lapsed | $432–455 (c: $443) | $440 (2 Jul, last pre-target) | ✓ circular | −$3.00 |
| 2026-07-28 | Pakistan HMS 80:20 | Proxy-derived | $378–396 (c: $388) | $388.75 (28 Jul) | ✓ | +$0.75 |
| 2026-07-17 | Turkey HMS 80:20 | Kallanish | $362–370 (c: $366) | $368.00 (16 Jul) | ✓ | +$2.00 |
| 2026-07-17 | Turkey HMS 80:20 | Fastmarkets MB-STE-0417 | $359–367 (c: $363) | $364.76 (16 Jul) | ✓ | +$1.76 |
| 2026-07-17 | Pakistan CFR Port Qasim | Fastmarkets MB-STE-0887 | $385–394 (c: $390) | $402.23 (17 Jul) | ✗ | +$12.23 |
| 2026-06-01 | Turkey shredded | Kallanish | $426–430 (c: $428) | $428.50 (4 Jun) | ✓ | +$0.50 |
| 2026-06-01 | Turkey shredded | Fastmarkets MB-STE-0095 | $418–425 (c: $421) | $421.64 (3 Jun) | ✓ | +$0.64 |
| 2026-06-01 | Pakistan CFR Port Qasim | Fastmarkets MB-STE-0887 | $420–426 (c: $422) | $420.21 (5 Jun) | ✓ | −$1.79 |
Why an unscoreable leg is not a miss
Our 6 September Pakistan shred 211 forecast declared two publishers as separate, equally-weighted legs: Fastmarkets (our primary benchmark) and Kallanish (secondary, flagged at the time as unreliable). Fastmarkets printed on schedule and landed inside our $408–423 range at $419.56, +$4.56 against our $415 central. Kallanish did not print again for this market after a print roughly five weeks before target, and it still had not printed ten business days after target either.
We are not scoring that as a miss, and we are not scoring it against the five-week-old number. Both would manufacture a result. Scoring a silent publisher against its last print assumes the market held still for five weeks, when our own primary benchmark moved meaningfully over the same span. We declared, in writing, before the outcome was knowable, that this specific leg would close unscoreable rather than be forced to a verdict either way if it went quiet. It is logged, it carries no error figure, and it is excluded from every ratio on this page — the same treatment we already apply when one of our own transaction-basis legs produces no observation at all.
What we changed as a result, already, not just noted for next time: new Pakistan shred 211 forecasts now drop a publisher leg from scoring in advance if it has gone quiet for roughly a month at the time we write the forecast, rather than declaring it and finding out five weeks later that it was never going to speak. Our 25 September Pakistan shred 211 forecast is the first written under that rule.
The same publisher leg closed unscoreable a second time on our 18 September call, which was written one week before that rule took effect. It declared Kallanish as a co-equal secondary leg, and Kallanish still has not printed on this market since the same print that caused the first gap. The primary Fastmarkets leg printed on schedule and landed inside our range at $418.68, −$1.32 against our $420 central. We are treating two-for-two unscoreable outcomes on the last two calls that declared this leg as confirmation the rule change above was the right call, not an overreaction to one gap.
The 9 October Turkey partial
We called the direction up off a month-long plateau, with a $415–430 range and a $422 central. Both publishers instead broke down: Kallanish printed $417.50 (inside the range, −$4.50 against our central) and Fastmarkets $409.91, $5.09 below the range floor and $12.09 under our central — our largest miss since 11 September, and the first to break low since 6 August.
The upside case we leaned on was a further chain-wide Fastmarkets re-basing of the kind that had recurred through September. It did not repeat; the step that arrived was down (−$4.28 on 2 October, −$5.00 on 8 October). The downside risks we listed — an iron-ore cost base weakening, regional demand staying capped, a plateau that breaks lower first — were the ones that played out, and we sized them too small relative to the upside. Kallanish eased far less (−$2.50) than Fastmarkets (−$9.28) over the week, so the two publishers ended $7.59 apart with the Kallanish shredded-over-HMS premium still locked at $20.00/t. Our next Turkey call (16 October) is written with direction down.
The 2 October Pakistan call
The published CFR Port Qasim shredded price (Fastmarkets MB-STE-0887) printed $417.31 on 9 October — the first print on or after our 2 October target — inside our $413–425 range and $0.69 under our $418 central. The call was a low-conviction, near-flat one: we declined to chase offer chatter far above the index, and the index stayed in the same $415–419 band it had held since 18 September. The direction call was marginal and we score it as right only against the anchor we declared when we wrote the forecast.
The 2 October Turkey call
Our $412–428 range held on both publishers, and the centre was exactly right on one of them. Kallanish printed $420.00, on our central. Fastmarkets printed $414.91, $5.09 below the central and $2.91 inside the range floor.
The direction call split. We expected a small step down on Kallanish and it came ($421.00 → $420.00 on the week). On Fastmarkets we expected a step up from $417.19 and it fell $2.28. We had named a further Fastmarkets re-basing as our upside case but no downward one, and the Fastmarkets leg’s one-session fall of 1.02% on 2 October is the move our central was on the wrong side of. Both legs stayed inside the range, which is what the range was built for.
The 18 September Turkey partial
Our $397–413 range held on one publisher and broke on the other: Fastmarkets printed $410.66 (inside the range, +$4.66 against our $406 central); Kallanish printed $417.00 ($4.00 above the range top, +$11.00).
The thesis was right — we had written the forecast into a breakout that both publishers were already confirming, and the worst-case reversal we flagged (the breakout giving its gain back within a week) did not happen. Neither leg gave anything back. Kallanish kept stepping up through the week while Fastmarkets held flat, so the two publishers diverged by $6.34 by the target date. The central again sat under the market, though the misses have narrowed from the week before (Kallanish +$17.50 → +$11.00, Fastmarkets +$8.84 → +$4.66). It is the third Turkey shredded resolution in a row with at least one leg breaking high.
The 11 September Turkey miss
Our $391–401 range missed on both legs, and both broke the same way the 4 September call did the week before: Kallanish printed $413.50 ($12.50 above our range top), Fastmarkets $404.84 ($3.84 above it).
We had already seen the breakout starting when we wrote the forecast — both publishers had jumped the week before (Kallanish +$6.50, Fastmarkets +$7.55) — and read it as a level to consolidate around rather than a trend to extend. A market participant we cited put the realistic further move at $2–3/t. The actual further move was $6.50 to $17.50/t depending on the leg — three to six times the ceiling we built our range around.
The pattern is now five misses deep and every one has broken the same direction. We named the upside risk correctly both times (4 September and 11 September) and sized it too small both times. We are treating “when we name an upside driver, size it toward the top of any range a market source gives us, not the bottom” as the standing correction to carry into the next call on this market.
The 4 September double miss
Both calls targeting 4 September missed, and both missed on the same side — every scored leg printed above the top of its range.
Turkey shredded scrap CFR ($386–396, c $391): Kallanish printed $397.50, Fastmarkets $397.64 — the two publishers converging to within $0.14 of each other, the tightest gap we’ve logged between them all year, both above our range. We had named the upside case — a second round of Fastmarkets re-basing its whole chain upward — but sized it at $1–3.74/t. It happened twice in four days and the market moved roughly three times what we sized.
Pakistan scrap import, MB-STE-0887 ($408–418, c $412.50): printed $420.16, the second-highest level in our dataset. We had explicitly argued that a dated September container-freight increase would land on the seller, not the price. The publisher’s own reporting for the target week says the opposite happened: offers rose “on higher freight costs and firmer global prices.” We were wrong about who absorbs a cost increase, not just about the price level.
The pattern to take from this: every material miss on our book to date has broken high. We named the mechanism twice — once for Turkey, once for Pakistan — and in both cases treated it as the smaller, less-likely alternative to a downside case that didn’t happen. Naming a risk is not the same as weighting it correctly, and we are carrying that forward as a standing correction rather than a one-off note.
Why the spread call is not in the average
On 18 August we forecast the gap between the LME copper cash settlement and its three-month settlement for 1 September. We said the gap would stay in steep backwardation, come off its 17 August peak of $535, and not collapse back to the $25–70 range it held through July. We put the range at $120–650 and the central at $300, with confidence marked Low.
It printed $180.50 — cash $14,395.50 against three-month $14,215.00. That is inside our range, so it scores as a hit. We do not think it should be read as one.
Three reasons, all of which cut against us:
- The central was wrong by $119.50, or 39.8%. Getting the shape right and the size that wrong is not a good forecast.
- The range was $530 wide on a quantity whose actual span over the fortnight was $56 to $290. A range that wide will contain most outcomes. The width produced the result, not the judgement.
- The date decided it. On 21 August the same spread printed $56.00 — $64 below our own floor. Had the target fallen four business days earlier, this would have been a clean miss. It then travelled $99 → $189 → $254 → $165 → $180.50 before settling on our target date.
We are not putting this error into our headline accuracy average, and we want to be explicit about why — because the reason could look self-serving. A $119.50 error on a $300 spread between two delivery dates for the same metal is not the same kind of number as an $18.00 error on a $388/tonne cargo price. Averaging them together would move our reported error from $4.98/t to roughly $12.61/t, and that jump would describe a change of unit, not a change in how well we forecast. So the spread call counts in our hit rate and our direction accuracy — those are just ratios — and its error is reported separately, in full, right here. Both averages are shown above.
This is now a permanent rule, not a one-off decision about an inconvenient number. Every forecast we write is tagged as either a price forecast (a traded level) or a spread forecast (a gap between two quoted things). Price and spread errors are averaged separately and always shown together. The rule cuts both ways, which is the point: a spread call with a flatteringly small error is kept out of the headline average too. We set it before we knew whether the next spread call would help us or hurt us.
What we are changing. A single settlement date is close to a coin flip on a quantity this volatile. Our next spread forecast will resolve against an average across the target week rather than one day’s print, so the result measures the call instead of the timing.
One thing that did work. When we wrote the forecast we specified, in advance, that it could only be scored against the exchange’s own published cash and three-month settlements — and that no market commentary could supply either number, even if the official prints were hard to retrieve. That mattered: a widely circulated commentary table dated “1 September” turned out to carry 28 August’s settlements, which would have produced a spread of $165.00 and an error of $135.00 against a date that never printed it. Naming the resolution source before the outcome exists is the cheapest discipline on this page, and this is the second time it has paid.
Our worst call, in full
On 30 July we forecast Pakistan shred 211 CFR Port Qasim on two bases: an index central of $413, and a transaction central of $388 with a range of $383–396.
The index leg hit at +$1.21. The transaction leg missed at +$18.00 — observed transaction levels reached around $406/t CFR Port Qasim on 19 August, ten dollars above the top of our range.
We had named the mechanism ourselves. Our primary upside case was a mill-restocking wave and we sized it at +$5–10/t. It delivered roughly double. Getting the driver right and the magnitude wrong is still a miss.
The standing rule we adopted: a transaction-basis central is only as fresh as the transaction series beneath it. Ours rested on observed levels from earlier in the summer while the published index had already moved. Where a benchmark and a transaction series disagree, the benchmark is the one with a refresh schedule. Since 18 August, every forecast on a market without a published index is written as a spread to a live anchor rather than off a standalone transaction level.
The 6 August Turkey miss
Our $390–400 range was correct on one publisher and wrong on the other, on the same grade and the same date: Kallanish printed $392.50 (in range, −$2.50), Fastmarkets $387.50 (out, −$7.50).
The cause was not a market move. Fastmarkets re-based every leg of its European deep-sea chain by the same amount on the same date. A publisher re-basing is not price transmission, and a forecast that declares a single publisher as its resolution basis will be scored on that publisher’s methodology decisions as well as on the market. We now declare both legs as separate sub-components before the print exists, so neither can be dropped after the outcome is known.
Calls we threw out
Three dark-market forecasts — rails, PNS and windmill plates — resolved by lapsing when no qualifying transaction appeared inside the window. All three are excluded from our headline statistics, and we think the reason matters more than the exclusion.
Two of them scored against the very observation their central tendency had been derived from. One produced a $0.00 “perfect” error that measures nothing except the size of its own anchor; the other scored against a level 150 days older than its target. A scoring rule that manufactures a perfect hit is broken, so we flagged both, excluded them, and changed the rule rather than banking a flattering number.
How we score
- Scoring window: the first published print on or after the target date. No pre-target interim scoring — an early print that happens to flatter a call is not a resolution.
- Hit: actual falls inside the forecast range. Partial: within $1/t of a boundary. Miss: more than $1/t outside.
- Error: signed, against central tendency. Positive means the market printed above our central.
- Resolution basis is declared in writing before the target print exists. Where two publishers assess the same grade, both are named as separate sub-components in advance, so neither can be selected after the fact.
- Lapsed calls on markets with no published index are excluded from the headline and shown separately, with the reason.