Transparency: the honest track record

Factual, verifiable performance evidence for EV-IDEX predictions — backtested, paper-traded and live results kept strictly apart, every number with its sample size and uncertainty, losses included. This page is updated automatically and shows no customer data.

How to read this page

Backtested = simulated on past data (can flatter). Paper-traded = Model v2's sealed shadow ledger, never shown to customers. Live = the public v1 ledger, sealed before outcomes were known. Every accuracy figure carries its correct/eligible counts and 95% interval; below 30 graded calls it is labelled "below the reporting threshold", and nothing is called meaningful below 100.

Cost model: Profitability after costs = signed return − 0.3% round trip (0.10% fee + 0.05% slippage per side), 1 unit per call. Neutral-band grading unchanged. Updated Fri, 13 Nov 2026 18:28:41 GMT.

insufficientinsufficient — fewer than 30 graded calls; no accuracy figure is reported
reportablereportable with wide uncertainty — 30–99 graded calls; NOT conclusive
meaningfulstatistically meaningful — ≥ 100 graded calls AND the 95% interval excludes the baseline
not meaningful≥ 100 graded calls but the 95% interval still includes the baseline — no demonstrated edge
Backtested (simulated)

2024-04-30 → 2026-10-26 · 14 coins · chronological walk-forward · reserved segment from 2026-04-28

7-day horizon
Model v2 (every 4 h)n graded 201 · calls 542
Accuracy55.7% (CI 48.8–62.4)Avg net / call+0.693% (CI -0.53–1.917)Calls beating costs47.6%Max drawdown−267.78%Profit factor1.18
Model v2 @ daily pointsn graded 92 · calls 260
Accuracy53.3% (CI 43.1–63.1)Avg net / call+0.489% (CI -1.115–2.094)Calls beating costs47.3%Max drawdown−173.8%Profit factor1.12
v1 rules @ daily pointsn graded 86 · calls 253
Accuracy57% (CI 46.4–66.9)Avg net / call+0.659% (CI -0.586–1.905)Calls beating costs45.8%Max drawdown−296.41%Profit factor1.21
Buy & hold (same windows)n graded 512 · calls 1385
Accuracy50.6% (CI 46.3–54.9)Avg net / call+0.644% (CI -0.031–1.319)Calls beating costs45.6%Max drawdown−745.68%Profit factor1.18
Buy & hold @ dailyn graded 485 · calls 1385
Accuracy52.2% (CI 47.7–56.6)Avg net / call+0.677% (CI 0.013–1.341)Calls beating costs46%Max drawdown−759.68%Profit factor1.19
Trend-followingn graded 512 · calls 1385
Accuracy46.7% (CI 42.4–51)Avg net / call-0.304% (CI -0.982–0.373)Calls beating costs43.9%Max drawdown−791.95%Profit factor0.93
24-hour horizon
Model v2 (every 4 h)n graded 406 · calls 1080
Accuracy54.9% (CI 50.1–59.7)Avg net / call-0.097% (CI -0.43–0.237)Calls beating costs41.9%Max drawdown−252.35%Profit factor0.95
Model v2 @ daily pointsn graded 122 · calls 351
Accuracy55.7% (CI 46.9–64.2)Avg net / call+0.077% (CI -0.518–0.671)Calls beating costs41%Max drawdown−82.91%Profit factor1.04
v1 rules @ daily pointsn graded 163 · calls 584
Accuracy54.6% (CI 46.9–62.1)Avg net / call-0.131% (CI -0.426–0.164)Calls beating costs42.6%Max drawdown−186.55%Profit factor0.9
Buy & hold (same windows)n graded 3400 · calls 9757
Accuracy50.7% (CI 49–52.4)Avg net / call-0.309% (CI -0.396–-0.221)Calls beating costs42.2%Max drawdown−3117.18%Profit factor0.81
Buy & hold @ dailyn graded 3145 · calls 9757
Accuracy50.2% (CI 48.5–52)Avg net / call-0.310% (CI -0.396–-0.224)Calls beating costs41.5%Max drawdown−3150.9%Profit factor0.8
Trend-followingn graded 3400 · calls 9757
Accuracy50.9% (CI 49.2–52.5)Avg net / call-0.405% (CI -0.493–-0.318)Calls beating costs41.7%Max drawdown−3996.02%Profit factor0.75
Decision: REMAIN IN SHADOW — DO NOT REPLACE v1
v2 vs v1 (7d, identical dates): no statistically meaningful difference
v2 vs buy & hold (7d): no statistically meaningful difference
Long-only study: DIRECTIONALLY BETTER, NOT STATISTICALLY ESTABLISHED: dropping SELL calls improves v2's average net after costs and beats buy & hold, but the net CI includes zero and none of the comparisons against BUY/SELL, buy & hold or v1 rules reaches p < 0.05. The improvement holds on the reserved segment.
Confidence calibration: NO MEANINGFUL DIFFERENCE: the calibrated map is not statistically better than raw confidence out of sample (paired CI includes zero) — raw confidence remains over-stated; calibrated numbers are closer to the truth on average but the gain is not established.
Paper-traded (v2 shadow ledger)

shadow — never shown to customers. 113 calls still pending.

1-hour callsn graded 360 · calls 714
Accuracy55% (CI 49.8–60.1)Avg net / call-0.146% (CI -0.227–-0.065)Calls beating costs36.3%
4-hour callsn graded 158 · calls 314
Accuracy51.9% (CI 44.2–59.6)Avg net / call-0.259% (CI -0.424–-0.094)Calls beating costs36.9%
24-hour callsn graded 99 · calls 148
Accuracy57.6% (CI 47.7–66.8)Avg net / call-0.231% (CI -0.758–0.297)Calls beating costs50%
7-day callsn graded 38 · calls 53
Accuracy57.9% (CI 42.2–72.1)Avg net / call-0.679% (CI -2.932–1.574)Calls beating costs52.8%
Live head-to-head vs v1: insufficient data
Live (public v1 ledger)

319 unique predictions (every record sealed) · 149 pending · 170 finished · 138 graded · 52 directional graded · 88 independent graded · 241 historical duplicates (audit only, never scored) · 171 superseded

3-day directional callsn graded 34 · calls 61
Accuracy61.8% (CI 45–76.1)Avg net / call+0.477% (CI -0.542–1.495)Calls beating costs49.2%
14-day directional callsn graded 6 · calls 7
Accuracynot reported (< 30)Avg net / call+3.863% (CI -2.636–10.362)Calls beating costs57.1%
30-day directional callsn graded 6 · calls 6
Accuracynot reported (< 30)Avg net / call-1.011% (CI -16.225–14.203)Calls beating costs50%
7-day directional callsn graded 6 · calls 8
Accuracynot reported (< 30)Avg net / call+0.718% (CI -3.367–4.803)Calls beating costs62.5%
Headline directional accuracy — market (AI outlook) predictions, independent sample (one call per asset, timeframe and horizon window), Bullish/Bearish calls only, WAIT/Neutral calls excluded: 54.8% · 17 correct of 31 eligible · 95% CI 37.8–70.8% · 19 neutral-band outcomes excluded from the denominator
All-calls accuracy (includes WAIT/Neutral calls graded by the ledger rules; shown separately, never labelled directional): 78% · 64/82 correct · 95% CI 67.9–85.6% · 20 neutral-band excluded
Full published record — every directional call, market and strategy (overlapping and superseded included), WAIT/Neutral calls excluded: 61.5% · 32/52 correct · 95% CI 48–73.5% · 30 neutral-band outcomes excluded from the denominator
Full published record — WAIT/Neutral calls only (reported separately, never blended into directional accuracy): 86% · 74/86 correct · 95% CI 77.2–91.8% · 2 neutral-band excluded
Public Prediction Accuracy Score: 40/100 (preliminary, 82 resolved)
Every sealed record, the hash chain check and the CSV export are available to any free account under Track Record — no paid plan required. Open the accuracy score & ledger →

Proof-of-concept status

Two different questions, answered separately: does the machine work (data in → sealed prediction → automatic grading) and are the predictions any good. Everything below is computed live from the public ledger; nothing is estimated.

Technical proof of concept
demonstrated
560 sealed records · 170 reached their horizon · 138 graded automatically by the daily job.
Prediction accuracy
not validated
31 of the 100 independent directional calls needed before accuracy can be judged. The interval below is too wide to tell skill from chance.
Advantage after costs
not demonstrated
Needs ≥100 calls, a positive average return after realistic fees and a t-statistic above 2. Not a promise of future results.
Independent directional accuracy
Correct / graded
17 / 31 = 54.8%
95% CI 37.8–70.8% · 19 neutral-band outcomes excluded
Live (sealed at prediction time)Recorded the moment a user requested the signal.
9 / 15 = 60%
CI 35.7–80.2% · 21 pending
Historical backfillSignals that existed before the ledger; sealed up to ~4 days after they were made, so not tamper-evident from the start.
0 / 6 = 0%
CI 0–39% · 0 pending
Scheduled (daily recorder)Recorded automatically each day at 05:00 UTC; first run 4 Nov 2026.
8 / 10 = 80%
CI 49–94.3% · 128 pending
Simple baselines on the same 31 windows
EV-IDEX v1 (the model)
54.8%
CI 37.8–70.8%
Coin flip
50%
Always say Bearish
90.3%
28/31 · CI 75.1–96.7%
Always say Bullish
9.7%
3/31 · CI 3.3–24.9%
Buy & hold, same windows
-6.26%
average move per window · 9.7% of windows rose

A model only shows skill if it beats the best of these over many calls. In a one-way market "always Bearish" or "always Bullish" can look brilliant by luck.

Average return per call after transaction costs

One unit per call: buy on Bullish, sell short on Bearish, close at the horizon. Costs are charged once per round trip. Spot-only users cannot short, so the long-only row is the realistic one for most people.

No fees
+0.84%
54.8% of calls positive · t = 0.56 · hold -6.26%
0.1% × 2 (low-fee exchange)
+0.64%
54.8% of calls positive · t = 0.43 · hold -6.26%
0.6% × 2 (Coinbase taker)
-0.36%
54.8% of calls positive · t = -0.24 · hold -6.26%
Long-only (Bullish calls only), 0.6% × 2
-8.83%
11 calls · 0% positive
Insufficient evidence: with 31 independent directional calls, neither predictive accuracy nor profitability is established. These figures describe the past sample only.

Independent sample: one call per asset, timeframe and horizon window; Bullish/Bearish market calls only; neutral-band outcomes excluded from the denominator; baselines graded with the same ±2% rule on the same windows.

Live evidence ledger — what is reportable today

Each row is a real, sealed sample. A tier describes how much the sample can say — it never certifies accuracy.

SampleTypeGradedAccuracy (95% CI)Avg net after costsBuy & hold, same windowsWorst callTier
v1 live engine · 7-day directional callslive ledger6 / 30not reported (< 30)——-10.1%insufficient
v1 live engine · 1-day directional callslive ledger0 / 30not reported (< 30)———insufficient
v2 shadow · 7-day BUY/SELL callspaper / shadow38 / 10057.9% (42.2–72.1)-0.679% (-2.932–1.574)-3.253%-19.46%reportable
v2 shadow · 7-day BUY-only subset (long-only study)paper / shadow12 / 30not reported (< 30)——-19.46%insufficient
v2 shadow · 7-day calls with stated confidence (calibration study)paper / shadow38 / 10057.9% (42.2–72.1)-0.679% (-2.932–1.574)-3.253%-19.46%reportable
v2 shadow · 24-hour BUY/SELL callspaper / shadow99 / 10057.6% (47.7–66.8)-0.231% (-0.758–0.297)-1.241%-9.56%reportable
Weekly evidence summary
Nov 6 → Nov 13 · last digest e-mailed 2026-W46

Identical to the Monday e-mail digest: counts are read from the sealed ledger for the last 7 days — nothing is estimated or back-filled. The Prediction Accuracy Score is published only once both thresholds below are met.

New sealed records
90
origin: live 4 · scheduled 86 · source: ai 85 · rule 5
Graded this week
67
correct 44 · incorrect 9 · neutral 14 · directional 34 · invalidated 0
Independent graded
88 / 30
threshold met
Independent directional
32 / 20
threshold met
Still pending
149
graded automatically at their horizon (05:00 UTC)
Last 12 weeks sealed graded
W35W37W39W41W43W45
Daily evidence recorder — run log last run OK

Every day at 05:00 UTC the recorder seals one AI signal and one rule-engine decision per covered coin (origin scheduled), using the same model, prompt and grading rules as user-triggered analyses. Retried runs only fill gaps — never duplicates. Credits are the market-data and AI calls each run spent.

next run · Nov 14, 05:00 AM UTC
Run (your time)StatusUniverseRecords createdDuplicatesSkipsCredits usedErrors
Nov 13, 05:02 AM UTC
cron
done13 coins13 · 11 AI / 2 rule4757 market-data · 13 AI0
Nov 12, 05:01 AM UTC
cron
done13 coins18 · 18 AI / 0 rule4251 market-data · 13 AI0
Nov 11, 05:01 AM UTC
cron
done13 coins8 · 8 AI / 0 rule52510 market-data · 13 AI0
Nov 10, 05:01 AM UTC
cron
done13 coins13 · 13 AI / 0 rule4758 market-data · 13 AI0
Nov 9, 05:01 AM UTC
cron
done13 coins9 · 8 AI / 1 rule4765 market-data · 12 AI0
Nov 8, 05:01 AM UTC
cron
done13 coins13 · 12 AI / 1 rule4751 market-data · 13 AI0
Nov 7, 05:01 AM UTC
cron
done13 coins12 · 11 AI / 1 rule48510 market-data · 13 AI0
Nov 6, 05:01 AM UTC
cron
done13 coins23 · 22 AI / 1 rule3848 market-data · 13 AI0
Nov 5, 05:03 AM UTC
cron
done13 coins22 · 21 AI / 1 rule38510 market-data · 13 AI0
Nov 4, 05:02 AM UTC
cron
done13 coins46 · 39 AI / 7 rule14510 market-data · 13 AI0
Skips by coin · across the last 10 runs
CoinSkippedAI / ruleMost common reason
FIGR_HELOC10×0 / 10no Coinbase USD market or not enough daily candles for the rule engine
TRX10×0 / 10no Coinbase USD market or not enough daily candles for the rule engine
WBT10×0 / 10no Coinbase USD market or not enough daily candles for the rule engine
XMR10×0 / 10no Coinbase USD market or not enough daily candles for the rule engine
LEO9×0 / 9no Coinbase USD market or not enough daily candles for the rule engine
figure-heloc1×1 / 0market data delayed or empty — not recorded

Retry no-ops ("already recorded today") are not counted here — a retried run only fills gaps and never duplicates. Coins without a Coinbase USD market get an AI record only; the rule engine needs exchange candles.

Covered coins · 13
coverage from scanner run · Nov 13, 06:08 AM

Coverage = coins in the latest v1 scanner run. Rule-engine decisions additionally require a Coinbase USD market with ≥ 260 daily candles. Coins without a Coinbase USD market therefore receive AI records only.

#CoinExchange marketRecords per runScheduled records sealed
4BNB · BNBCoinbase USDAI + rule19
1BTC · BitcoinCoinbase USDAI + rule9
12DOGE · DogecoinCoinbase USDAI + rule14
2ETH · EthereumCoinbase USDAI + rule11
9FIGR_HELOC · Figure Helocno Coinbase marketAI only14
10HYPE · HyperliquidCoinbase USDAI + rule13
15LEO · LEO Tokenno Coinbase marketAI only15
13XMR · Monerono Coinbase marketAI only13
6XRP · XRPCoinbase USDAI + rule13
7SOL · SolanaCoinbase USDAI + rule11
8TRX · TRONno Coinbase marketAI only14
16WBT · WhiteBIT Coinno Coinbase marketAI only13
11ZEC · ZcashCoinbase USDAI + rule13

Methodology

  • Live ledger (v1): every public prediction is written to a hash-chained ledger when it is made, before the outcome is known, and graded once at its horizon by fixed rules (neutral band for small moves). Any free account can verify the chain under Track Record.
  • Shadow ledger (v2): Model v2 (deterministic quant engine) runs alongside v1 and records sealed predictions nobody sees. It is compared against v1 on identical dates; it is not used for any customer recommendation.
  • Backtest: strictly chronological walk-forward over the full coin universe — indicators use only data available at each decision point, every model is scored on the same windows with the same costs, and a reserved final 20% of the period was never used to choose anything.
  • Baselines: coin flip (50%) for accuracy, cash (0%) and buy & hold on the same windows for net return. A model only matters if it beats these after costs with an interval that excludes them.
  • Statistics: Wilson 95% intervals for proportions, bootstrap intervals (2000 resamples) for mean net, two-proportion z-tests and paired bootstraps for model comparisons, Brier/ECE for confidence calibration.

Risk disclosures

  • EV-IDEX is a research and paper-trading platform. Nothing on this page is investment advice, an offer, or a solicitation. No real trades are executed by EV-IDEX.
  • Backtested results are simulations on past data with assumed fills and fixed 0.30% round-trip costs (0.10% fee + 0.05% slippage per side). They were produced after the fact and can look better than live trading would have.
  • Paper-traded (shadow) and live ledger results are real, sealed, time-stamped predictions graded once by fixed rules — but they are still not real-money results: no capital was at risk, no market impact, no missed fills.
  • Small samples carry wide uncertainty. We show the 95% Wilson interval next to every accuracy figure. Below 30 graded calls a percentage is shown only together with its correct/eligible counts and is labelled 'below the reporting threshold'; even 100 calls do not prove an edge unless the interval excludes the baseline.
  • Directional accuracy ignores calls that ended inside the neutral band; sign-after-cost and average net per call count every call and are the honest profitability view. Losses, worst calls and drawdowns are shown deliberately.
  • Cryptocurrency prices are highly volatile. You can lose all the money you invest. Past performance — live, paper or simulated — never guarantees future results.

Want the full statistical detail (walk-forward tables, regime tests, confidence calibration, long-only study)? Create a free account and open Prediction quality tab. Nothing on this page is investment advice.