Does the vector filter help?
Compare the original bullish-vector breakout and midpoint retest with the same price pattern allowing any volume.
CRYPTO RESEARCH, IN PLAIN SIGHT Research & paper only
Visitor feedback · owner inbox
Private feedback submitted through the public form. Email addresses are supplied by visitors and are not verified. Review messages here; no notification emails are sent automatically.
Load the inbox to read submitted feedback.
VOLUME. CONTEXT. EVIDENCE.
Explore the pattern.
Follow the evidence.
When trading activity rises sharply, does price behave differently afterwards? We test that question and whether fixed trading rules make money after costs. The research runs automatically; you can explore the results and help us explain them clearly.
Free to explore. No account or wallet needed.
Checking report freshness…
Collection targetEvery 4 hours
Hourly catch-up checks if a report is late.
Next target · local timeChecking…
Scheduling can be delayed.
Fixed forward studyLoading…
Review on 19 November 2026 UTC.
01
Start with the basicsOpen How it works. No trading experience needed.
02
Read the findingsCheck the date and what the results can tell us.
03
Tell us what confused youA short comment is a useful contribution.
HELP US MAKE THIS USEFUL
Help improve Vector Desk
You do not need trading experience. Try the page for a few minutes and tell us:
- Which word, number or instruction was hard to understand?
- What did you try to do, and what happened?
- What would make this worth visiting again?
Your feedback helps us improve the explanation and find errors. Reading the site or saving a practice plan does not add your results to the shared research.
Write feedback · no account needed
Prefer email? Vectordesk@frogclub.io
Prompts to copy into an email
Copying prepares a draft; it does not send an email.
THE ONGOING EXPERIMENT
Does the volume filter improve results?
Reading the latest evidence…
The same trading rules are tested with and without a volume filter, on new data and after costs. Reports target a four-hour cadence, with hourly catch-up checks. You do not need to run these comparisons yourself.
This report is older than expected. Treat these results as a dated snapshot.
A download was requested. If your browser blocks it, select and copy the exact displayed report below.
WHY THE RULES ARE WAITING
Latest hourly setup snapshot
From the saved report: 1h candles with 4h context. These are pattern checks, not live entry instructions or your personal trading record. Entry costs, current quotes and position limits are checked separately.
Loading setup explanations…
WHAT THE EARLIER SAMPLE SHOWED
Vectors alongside ordinary candles
Loading the saved historical comparison…
Inspect the saved study ↗ · Research scope & next decisions ↗
CURRENT SCOPE
Two questions. A visible record.
Do defined vector events differ from comparable ordinary candles? Can fixed trading rules profit after modeled costs? We measure these separately, with automatic collection and paper-only tools.
We cannot identify which addresses produced the activity from these candles. Participant Study B is deferred until a scope review on 19 November 2026, pending a coverage check, research design and resource budget. Accounts, new strategies and live execution are outside this phase.
ABOVE & BELOW PRICE
Historical vector-candle ranges
Unvisited portions of historical vector-candle ranges, ranked by estimated whole-candle traded value. A large amount belongs to the original candle, even when only a small price range remains. It is not value remaining in that range, outstanding orders or market-maker inventory.
Are vectors recovered in both directions?View the study +
Green vectors are checked for recovery downward; red vectors for recovery upward. Each is compared with ordinary candles in similar conditions.
Checking completed follow-up…
Loading the two-way study…
Starts 21 September 2026 UTC. Each event needs 24 hours of follow-up. Midpoint recovery is also measured before another 1 ATR move away; ambiguous same-bar order counts as non-success. Young events are pending, never failed.
Recovery is not a trading return. Trade simulations remain spot long-only. A short strategy needs separate perpetual-market data, funding and execution tests. Candles cannot reveal a market maker’s inventory.
Data, automation & limitations
Shared research covers UBTC, UETH and HYPE spot markets using Hyperliquid public data. UBTC and UETH are venue representations of BTC and ETH. This website is a research interface; it does not connect a wallet or place orders.
The fixed comparisons target four-hour updates, with hourly catch-up checks. This page checks for updates every five minutes and when you return to it. The market desk requests current data when you open it or refresh. Simulated results use assumed fees and slippage and can include a provisional exit at the sample end. Historical recovery does not establish future profitability.
Your paper journal is saved only in this browser. Export it before clearing browser data or changing devices. Visitors do not have automatic cloud recording, accounts or cross-device sync. The owner's cloud recorder is separate and private.
Read the method audit ↗ · Quick start guide ↗
Download the shared report ↗ · Download the fixed protocol ↗
Public Hyperliquid spot feed · no account neededRefresh to load
Data quality is checked on every load.
UBTC / USDC
—USDC
Awaiting data
Context uses completed candles
Load a market to see price, volume and vector candles.
Climax upClimax downElevated upElevated down50 EMA
Hover a candle for its price and relative volume.Spread —
ABOVE & BELOW PRICE
Historical vector-candle ranges
Load a market to find the five largest zones on each side.
Last 500 candles. Zones shrink as price revisits them. Amounts estimate the whole original candle’s traded value, not value in the remaining fragment, outstanding orders, liquidation value or inventory. Partly recovered zones keep their original volume for ranking. Body / wick settings affect this map only; strategy rules always use bodies.
POSITION PLANNER
Know the loss before the trade.
Planned stop loss—
Position value—
Reward / risk after costs—
Token quantity—
Enter an entry, stop and target to calculate your position.
Fees are editable assumptions. Sizing includes both fees and adverse slippage. Gaps can exceed the planned loss.
UBTC and UETH are Unit representations of BTC and ETH, with bridge and custody dependencies. Market data comes from Hyperliquid; this desk does not connect to a wallet.
TEST THE PREMISE
Do vectors recover more often?
Compare vector candles with ordinary candles of similar body size and market conditions. A return to a price level is different from a profitable trade.
Load a market to test the recovery idea against a baseline.
How the comparison works
Uses only the last 30% of eligible loaded history. Each vector is matched to one unused ordinary candle within 30 days, with the same direction, completed higher-timeframe regime and body-size / ATR band (0.25–0.5, 0.5–1, 1–2 or 2+). Match selection never uses the subsequent outcome.
A 50% recovery means a later wick touches the body midpoint; full recovery means it touches the candle open. “Before moving away” requires midpoint recovery before another 1 ATR in the original direction. If both barriers touch in one candle and the open does not settle their order, it counts as ambiguous, not a success. Incomplete windows and windows with missing candles are excluded.
Matching is approximate. Observations overlap and are not independent. These are descriptive historical rates, not probabilities for the next trade, evidence of market-maker positions, or returns after costs.
MEASURE BEFORE YOU BELIEVE
Does this rule survive its costs?
Uses the loaded market and planner’s risk, fee and slippage settings. The last 30% of available history is reserved for a chronological holdout. Changing rules after reading it makes it part of your research.
Load a market, then run a test. No performance claims are preloaded.
Next-bar entries. One position at a time. Stop first if both exits touch in one candle. Maximum 24 bars per trade. Daily 1% loss lock and two-loss lock, measured on closed trades in UTC. Available history may still be short. Repeatedly inspected holdouts become research data.
FOLLOW EACH DECISION
Why did the setup qualify—or fail?
Uses the loaded market, setup and planner costs. The trace covers the last 30% of eligible history. It explains the existing rules without changing the fixed experiments.
Load a market, then analyze its decisions.
The chart stops at the decision time. An entry decision displays the next candle's opening price, without its later high or low. Outcomes are revealed separately. Historical spread and book depth are unavailable; these entry checks model price, costs and risk only.
TWO QUESTIONS. FIXED RULES.
Does volume improve the trade?
UBTC is the primary market. UETH and HYPE check whether the result repeats elsewhere. Every comparison uses 1h candles, 4h context and the same position sizing and costs.
Does climax volume help?
Compare the original bearish-climax sweep and confirmation with the same pattern allowing any volume. Both retain the same vector-zone target rule.
Loading the saved experiment results…
NEW DATA ONLY
Fixed forward observation
20 September–19 November 2026, UTC · end exclusive · 60 days.
The same two experiments are frozen for this future window. Evaluate the market currently loaded in the desk using 1h / 4h and extended history. This optional view evaluates loaded candles. The four-hour research job runs these comparisons automatically for all three markets; see Overview for its latest report.
Decision rules and limits
Keep the two hypotheses, markets, dates and thresholds fixed. A historical or forward screen needs at least 30 closed trades per arm in UBTC, positive vector-strategy return at normal and doubled costs, and higher average net R than the control at both cost levels. The same checks must pass in at least one replication market. Until the entire future window is complete, every forward result is provisional.
Thirty trades is a minimum checkpoint, not a statistical guarantee. A passing screen only earns a manual review. These observations can overlap and be correlated across markets. No live trading is authorized by a score.
The control allows any volume, so it can include vectors. Each arm follows its own entries, one-position limit and daily loss locks; this compares complete strategy paths, not identical matched trades. Costs are fixed at 0.08% fee and 0.05% slippage per side, with 0.25% risk and 25% maximum allocation from 1,000 USDC. The planner's editable settings do not alter these experiments.
The historical period was already examined in earlier research. It is a development sample. Changing a rule requires a new protocol and a new future window; do not select a winning market after seeing results.
Read the experiment report ↗Download frozen results and trades ↗
PRIVATE, AUTOMATIC PAPER RECORD
Observe the next setup.
Saved privately on Cloudflare, across devices. A background service checks once a minute even with the browser closed. Missed or delayed checks remain visible as coverage gaps. Captures observed bid/ask quotes, modeled entries, rejections and exits. No real orders.
Connecting to private paper storage…
Start uses the selected market, setup and planner costs, with 1h candles / 4h context. Settings are fixed for this recording session. Entries require a fresh quote, sufficient top-of-book size, ≤20 bps spread and ≥1.5R after costs. The first observation never backfills a trade.
How automatic outcomes are labeled
Entries use an observed ask plus assumed slippage and fees. Exits can use an observed bid or be reconstructed from completed one-minute candles, with stop first when both barriers touch. These are simulations, not broker fills. A missing price path or an ambiguous touch in the entry minute pauses the position for manual review. Missed entry opportunities are counted as gaps; they are never invented.
This recorder is separate from the frozen forward experiment. Its quote-based fills differ from historical next-open fills. Pausing stops new entries while continuing to check an open paper position.
YOUR EVIDENCE
Your paper journal
Practice without placing a trade. Open a market, enter an entry, stop and target in the position planner, then save the plan here.
Plans and manually entered outcomes are saved only in this browser. They are not broker fills. Export a backup before changing devices or clearing browser data. Import journal restores an Export journal JSON file after you review it. Files stay in your browser; the limit is 2 MiB and 2,000 plans.
Review your backup
Matching IDs are skipped. Existing plans keep their current outcomes, even if the backup contains a newer exit. Open imported plans count toward your journal’s usual limits.
NEW HERE? START HERE
You can help without being a trader.
Vector Desk is an experiment about trading activity and price. It checks whether unusually active periods are followed by different price movements, and separately whether fixed trading rules would make money after estimated costs.
A candle is a summary of prices over a set time, such as one hour. A vector candle is one our rule marks as unusually active. A paper trade is practice with no real purchase or sale.
We have not established a dependable trading advantage. You can read the results without placing a trade, connecting a wallet or creating an account. Data collection runs automatically, even while you are away.
1 · Read the date and finding
On Overview, check when the report was collected. “Too early” means the study cannot answer yet. The earlier historical comparison and the ongoing study are separate pieces of evidence.
2 · Tell us where you got stuck
Try to explain one result in your own words. Tell us which word, number or button made that difficult. A useful comment could be: “I thought 92% meant I would win 92% of trades. Is that what it means?” It does not.
3 · Send a short reply
Use the “Help improve Vector Desk” form on Overview and press Send feedback. Your message goes privately to the project owner; an email address is optional if you would like a reply. You can also email Vectordesk@frogclub.io. Your practice journal is not sent with your feedback or pooled into the research.
A simple example: returning to a price is not earning a profit
Suppose a candle opens at 100 and closes at 110. Halfway between them is 105: the body midpoint. If a later price touches 105 within the study’s time limit, that counts as a midpoint revisit. The study also checks how often this happens after comparable ordinary candles.
This does not tell us that buying or selling would have made money. That depends on the entry price, exit price and costs. A reported 90% revisit rate describes 90 out of every 100 observed cases in that sample; it is not a 90% chance of winning your next trade.
Plain-English glossary: the words used in the results
- Candle, body and wick
- A candle summarises the opening, highest, lowest and closing prices for a period. The body runs from open to close; wicks show prices reached beyond the body. “1h” means one hour; “4h” means four hours.
- Volume, vector and ordinary candle
- Volume is the amount traded during that period. Our fixed rule flags unusually active candles as vectors. An ordinary candle does not meet that rule. Green means the close was above the open; red means it was below.
- Recovery / midpoint revisit
- Recovery means a later return to an earlier price. In the Overview tables, a midpoint revisit means reaching halfway between the candle’s open and close. It does not mean recovering a financial loss.
- Matched pair
- One vector candle compared with one ordinary candle with similar recorded conditions. This makes the comparison more useful but does not make the observations independent or prove a cause.
- Vector filter / any-volume comparison
- Two versions of a trading rule: one requires the specified unusual activity at entry; the other does not. The second version can still include vector candles. This is different from the vector-versus-ordinary recovery comparison.
- Historical / forward study
- Historical results use earlier data already examined. A forward study uses new data collected after its rules were fixed.
- Settled / provisional
- A settled simulated trade has closed under the test rules. A provisional result uses the last available price for a still-open simulated trade and can change.
- Net return / double costs
- Net return is the simulated percentage gain or loss after estimated trading costs. Double costs repeats the calculation with higher assumed costs. Neither is money actually earned.
- ATR / EMA
- Average true range (ATR) measures recent price movement in price units. Exponential moving average (EMA) is a smoothed price average giving more weight to recent prices. The detailed rules below explain where they are used.
- Percentage points (pp)
- The difference between two percentages. For example, 92% minus 88% is 4 percentage points.
- UBTC, UETH, HYPE / USDC
- The three assets studied on the Hyperliquid market. UBTC and UETH represent Bitcoin and Ether on that venue; HYPE is its native token. Prices are quoted in USDC. These data do not cover every exchange.
- Waiting / pattern forming / pattern confirmed
- The full pattern is absent, developing, or present in the saved data. Even “confirmed” is a research check, not an instruction to trade.
Exact calculation and strategy rules (optional)
What a vector actually tells you
A climax candle has at least twice the previous 10 bars’ average volume, or a range × volume value at least as high as any of those bars. Elevated volume begins at 1.5×. Up and down describe the candle’s close versus open, not who bought or sold.
This reveals unusual activity. It cannot reveal a particular market maker’s inventory or prove that a price level must be revisited.
Choose a pace you can repeat
Start with 1h candles and 4h context. It gives you fewer decisions than scalping. Try 4h / daily if you cannot check hourly. Test 15m / 1h separately; smaller moves make costs more important.
The app does not select whichever timeframe happened to win recently. Use one market, one pace and one setup for a fixed evaluation period.
01 · Trend retest
Higher timeframe: close above a rising 50 EMA, which is above the 200 EMA, with at least 0.25 ATR separation. A bullish vector breaks the previous 10-bar high. Within eight bars, price touches the vector body midpoint and closes bullish above it and the 50 EMA.
Enter on the next bar. Stop below the lower of the vector and retest lows, minus 0.2 ATR. Target 2.5× the planned price risk. Cancel if price closes below the vector low or the trend fails.
02 · Sweep & recovery
Only outside a higher-timeframe downtrend. A bearish climax candle sweeps the previous 20-bar low, closes back above it, and has a lower wick at least 35% of its range. Within five bars, a close above that candle’s high confirms the setup.
Target the nearest still-untouched bearish-vector body midpoint above entry from the last 100 bars. Stop below the sweep low minus 0.2 ATR. Skip if no target offers at least 1.5R after costs.
When the answer is no trade
Skip stale or missing data, unconfirmed candles, wide spreads, a target too close after fees, or an exhausted setup. In a downtrend, this beginner version holds cash. Neither strategy averages into losses or uses leverage.
Backtests pause new entries after two closed losses or 1% realized loss in a UTC day. A stop is an intended exit, not a guaranteed execution price.
What sustainable would require
Positive average net return per unit of risk, tolerable drawdown, and results that persist on unseen periods and after higher costs. The current tests have not established that edge. Compare against ordinary candles first, then freeze a promising rule and test it on new paper outcomes.
Keep rules fixed while collecting evidence. Do not increase risk to recover losses. This is an original research implementation inspired by Traders Reality, not its official system or a validated source of income.