HEXATAPE
Tools
Not financial adviceData and analytics, for information only

You're seeing what each file holds. The files themselves, and every tool's hourly history, come with the In The Trenches plan.

See the plansMembers: open the full version
In The Trenches

Data pack

The numbers behind LIQ MAP, WIN LOSE and LUCK SCREEN in a clean Excel workbook, or as plain CSV files for your own code. Keep your own history and run your own tests on the same scan the pages show. Plus every tool's figures, hour by hour, the oldest from 30 September 2026.

7
Tables, each in Excel and CSV
2,473
Rows of data in this build
20
Coins on the liquidation map
226
Vaults screened
The filesCSV · one table per file

Times are UTC. Dollar amounts are in US dollars, rounded to the dollar. Prices are not rounded: they keep up to 8 significant digits. A column ending in _pct holds percent, so 5 means 5%. Decimals use a dot. Each file is UTF-8 with a header row, marked so Excel shows names in any language correctly. Open What each column means under a file for its columns, and A look inside for its first rows.

Everything in one Excel workbook

data_pack.xlsx

Every table below on its own sheet, ready to read: plain column names with their units, dollars and percents formatted, the header row frozen, filters on. A Read me sheet explains every column.

Sheets 8Size 160 KB
Get the files
Open the map

Liquidation levels, band by band

liqmap_bands.csv

Where forced selling and forced buying sit, coin by coin, in 0.5% steps of price out to 25% either side. Longs sit below the price and shorts above it, so in any band one side is zero. One row per coin per band.

Rows 2,000Columns 13Size 143 KBScan 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan ran.
coinThe Hyperliquid perpetual, named the way Hyperliquid names it.
price_usdThe last price the scan saw for the coin. Every band and move is measured from it.
band_from_pctWhere the band starts, as a move from the price in percent. Below the price is negative.
band_to_pctWhere the band ends, as a move from the price in percent.
band_from_price_usdThe price where the band starts.
band_to_price_usdThe price where the band ends.
longs_usd_measuredSize of the Hyperliquid longs the scan read whose liquidation price is in the band.
longs_count_measuredHow many long positions that is.
shorts_usd_measuredSize of the Hyperliquid shorts the scan read whose liquidation price is in the band.
shorts_count_measuredHow many short positions that is.
longs_usd_whole_market_estLongs in the band across every major venue: the measured Hyperliquid longs plus our estimate for the rest. Blank when no venue figures were fresh enough.
shorts_usd_whole_market_estShorts in the band across every major venue, measured plus estimated.
A look inside
scan_utccoinprice_usdband_from_pctband_to_pctband_from_price_usdband_to_price_usdlongs_usd_measuredlongs_count_measuredshorts_usd_measuredshorts_count_measuredlongs_usd_whole_market_estshorts_usd_whole_market_est
2026-10-07 12:35BTC83358-5.5-5.078773.3179190.13101931730011497849110
2026-10-07 12:35BTC83358-3.5-3.080440.4780857.26197368473007315805310
2026-10-07 12:35BTC833582.53.085441.9585858.7400854049610293784375
2026-10-07 12:35BTC833587.07.589193.0689609.8500351117810120780936

If the price moves

liqmap_moves.csv

How much gets force-closed if the price moves 1, 2, 5, 10 or 20%, counting every position whose liquidation price sits between here and there. One row per coin per move size.

Rows 100Columns 10Size 7 KBScan 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan ran.
coinThe Hyperliquid perpetual, named the way Hyperliquid names it.
price_usdThe last price the scan saw for the coin. Every band and move is measured from it.
move_pctThe size of the move, in percent.
price_down_usdThe price after a fall of that size.
price_up_usdThe price after a rise of that size.
longs_usd_if_down_measuredThe Hyperliquid longs the scan read that are liquidated on the way down to that price.
shorts_usd_if_up_measuredThe Hyperliquid shorts the scan read that are liquidated on the way up to that price.
longs_usd_if_down_whole_market_estLongs liquidated on the way down across every major venue, measured plus estimated.
shorts_usd_if_up_whole_market_estShorts liquidated on the way up across every major venue, measured plus estimated.
A look inside
scan_utccoinprice_usdmove_pctprice_down_usdprice_up_usdlongs_usd_if_down_measuredshorts_usd_if_up_measuredlongs_usd_if_down_whole_market_estshorts_usd_if_up_whole_market_est
2026-10-07 12:35BTC83358182524.4284191.58430826801596934440
2026-10-07 12:35BTC83358281690.8485025.161226647104546780740
2026-10-07 12:35BTC83358579190.187525.93540062385404961312185621293784375
2026-10-07 12:35BTC833581075022.291693.8147998308177691265485814567611239831
2026-10-07 12:35BTC833582066686.4100029.6407391030157182594151006567465406921109

What the scan read, coin by coin

liqmap_coins.csv

The figures each coin's map stands on: the price, open interest, how much of it the scan read, and the positions that sit outside the bands. One row per coin.

Rows 20Columns 19Size 4 KBScan 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan ran.
coinThe Hyperliquid perpetual, named the way Hyperliquid names it.
price_usdThe last price the scan saw for the coin. Every band and move is measured from it.
price_move_during_scan_pctHow far the price moved while the scan ran, in percent.
hl_open_interest_usdHyperliquid's open interest in the coin, in dollars.
hl_share_read_pctThe share of Hyperliquid's open interest the scan read, in percent.
positions_readPositions the scan read in this coin.
positions_with_liq_priceOf those, the positions with a liquidation price. The rest have margin for any move.
longs_held_usdTotal size of the long positions read, in dollars.
shorts_held_usdTotal size of the short positions read, in dollars.
isolated_margin_pctThe share of that size on isolated margin, in percent. The rest is cross margin.
crossed_countPositions whose liquidation price the price passed while the scan ran. They are not in the bands.
crossed_usdThe size of those positions, in dollars.
longs_outside_bands_usdLongs whose liquidation price is further below than the last band reaches.
shorts_outside_bands_usdShorts whose liquidation price is further above than the last band reaches.
oi_snapshot_utcWhen the other venues' open interest was read, from CoinGecko. Blank without venue figures.
whole_market_oi_est_usdOpen interest across Hyperliquid and every other venue in the estimate, in dollars.
measured_share_of_market_pctThe share of that open interest measured directly from Hyperliquid positions, in percent.
venues_in_market_estHow many venues the whole-market figures cover, Hyperliquid included.
A look inside
scan_utccoinprice_usdprice_move_during_scan_pcthl_open_interest_usdhl_share_read_pctpositions_readpositions_with_liq_pricelongs_held_usdshorts_held_usdisolated_margin_pctcrossed_countcrossed_usdlongs_outside_bands_usdshorts_outside_bands_usdoi_snapshot_utcwhole_market_oi_est_usdmeasured_share_of_market_pctvenues_in_market_est
2026-10-07 12:35BTC83358-0.31337661876434.46024531117277933120652096110.10058653151710064301032026-10-07 12:30414584704452.814

Leverage, longs against shorts

liqmap_leverage.csv

Position size by the leverage each account set for the coin, read from Hyperliquid positions. One row per coin per leverage bucket.

Rows 100Columns 8Size 6 KBScan 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan ran.
coinThe Hyperliquid perpetual, named the way Hyperliquid names it.
leverage_fromThe lowest leverage in the bucket. Blank for the lowest bucket.
leverage_belowThe bucket holds leverage below this. Blank for the highest bucket.
longs_usdSize of the long positions set to leverage in this bucket, in dollars.
shorts_usdSize of the short positions set to leverage in this bucket, in dollars.
longs_share_pctThis bucket's share of all long size in the coin, in percent.
shorts_share_pctThis bucket's share of all short size in the coin, in percent.
A look inside
scan_utccoinleverage_fromleverage_belowlongs_usdshorts_usdlongs_share_pctshorts_share_pct
2026-10-07 12:35BTC3469009894417950.40.8
2026-10-07 12:35BTC351546616134550723613.83.8
2026-10-07 12:35BTC510432315495279169163.943.8
2026-10-07 12:35BTC10202267568624358326520.33.6
2026-10-07 12:35BTC2068793781158007175061.648.1
Open the board

Winners vs losers, every coin

winlose_board.csv

The full board: what Hyperliquid's biggest winners, its biggest losers and this month's best hold in each coin, long and short. One row per coin.

Rows 24Columns 21Size 4 KBScan 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan ran.
leaderboard_utcWhen the leaderboard that picks the groups was read.
coinThe Hyperliquid perpetual, named the way Hyperliquid names it.
mark_at_scan_start_usdThe coin's mark price when the scan started. The map files' price_usd is the last price the scan saw.
winners_long_usdDollars the biggest winners hold long in the coin.
winners_short_usdDollars the biggest winners hold short in the coin.
winners_net_usdLong minus short for the biggest winners. Above zero they are betting up, below zero betting down.
winners_long_countHow many of the biggest winners are long.
winners_short_countHow many of the biggest winners are short.
losers_long_usdDollars the biggest losers hold long in the coin.
losers_short_usdDollars the biggest losers hold short in the coin.
losers_net_usdLong minus short for the biggest losers. Above zero they are betting up, below zero betting down.
losers_long_countHow many of the biggest losers are long.
losers_short_countHow many of the biggest losers are short.
month_long_usdDollars this month's best hold long in the coin.
month_short_usdDollars this month's best hold short in the coin.
month_net_usdLong minus short for this month's best. Above zero they are betting up, below zero betting down.
month_long_countHow many of this month's best are long.
month_short_countHow many of this month's best are short.
winners_losers_oppositeYes when the biggest winners and the biggest losers are betting opposite ways.
winners_month_oppositeYes when the biggest winners and this month's best are betting opposite ways.
A look inside
scan_utcleaderboard_utccoinmark_at_scan_start_usdwinners_long_usdwinners_short_usdwinners_net_usdwinners_long_countwinners_short_countlosers_long_usdlosers_short_usdlosers_net_usdlosers_long_countlosers_short_countmonth_long_usdmonth_short_usdmonth_net_usdmonth_long_countmonth_short_countwinners_losers_oppositewinners_month_opposite
2026-10-07 12:352026-10-07 00:30BTC83616375281711650097220-2748155093631146673066280041059-1333679932537488777908787102574100676511910noyes
2026-10-07 12:352026-10-07 00:30ETH2577.8471958143665039897-1930817533027189786902337020360-147233458164486022643553135923807090512279noyes
2026-10-07 12:352026-10-07 00:30HYPE89.329171347086304276249-13292916338332930582495192682-65886858153627793495219823315258111636216noyes

Who is in each group

winlose_groups.csv

Each group's size and how much of it the scan saw. The board page explains how each group is picked. One row per group.

Rows 3Columns 7Size 1 KBScan 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan ran.
leaderboard_utcWhen the leaderboard that picks the groups was read.
groupwinners, losers or month: the prefix the board file uses.
group_nameThe group as the board page names it.
traders_in_groupHow many traders the group holds.
traders_in_scanHow many of them the positions scan read.
traders_with_positionsHow many of those hold an open position.
A look inside
scan_utcleaderboard_utcgroupgroup_nametraders_in_grouptraders_in_scantraders_with_positions
2026-10-07 12:352026-10-07 00:30winnersBiggest winners400156156
2026-10-07 12:352026-10-07 00:30losersBiggest losers300119119
2026-10-07 12:352026-10-07 00:30monthThis month's best2008181
Open the screen

Every screened vault

luck_vaults.csv

Every open Hyperliquid vault the screen tested, with its record, its luck test, its flags and its verdict. Largest first. One row per vault.

Rows 226Columns 12Size 28 KBRecords 2026-10-06
Get the files
What each column means
records_dateThe date of the vault records the screen used.
vault_nameThe vault's name, as its owner set it.
vault_addressThe vault's address on Hyperliquid.
size_usdMoney in the vault, in dollars.
age_daysHow long the vault's record runs, in days.
return_per_year_pctReturn per year, in percent, with deposits and withdrawals taken out.
volatility_per_year_pctHow much the return swings in a year, in percent.
luck_probability_pctHow likely a trader with no skill would be to post a record this good, in percent. Under 5 is hard to put down to luck.
after_the_crowd_pctThe same chance, adjusted because hundreds of vaults are tested at once, in percent.
martingaleYes when the vault keeps adding to losing trades with bigger size.
smooth_recordYes when its ups and downs are very small, which can hide a tail risk.
verdictNot luck, Luck or Too short, as the screen shows it.
A look inside
records_datevault_namevault_addresssize_usdage_daysreturn_per_year_pctvolatility_per_year_pctluck_probability_pctafter_the_crowd_pctmartingalesmooth_recordverdict
2026-10-06Growi HF0x1e37a337ed460039d1b15bd3bc489de789768d5e13372500817.143.0023.200.379713.4164nonoLuck
2026-10-06drkmttr0xc179e03922afe8fa9533d3f896338b9fb87ce0c811191900299.1161.55132.7314.351336.6787nonoLuck
2026-10-06[ Systemic Strategies ] L/S Grids0x07fd993f0fa3a185f7207adccd29f7a87404689d10718800614.176.3953.683.717525.2639nonoLuck

Every tool's main figures, hour by hour, coin by coin: what each page showed at the time, kept as it was. Each file starts at the first hour we recorded for its tool, the oldest on 30 September 2026. Nothing is worked out again later with hindsight, and a blank cell means the figure did not exist yet at that hour.

The history in one Excel workbook

history.xlsx

Every tool below on its own sheet, hour by hour, ready to read. It holds up to the newest 14 days; the CSV files go further back.

Sheets 8Size 1107 KBFrom 2026-09-30 04:28 UTCTo 2026-10-07 13:10 UTC
Get the files

LIQ MAP, hour by hour

history_liqmap.csv

What a 1, 2, 5, 10 or 20% move would liquidate, longs on the way down and shorts on the way up, at every hourly scan. One row per coin per scan.

Hours 134Rows 2,680Size 252 KBFrom 2026-10-01 23:01 UTCTo 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan behind the map ran.
coinThe Hyperliquid perpetual, named the way Hyperliquid names it.
price_usdThe last price the scan saw. Every move below is measured from it.
hl_share_read_pctThe share of Hyperliquid's open interest the scan read.
longs_1pct_down_usd_measuredHyperliquid longs the scan read that a 1% fall would liquidate.
shorts_1pct_up_usd_measuredHyperliquid shorts the scan read that a 1% rise would liquidate.
longs_2pct_down_usd_measuredHyperliquid longs the scan read that a 2% fall would liquidate.
shorts_2pct_up_usd_measuredHyperliquid shorts the scan read that a 2% rise would liquidate.
longs_5pct_down_usd_measuredHyperliquid longs the scan read that a 5% fall would liquidate.
shorts_5pct_up_usd_measuredHyperliquid shorts the scan read that a 5% rise would liquidate.
longs_10pct_down_usd_measuredHyperliquid longs the scan read that a 10% fall would liquidate.
shorts_10pct_up_usd_measuredHyperliquid shorts the scan read that a 10% rise would liquidate.
longs_20pct_down_usd_measuredHyperliquid longs the scan read that a 20% fall would liquidate.
shorts_20pct_up_usd_measuredHyperliquid shorts the scan read that a 20% rise would liquidate.
longs_1pct_down_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.
shorts_1pct_up_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.
longs_2pct_down_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.
shorts_2pct_up_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.
longs_5pct_down_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.
shorts_5pct_up_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.
longs_10pct_down_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.
shorts_10pct_up_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.
longs_20pct_down_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.
shorts_20pct_up_usd_whole_market_estThe same across every major venue, measured plus our estimate. Blank before 2026-10-05.

WIN LOSE, hour by hour

history_winlose.csv

What Hyperliquid's biggest winners and losers held, coin by coin, at every hourly scan. One row per coin per scan.

Hours 134Rows 3,216Size 321 KBFrom 2026-10-01 23:01 UTCTo 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan behind the board ran.
coinThe Hyperliquid perpetual, named the way Hyperliquid names it.
mark_at_scan_start_usdHyperliquid's mark price when the scan started.
winners_long_usdDollars the biggest all-time winners hold long in the coin.
winners_short_usdDollars the biggest all-time winners hold short in the coin.
winners_net_usdLong minus short for the biggest all-time winners.
winners_long_countHow many of the biggest all-time winners are long.
winners_short_countHow many of the biggest all-time winners are short.
losers_long_usdDollars the biggest all-time losers hold long in the coin.
losers_short_usdDollars the biggest all-time losers hold short in the coin.
losers_net_usdLong minus short for the biggest all-time losers.
losers_long_countHow many of the biggest all-time losers are long.
losers_short_countHow many of the biggest all-time losers are short.
month_net_usdLong minus short for this month's best traders.

LEV CENSUS, hour by hour

history_leverage.csv

How much of each coin's money sat within 5% of liquidation, and how much at 20x or more, at every hourly scan. One row per coin per scan.

Hours 134Rows 2,680Size 172 KBFrom 2026-10-01 23:01 UTCTo 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan behind the census ran.
coinThe Hyperliquid perpetual, named the way Hyperliquid names it.
price_usdThe price each position's liquidation distance is measured from.
longs_read_usd_measuredTotal size of the Hyperliquid longs the scan read.
shorts_read_usd_measuredTotal size of the Hyperliquid shorts the scan read.
longs_near_liq_pct_measuredThe share of those longs whose liquidation price is within 5% of the price, or already passed.
shorts_near_liq_pct_measuredThe share of those shorts whose liquidation price is within 5% of the price, or already passed.
longs_near_liq_pct_whole_market_estThe same share across every major venue, measured plus our estimate. Blank before 2026-10-05.
shorts_near_liq_pct_whole_market_estThe same share across every major venue, measured plus our estimate. Blank before 2026-10-05.
size_at_20x_plus_pct_measuredThe share of the size read whose account set 20x leverage or more.

CARRY SCAN, hour by hour

history_carry.csv

Each coin's best two-exchange funding carry, after both legs' fees, every hour. One row per coin per hour.

Hours 56Rows 1,120Size 90 KBFrom 2026-10-05 06:14 UTCTo 2026-10-07 12:30 UTC
Get the files
What each column means
rates_utcWhen the funding rates were read.
coinThe coin, named the way most exchanges name it.
long_venueThe exchange to hold the long leg on: its funding is the lowest.
short_venueThe exchange to hold the short leg on: its funding is the highest.
long_funding_pct_a_yearFunding on the long leg, in percent a year.
short_funding_pct_a_yearFunding on the short leg, in percent a year.
gap_pct_a_yearShort leg funding minus long leg funding, in percent a year.
fees_round_trip_pctTaker fees to open and close both legs, in percent of one leg's size.
breakeven_hoursHow long the gap must last to pay those fees.
net_1d_pct_a_yearThe gap less the fees spread over a 1 day hold, in percent a year.
net_7d_pct_a_yearThe gap less the fees spread over a 7 day hold, in percent a year.
net_30d_pct_a_yearThe gap less the fees spread over a 30 day hold, in percent a year.
pay_clock_estimatedYes when a leg's funding clock is our estimate, not the exchange's own.

LIQ TAPE, hour by hour

history_liqtape.csv

The liquidations OKX published for the 24 hours to each hour, longs and shorts, with the cascades. One row per coin per hour.

Hours 97Rows 2,425Size 116 KBFrom 2026-10-03 13:09 UTCTo 2026-10-07 13:10 UTC
Get the files
What each column means
to_utcThe end of the 24 hours counted.
coinThe coin, as OKX names its contracts.
longs_24h_usdLongs liquidated on OKX in the 24 hours, from the sample OKX publishes.
shorts_24h_usdShorts liquidated on OKX in the 24 hours, from the same sample.
longs_1h_usdLongs liquidated in the last of those hours.
shorts_1h_usdShorts liquidated in the last of those hours.
orders_24hHow many liquidation orders that is.
cascades_24hBursts of liquidations far above the coin's usual pace, in the 24 hours.

CROWD MOOD, hour by hour

history_crowd.csv

The mood gauge and its seven readings, each with its rank against its own history, at every hourly scan. One row per coin per scan.

Hours 56Rows 1,120Size 109 KBFrom 2026-10-05 05:35 UTCTo 2026-10-07 12:35 UTC
Get the files
What each column means
scan_utcWhen the positions scan behind the gauge ran.
coinThe Hyperliquid perpetual, named the way Hyperliquid names it.
mood_scoreThe gauge: 0 is very fearful, 100 is very greedy. Blank when too few readings rank.
moodThe gauge in words, as the page shows it.
readings_usedHow many readings the gauge averages.
losers_minus_winners_long_ptsHow much more of their money the biggest losers have long than the biggest winners, in percentage points.
losers_minus_winners_long_pts_rankWhere that reading sits against its own history, 0 the lowest and 100 the highest.
near_liq_pctThe share of the money read within 5% of liquidation.
near_liq_pct_rankWhere that reading sits against its own history, 0 the lowest and 100 the highest.
long_share_pctThe share of the money read that is long.
long_share_pct_rankWhere that reading sits against its own history, 0 the lowest and 100 the highest.
pain_longs_pctOf what a 10% move would liquidate, the share that is longs.
pain_longs_pct_rankWhere that reading sits against its own history, 0 the lowest and 100 the highest.
funding_pct_a_yearFunding across the big exchanges, weighted by open interest, in percent a year.
funding_pct_a_year_rankWhere that reading sits against its own history, 0 the lowest and 100 the highest.
put_call_ratioPuts open for every call open on Deribit.
put_call_ratio_rankWhere that reading sits against its own history, 0 the lowest and 100 the highest.
skew_25d_pts30-day puts' minus calls' implied volatility at 25-delta, in volatility points.
skew_25d_pts_rankWhere that reading sits against its own history, 0 the lowest and 100 the highest.

VOL DESK, hour by hour

history_vol.csv

Implied against realized volatility, skew, put/call and dealer gamma for BTC and ETH, every hour. One row per coin per hour.

Hours 177Rows 354Size 24 KBFrom 2026-09-30 04:28 UTCTo 2026-10-07 13:05 UTC
Get the files
What each column means
chain_utcWhen Deribit's options were read.
coinBTC or ETH.
index_usdDeribit's index price.
iv_30d_pctThe 30-day implied volatility, a year, from the options.
realized_7d_pctHow much the price really moved over the past 7 days, a year.
realized_30d_pctThe same over 30 days. Blank until 30 days of readings exist.
iv_minus_realized_ptsThe 30-day implied vol minus realized vol, in volatility points: 30 days when there, else 7.
skew_25d_ptsCalls' minus puts' implied volatility at 25-delta near 30 days, in points.
skew_expiry_daysDays to the expiry the skew is read on.
put_call_oiPuts open for every call open.
dealer_gamma_usd_per_1pctWhat dealers trade to stay hedged for each 1% move. Blank when the trade record is missing.
gamma_flip_usdThe price where dealer gamma changes sign. Blank when there is none.

The same scan as the pages

Each file is written in the same build as the page it comes from, then read back and checked against that page's own numbers, cell by cell. If one cell disagrees, nothing is published.

Measured, or estimated

Columns ending in _measured come from the Hyperliquid positions the scan read: part of Hyperliquid, not scaled up. Columns ending in _whole_market_est add the rest of Hyperliquid and the other venues, estimated from the leverage Hyperliquid traders actually use. Read those as an estimate.

Safe to open

A name that starts with = + - or @ gets an apostrophe in front, so a spreadsheet reads it as text and never runs it as a formula. Numbers are left as numbers.

As it was at the time

Each hour's row is what that page showed then. Nothing is redone later with what we know now, and an hour the page skipped stays missing rather than filled in.