Headlines say Iran-US tensions escalating. The Polymarket question is more specific: will Strait of Hormuz traffic
return to normal by May 15? The market sits at 13.5% YES — already pricing that conflict outlasts the resolution date.
All 8 of our AI models converge on 11c. That's a 2.5-cent gap on the market price, with 9 cents of upside to target if
you're buying NO at 86c.
$1.4M of volume, 16 days to resolution. Each model uses different inputs — news cycle, historical conflict
resolutions, logistics signals. When 8 of 8 line up, it's not consensus by accident. It's the math literally not
finding a path where Hormuz normalizes in 16 days.
Not a contrarian call. A status quo call. Boring works because most things stay the same most of the time — and this
resolution date is too soon for things to change.
Full breakdown: https://oddsshift.com/radar/strait-of-hormuz-traffic-returns-to-normal-by-may-15
return to normal by May 15? The market sits at 13.5% YES — already pricing that conflict outlasts the resolution date.
All 8 of our AI models converge on 11c. That's a 2.5-cent gap on the market price, with 9 cents of upside to target if
you're buying NO at 86c.
$1.4M of volume, 16 days to resolution. Each model uses different inputs — news cycle, historical conflict
resolutions, logistics signals. When 8 of 8 line up, it's not consensus by accident. It's the math literally not
finding a path where Hormuz normalizes in 16 days.
Not a contrarian call. A status quo call. Boring works because most things stay the same most of the time — and this
resolution date is too soon for things to change.
Full breakdown: https://oddsshift.com/radar/strait-of-hormuz-traffic-returns-to-normal-by-may-15
Smart money on Polymarket doesn't sit still. Here's what one wallet did in two weeks.
April 15. Sells half their NO position on US-invade-Iran at 70c — entry was 44c. Booked profit: $7,366.
April 17 — two days later. Same wallet drops $20,000 on YES for Warsh as next Fed Chair.
The wallet's track record: 88.9% accuracy on contested calls, $49M traded lifetime, 1,076 resolved trades.
This is the pattern most people miss. By the time someone tweets "this whale just made $7K," the wallet has already
redeployed into the next conviction trade. The follow-up move is where the edge is — not the celebration of the closed
one.
Our copy-trading enters when they enter. Same second, same direction, scaled to your stake. 88.9% isn't 100%, but it's
a different game than picking based on headlines.
https://oddsshift.com/copy-trading/
April 15. Sells half their NO position on US-invade-Iran at 70c — entry was 44c. Booked profit: $7,366.
April 17 — two days later. Same wallet drops $20,000 on YES for Warsh as next Fed Chair.
The wallet's track record: 88.9% accuracy on contested calls, $49M traded lifetime, 1,076 resolved trades.
This is the pattern most people miss. By the time someone tweets "this whale just made $7K," the wallet has already
redeployed into the next conviction trade. The follow-up move is where the edge is — not the celebration of the closed
one.
Our copy-trading enters when they enter. Same second, same direction, scaled to your stake. 88.9% isn't 100%, but it's
a different game than picking based on headlines.
https://oddsshift.com/copy-trading/
9 markets resolved with prior AI verdicts. 8 right, 1 wrong.
The miss first: we ran BUY YES on Kon Knueppel for NBA Rookie of the Year. He didn't win. That's our 1 of 9.
The 8 we got right:
— Iran ceasefire by June 30 (YES)
— Iran-Israel/US conflict by May 15 (consensus YES → won YES)
— Bank of Japan no rate change (consensus YES → won YES)
— Cooper Flagg Rookie of the Year (BUY YES → won YES)
— Keldon Johnson Sixth Man of the Year (BUY YES → won YES)
— Barcelona won't win Champions League (BUY NO → won NO)
— Ottawa Senators won't win Eastern Conference (consensus NO → won NO)
— Ukraine retakes Rodynske by Apr 30 (consensus YES → won YES)
Knueppel context: we also ran a verdict on Cooper Flagg for the same Rookie award. Two markets, same race. We backed
both. Flagg took it, Knueppel didn't. So 1 for 2 on that specific race, 8 of 9 across the full board.
Why publish misses: prediction markets aren't a "we always win" service. Anyone selling 95%+ accuracy out of the gate
is fitting curves to history. Real track records have losses. Public log compounds only if the misses are visible
alongside the wins.
Going forward: every resolved market with a prior AI verdict lands on the ledger. Each new resolution updates the
table, not the marketing copy. 8 right, 1 wrong, 9 on the board so far.
🔗 https://oddsshift.com/radar
The miss first: we ran BUY YES on Kon Knueppel for NBA Rookie of the Year. He didn't win. That's our 1 of 9.
The 8 we got right:
— Iran ceasefire by June 30 (YES)
— Iran-Israel/US conflict by May 15 (consensus YES → won YES)
— Bank of Japan no rate change (consensus YES → won YES)
— Cooper Flagg Rookie of the Year (BUY YES → won YES)
— Keldon Johnson Sixth Man of the Year (BUY YES → won YES)
— Barcelona won't win Champions League (BUY NO → won NO)
— Ottawa Senators won't win Eastern Conference (consensus NO → won NO)
— Ukraine retakes Rodynske by Apr 30 (consensus YES → won YES)
Knueppel context: we also ran a verdict on Cooper Flagg for the same Rookie award. Two markets, same race. We backed
both. Flagg took it, Knueppel didn't. So 1 for 2 on that specific race, 8 of 9 across the full board.
Why publish misses: prediction markets aren't a "we always win" service. Anyone selling 95%+ accuracy out of the gate
is fitting curves to history. Real track records have losses. Public log compounds only if the misses are visible
alongside the wins.
Going forward: every resolved market with a prior AI verdict lands on the ledger. Each new resolution updates the
table, not the marketing copy. 8 right, 1 wrong, 9 on the board so far.
🔗 https://oddsshift.com/radar
$6 million is sitting on one question this week: how many times Elon Musk will tweet. Polymarket breaks the question
into brackets — ranges like 140-159, 160-179, 180-199 — and each bracket trades as its own YES/NO market. By Friday
close, the bracket that contains his actual tweet count resolves YES, every other one resolves NO.
The interesting part isn't "guess the number." It's the structural probability distribution. Money flows into brackets
based on how traders read the week — is Elon's news cycle loud, is there a Tesla earnings call, court hearings,
political flashpoints. By midweek one bracket usually starts to dominate, but the spread between brackets tells you
how confident the room actually is.
We built a model that forecasts which bracket wins. It runs across nine personas, not just Elon — Trump, Andrew Tate,
Ted Cruz, CZ, the White House account, Khamenei, the NYC Mayor account, Zelenskyy. Each has its own weekly bracket
market on Polymarket. The model pulls historical posting cadence, weekday distribution, correlation with public
events, and outputs a probability across the brackets for the next resolution.
It's in training mode. Every week we compare the model's predicted bracket against the actual resolved one, log the
gap, and recalibrate. We're not claiming accuracy yet — we're publicly tracking how the model learns. After a few more
weeks of resolved data, the picture gets clearer on whether there's real edge here or whether the market is already
pricing the pattern correctly.
Hub with all nine personas and live forecasts: oddsshift.com/xtweet-market. The market guesses. We model the pattern
into brackets — ranges like 140-159, 160-179, 180-199 — and each bracket trades as its own YES/NO market. By Friday
close, the bracket that contains his actual tweet count resolves YES, every other one resolves NO.
The interesting part isn't "guess the number." It's the structural probability distribution. Money flows into brackets
based on how traders read the week — is Elon's news cycle loud, is there a Tesla earnings call, court hearings,
political flashpoints. By midweek one bracket usually starts to dominate, but the spread between brackets tells you
how confident the room actually is.
We built a model that forecasts which bracket wins. It runs across nine personas, not just Elon — Trump, Andrew Tate,
Ted Cruz, CZ, the White House account, Khamenei, the NYC Mayor account, Zelenskyy. Each has its own weekly bracket
market on Polymarket. The model pulls historical posting cadence, weekday distribution, correlation with public
events, and outputs a probability across the brackets for the next resolution.
It's in training mode. Every week we compare the model's predicted bracket against the actual resolved one, log the
gap, and recalibrate. We're not claiming accuracy yet — we're publicly tracking how the model learns. After a few more
weeks of resolved data, the picture gets clearer on whether there's real edge here or whether the market is already
pricing the pattern correctly.
Hub with all nine personas and live forecasts: oddsshift.com/xtweet-market. The market guesses. We model the pattern
Three wallets just put $498,000 on the same Polymarket position: QatarEnergy LNG, YES side. Aenews2 took $400k, Bobe2
took $100k, the third wallet rounded out the rest. The setup fits a pattern we've now logged four times — Hezbollah
ceasefire, Hungary election, the April 22 cluster, and now this. Same small group of wallets, same coordination
window, same post-event re-pricing dynamic.
This is what we mean by alpha. Not a single whale taking a flier — a small group moving together within hours of a
structural shift in the underlying. Once three of them touch the same market in the same direction, the move usually
completes inside a week.
Separately, on Iran regime change before June 30: the NO side is stacked $323k across four weeks of accumulation.
Market sits at 0.935. That's 41% annualized if you read the position cold and let it run to resolution. 58 days. Not a
thesis, just math on the order book.
ChinaWhale's $1.17m on Taiwan invasion NO 2026 has been dormant four sessions — not a buy, not a sell, just held.
Eatpraylove rotated $44k of Iran-NO June 30 winners into shorter-horizon plays this week. Bobe2 redeploys $6m+ at
about 1.3% per day. The desk is moving.
End-of-month ledger on Substack. The wins and the misses, both. oddsshift.com.
took $100k, the third wallet rounded out the rest. The setup fits a pattern we've now logged four times — Hezbollah
ceasefire, Hungary election, the April 22 cluster, and now this. Same small group of wallets, same coordination
window, same post-event re-pricing dynamic.
This is what we mean by alpha. Not a single whale taking a flier — a small group moving together within hours of a
structural shift in the underlying. Once three of them touch the same market in the same direction, the move usually
completes inside a week.
Separately, on Iran regime change before June 30: the NO side is stacked $323k across four weeks of accumulation.
Market sits at 0.935. That's 41% annualized if you read the position cold and let it run to resolution. 58 days. Not a
thesis, just math on the order book.
ChinaWhale's $1.17m on Taiwan invasion NO 2026 has been dormant four sessions — not a buy, not a sell, just held.
Eatpraylove rotated $44k of Iran-NO June 30 winners into shorter-horizon plays this week. Bobe2 redeploys $6m+ at
about 1.3% per day. The desk is moving.
End-of-month ledger on Substack. The wins and the misses, both. oddsshift.com.
18-25% monthly. That's what alpha wallets cleared on Polymarket through April — not from thesis bets that resolved in
their favor, but from re-pricing on positions still open. The US-Iran invasion market this week shows the mechanic.
In March, a cohort of alpha wallets bought NO on "Will the US invade Iran before 2027?" at an average of 55c. The
market was still digesting Trump's rhetoric, the Pentagon's $1.45 trillion budget request, and the 60-day War Powers
timeline. Headlines pushed YES briefly into the mid-50s, then drifted. By early May, NO was sitting at 69c. This week
the same cohort started selling.
22% to 55% return in 30 days, depending on entry price. On a binary that resolves in December 2026 — seven months out.
The thesis didn't resolve. The market didn't resolve. They just took the move.
This is the part most retail traders miss on prediction markets. You don't need the underlying event to resolve in
your favor to get paid. You need the market to re-price toward your direction, then the discipline to lock the gain.
Alpha wallets did exactly that — entered against the headline narrative, held through the drift, and started exiting
before resolution.
Models still lean NO. The market still prices NO at 69c. The thesis is intact. The trade is just done.
their favor, but from re-pricing on positions still open. The US-Iran invasion market this week shows the mechanic.
In March, a cohort of alpha wallets bought NO on "Will the US invade Iran before 2027?" at an average of 55c. The
market was still digesting Trump's rhetoric, the Pentagon's $1.45 trillion budget request, and the 60-day War Powers
timeline. Headlines pushed YES briefly into the mid-50s, then drifted. By early May, NO was sitting at 69c. This week
the same cohort started selling.
22% to 55% return in 30 days, depending on entry price. On a binary that resolves in December 2026 — seven months out.
The thesis didn't resolve. The market didn't resolve. They just took the move.
This is the part most retail traders miss on prediction markets. You don't need the underlying event to resolve in
your favor to get paid. You need the market to re-price toward your direction, then the discipline to lock the gain.
Alpha wallets did exactly that — entered against the headline narrative, held through the drift, and started exiting
before resolution.
Models still lean NO. The market still prices NO at 69c. The thesis is intact. The trade is just done.
For 30 days this wallet held only cash. No new positions, no exits, no rebalancing — just dollars sitting in the
address. We log this kind of behavior internally as a sleeper phase: a holding pattern where conviction is being
assembled but not yet expressed. It's the part of alpha most observers miss, because there's nothing to copy.
Yesterday, that changed. $90,000 went out the door in 24 hours across three markets — $48,000 of it on "Iran closes
its airspace by May 8?" NO side, entered at 75c. Plus $17,000 on an NBA market and $5,000 on Steyer for California
Governor.
The Iran airspace deploy was live on OddsShift the moment it hit chain. Anyone following this wallet had it in their
daily digest this morning. By the time most readers see this, the move has already started: market is at 83c, up 11%
in 24 hours, and one tracked position on this market is sitting on $12,000 unrealized.
This is the part of prediction markets retail rarely sees. The wallet wasn't holding cash because it was lazy — it was
positioning to deploy when a specific trigger arrived. Whether the trade resolves in its favor by Friday or not, the
cohort that's been right before is now expressing a directional view in real time, and the price is moving with it.
We're watching whether the same wallet adds to this position in the next 24 hours. If it does, the signal compresses
further.
Market here: https://oddsshift.com/radar/iran-closes-its-airspace-by-may-8
address. We log this kind of behavior internally as a sleeper phase: a holding pattern where conviction is being
assembled but not yet expressed. It's the part of alpha most observers miss, because there's nothing to copy.
Yesterday, that changed. $90,000 went out the door in 24 hours across three markets — $48,000 of it on "Iran closes
its airspace by May 8?" NO side, entered at 75c. Plus $17,000 on an NBA market and $5,000 on Steyer for California
Governor.
The Iran airspace deploy was live on OddsShift the moment it hit chain. Anyone following this wallet had it in their
daily digest this morning. By the time most readers see this, the move has already started: market is at 83c, up 11%
in 24 hours, and one tracked position on this market is sitting on $12,000 unrealized.
This is the part of prediction markets retail rarely sees. The wallet wasn't holding cash because it was lazy — it was
positioning to deploy when a specific trigger arrived. Whether the trade resolves in its favor by Friday or not, the
cohort that's been right before is now expressing a directional view in real time, and the price is moving with it.
We're watching whether the same wallet adds to this position in the next 24 hours. If it does, the signal compresses
further.
Market here: https://oddsshift.com/radar/iran-closes-its-airspace-by-may-8
For six months, the single largest position on Polymarket sat untouched. ChinaWhale holding $1.17 million on "Will
China invade Taiwan by end of 2026" — NO side. No new trades. No exits. Just dollars in the address, day after day.
Today the wallet trimmed it. $31,000 sold — about 3% of the total position. In absolute terms, tiny. As a signal, the
first scaling action in half a year.
This is the part of prediction-market tracking that doesn't get covered. When a whale enters a position, that's the
easy headline. When a whale exits in a hurry, that's the obvious panic story. But when the largest single position on
the entire platform sits frozen for half a year and then makes a small calibrated trim — that's structural
information. Either the wallet's read on the underlying has shifted, or risk management on a long hold is finally
rotating.
The market itself sits at the high end of consensus on NO. Tracked alpha cohort all aligned that direction.
ChinaWhale's $1.17M was anchoring the conviction floor. A 3% trim doesn't break the thesis — it signals the wallet is
starting to rebalance.
We're watching whether this is the start of a scaling cycle. If the next 7-14 days show further trims, conviction is
rotating. If it stops at 3% and stays, this was a one-time risk-management touch. Either pattern tells you something
the market price alone won't.
https://oddsshift.com/radar/will-china-invade-taiwan-before-2027
China invade Taiwan by end of 2026" — NO side. No new trades. No exits. Just dollars in the address, day after day.
Today the wallet trimmed it. $31,000 sold — about 3% of the total position. In absolute terms, tiny. As a signal, the
first scaling action in half a year.
This is the part of prediction-market tracking that doesn't get covered. When a whale enters a position, that's the
easy headline. When a whale exits in a hurry, that's the obvious panic story. But when the largest single position on
the entire platform sits frozen for half a year and then makes a small calibrated trim — that's structural
information. Either the wallet's read on the underlying has shifted, or risk management on a long hold is finally
rotating.
The market itself sits at the high end of consensus on NO. Tracked alpha cohort all aligned that direction.
ChinaWhale's $1.17M was anchoring the conviction floor. A 3% trim doesn't break the thesis — it signals the wallet is
starting to rebalance.
We're watching whether this is the start of a scaling cycle. If the next 7-14 days show further trims, conviction is
rotating. If it stops at 3% and stays, this was a one-time risk-management touch. Either pattern tells you something
the market price alone won't.
https://oddsshift.com/radar/will-china-invade-taiwan-before-2027
World Cup 2026 — group advance is a re-pricing event.
We tracked the 7 tournament favorites and measured how much each one's title odds jump the moment they escape the
group stage.
Pre-tourney → Post-group advance (relative jump):
— Germany: 5% → 7% (+33%)
— Argentina: 8% → 10% (+25%)
— Portugal: 9% → 11% (+22%)
— Brazil: 9% → 11% (+22%)
— Spain: 17% → 20% (+18%)
— England: 11% → 13% (+18%)
— France: 18% → 20% (+14%)
The pattern: R² = 0.94 inverse fit.
The smaller a team's pre-tournament base rate, the bigger their relative re-pricing on group advance. Germany leaps
2.3× harder than France.
If you want torque on a group-stage trade, buy low-base favorites.
Data: Polymarket + Kalshi, weighted means. T-21 to kick-off.
Daily breakdown starts now.
oddsshift.com/radar
We tracked the 7 tournament favorites and measured how much each one's title odds jump the moment they escape the
group stage.
Pre-tourney → Post-group advance (relative jump):
— Germany: 5% → 7% (+33%)
— Argentina: 8% → 10% (+25%)
— Portugal: 9% → 11% (+22%)
— Brazil: 9% → 11% (+22%)
— Spain: 17% → 20% (+18%)
— England: 11% → 13% (+18%)
— France: 18% → 20% (+14%)
The pattern: R² = 0.94 inverse fit.
The smaller a team's pre-tournament base rate, the bigger their relative re-pricing on group advance. Germany leaps
2.3× harder than France.
If you want torque on a group-stage trade, buy low-base favorites.
Data: Polymarket + Kalshi, weighted means. T-21 to kick-off.
Daily breakdown starts now.
oddsshift.com/radar
Germany. The biggest leap in our cluster.
Yesterday we showed the pattern: smaller pre-tournament favorite = bigger relative jump on group advance.
Germany sits at the extreme end:
Pre-tournament: 5%
Conditional on group advance: 7%
Relative jump: +33%
Context — Germany is ranked 7th of the 7 favorites we track. The lowest base rate of any team with serious World Cup
pedigree (4× champion: 1954, 1974, 1990, 2014).
Why +33% is the largest leap of the cluster:
The market prices a team on visible form. Germany's recent form has been mediocre — group-stage exits in 2018 and
2022. The base rate of 5% reflects that pessimism.
But the moment they survive the group, the market re-prices fast. Pedigree converts back into a price signal. The
discount on doubt evaporates.
Smaller the favorite, the bigger the re-pricer. Germany is the extreme case.
Tomorrow: the R²=0.94 reveal — why this pattern holds across all 7.
oddsshift.com/radar
Yesterday we showed the pattern: smaller pre-tournament favorite = bigger relative jump on group advance.
Germany sits at the extreme end:
Pre-tournament: 5%
Conditional on group advance: 7%
Relative jump: +33%
Context — Germany is ranked 7th of the 7 favorites we track. The lowest base rate of any team with serious World Cup
pedigree (4× champion: 1954, 1974, 1990, 2014).
Why +33% is the largest leap of the cluster:
The market prices a team on visible form. Germany's recent form has been mediocre — group-stage exits in 2018 and
2022. The base rate of 5% reflects that pessimism.
But the moment they survive the group, the market re-prices fast. Pedigree converts back into a price signal. The
discount on doubt evaporates.
Smaller the favorite, the bigger the re-pricer. Germany is the extreme case.
Tomorrow: the R²=0.94 reveal — why this pattern holds across all 7.
oddsshift.com/radar