How the fund decides
Finance, politics and a lot of other complex things are extremely hard to predict, and most of the time they can’t be predicted at all. The way people react to them is far more predictable. They give some news too much weight and other news too little, they forget what came before, and they lose the context. Around every event, markets and the media fill with noise, and prices stop matching the facts.
We collect data of many kinds, and AI helps us work through it. We use it to separate the noise from what matters.
Then comes analysis: technical and fundamental analysis, valuation, options pricing and position sizing.
On top of that comes science: information theory, chaos theory, game theory, quantum mechanics and Monte Carlo simulation.
We build our own tools: to scan markets, to turn options prices into odds, to keep score of every call, and to spot insiders on Polymarket and Kalshi. We skip those markets more often than others. Some of our tools we are not ready to disclose yet.
When the data has said all it can, the fund decides: trade, or skip.
How we invest
A forecast shows one future. The real future can go thousands of ways. We invest for many outcomes, not one.
We do not build on a single forecast. We doubt the consensus. We update our view often and trade rarely.
The fund
Founded by Dastan Abikov. The fund invests in stocks, and in prediction markets on politics and other events. It is open to qualified investors.
Odds on what comes next
As of Nov 15, 2026
Jan 202622 callsJan 2026, 2 calls. The odds moved 18 points toward the calls since they were made.
The war in Ukraine stays hard to resolvePolymarket: 96 percent. Down 3 points since the call
The war between Ukraine and Russia will be hard to resolve, and there is too much noise around it as well.
Percentages are the chance each market gives the call.
- PolymarketUkraine signs peace deal with Russia by December 31, 2026? (opens in a new tab)96% now · 47% at the call, Jan 1
- At call
- 47%
- Jan 1
- +1 mo
- 54%
- Feb 1
- +3 mo
- 76%
- Apr 1
- Low
- 45%
- Jan 4
- High
- 96%
- Nov 14
- Now
- 96%
- Nov 15
- PolymarketRussia x Ukraine ceasefire by end of 2026? (opens in a new tab)Agreement announced · 56% at the call, Jan 1
- At call
- 56%
- Jan 1
- +1 mo
- 55%
- Feb 1
- +3 mo
- 72%
- Apr 8
- Low
- 0%
- May 9
- High
- 75%
- Apr 17
- Result
- 0%
- May 9
- PolymarketRussia x Ukraine ceasefire by December 31, 2026? (opens in a new tab)93% now · 68% when it opened, May 16
This market needs a ceasefire that takes effect and holds for ten days.
- Opened
- 68%
- May 16
- +1 mo
- –
- +3 mo
- –
- Low
- 65%
- May 18
- High
- 94%
- Nov 11
- Now
- 93%
- Nov 15
Russia and Ukraine announced a ceasefire agreement in May 2026, but no ceasefire has held for ten days since.
- PolymarketUkraine signs peace deal with Russia by December 31, 2026? (opens in a new tab)96% now · 47% at the call, Jan 1
Europe’s current leaders lose most of their electionsIncumbents lost: 4 of 6. Up 39 points since the call
Europe’s current leaders will lose most of their elections.
No single market prices this call. Below is every national parliamentary election in the EU since, with the market on whether the government in office in January held on.
Percentages are the chance each market gives the call.
- Slovenia, Mar 22 · PolymarketWill Robert Golob be the next Prime Minister of Slovenia? (opens in a new tab)Incumbent lost · 89% when it opened, Jan 19
- Opened
- 89%
- Jan 19
- +1 mo
- 86%
- Feb 1
- +3 mo
- 31%
- Apr 1
- Low
- 21%
- Mar 22
- High
- 100%
- May 15
- Result
- 100%
- May 31
- Denmark, Mar 24 · PolymarketWill Mette Frederiksen be the next prime minister of Denmark after the 2026 parliamentary elections? (opens in a new tab)Incumbent held on · 30% when it opened, Feb 27
- Opened
- 30%
- Feb 27
- +1 mo
- –
- +3 mo
- 6%
- Apr 8
- Low
- 0%
- Jun 4
- High
- 37%
- May 10
- Result
- 0%
- Jun 4
- Hungary, Apr 12 · PolymarketWill Fidesz–KDNP win the most seats in the next Hungarian parliamentary election? (opens in a new tab)Incumbent lost · 62% at the call, Jan 1
- At call
- 62%
- Jan 1
- +1 mo
- 61%
- Feb 1
- +3 mo
- 75%
- Apr 8
- Low
- 56%
- Jan 29
- High
- 100%
- Apr 13
- Result
- 100%
- Apr 13
- Bulgaria, Apr 19 · PolymarketWill Rosen Zhelyazkov be the next prime minister of Bulgaria after the 2026 parliamentary election? (opens in a new tab)Incumbent lost · 83% when it opened, Mar 13
- Opened
- 83%
- Mar 13
- +1 mo
- –
- +3 mo
- 100%
- Apr 8
- Low
- 83%
- Mar 13
- High
- 100%
- Apr 8
- Result
- 100%
- May 8
- Malta, May 30 · PolymarketWill Robert Abela be the next Prime Minister of Malta following the 2026 Maltese general election? (opens in a new tab)Incumbent held on · 35% when it opened, May 1
- Opened
- 35%
- May 1
- +1 mo
- –
- +3 mo
- –
- Low
- 0%
- May 31
- High
- 35%
- May 1
- Result
- 0%
- Jun 1
- Sweden, Sep 13 · PolymarketWill Ulf Kristersson be the next Prime Minister of Sweden? (opens in a new tab)76% now · 88% when it opened, Jan 19
Sweden voted on Sep 13; the next government is still being formed.
- Opened
- 88%
- Jan 19
- +1 mo
- 85%
- Feb 1
- +3 mo
- 64%
- Apr 1
- Low
- 34%
- Nov 9
- High
- 96%
- Sep 17
- Now
- 76%
- Nov 15
- Latvia, Oct 3 · PolymarketWill Jaunā Vienotība (JV) win the most seats in the 2026 Latvian parliamentary election? (opens in a new tab)Incumbent lost · 60% at the call, Jan 1
The United List won the most seats; the next government is still being formed.
- At call
- 60%
- Jan 1
- +1 mo
- 59%
- Feb 1
- +3 mo
- 78%
- Apr 4
- Low
- 51%
- Feb 14
- High
- 100%
- Aug 24
- Result
- 100%
- Oct 4
- Spain, Nov 29 · KalshiWill Pedro Sánchez become Prime Minister of Spain following the next Spanish general election? (opens in a new tab)83% now · 89% when it opened, Jul 9
Spain votes on Nov 29.
- Opened
- 89%
- Jul 9
- +1 mo
- –
- +3 mo
- –
- Low
- 77%
- Oct 24
- High
- 93%
- Jul 17
- Now
- 83%
- Nov 15
- Slovenia, Mar 22 · PolymarketWill Robert Golob be the next Prime Minister of Slovenia? (opens in a new tab)Incumbent lost · 89% when it opened, Jan 19
Feb 20260Biggest movesFeb 2026, no calls. Biggest moves: 2 markets.
No call this month. What moved most in politics, world events and AI:
- PolymarketUS strikes Iran by February 28, 2026? (opens in a new tab)33% on Feb 1 → Yes on Feb 28
- PolymarketWill Anthropic have the best AI model at the end of February 2026? (opens in a new tab)3% on Feb 1 → Yes on Feb 28
Mar 202611 callMar 2026, 1 call. The odds moved 8 points toward the calls since they were made.
The US–Iran conflict stays hard to resolveKalshi: 94 percent, Polymarket: 93 percent. Up 8 points since the call
The conflict between the United States and Iran will be hard to resolve, and Trump is just making too much noise.
Percentages are the chance each market gives the call.
- KalshiWill the US agree to a new Iranian nuclear deal before Jan 1, 2027? (opens in a new tab)94% now · 37% at the call, Mar 1
- At call
- 37%
- Mar 1
- +1 mo
- 61%
- Apr 1
- +3 mo
- 45%
- Jun 1
- Low
- 22%
- Apr 17
- High
- 97%
- Nov 13
- Now
- 94%
- Nov 15
- PolymarketUS-Iran nuclear deal by June 30? (opens in a new tab)Deal reached · 41% at the call, Mar 1
- At call
- 41%
- Mar 1
- +1 mo
- 76%
- Apr 1
- +3 mo
- 67%
- Jun 1
- Low
- 0%
- Jun 15
- High
- 84%
- Jun 10
- Result
- 0%
- Jun 15
- PolymarketUS–Iran final nuclear deal by December 31, 2026? (opens in a new tab)93% now · 53% when it opened, Jun 26
- Opened
- 53%
- Jun 26
- +1 mo
- –
- +3 mo
- –
- Low
- 53%
- Jun 26
- High
- 94%
- Nov 4
- Now
- 93%
- Nov 15
- PolymarketUS announces end of Iranian blockade by December 31? (opens in a new tab)81% now · 16% when it opened, Jul 27
- Opened
- 16%
- Jul 27
- +1 mo
- –
- +3 mo
- –
- Low
- 4%
- Aug 5
- High
- 81%
- Nov 14
- Now
- 81%
- Nov 15
In mid-June the US and Iran agreed a nuclear deal and a permanent peace deal, so the June market resolved yes.
It didn’t last. The US reinstated its naval blockade on July 13, and all 21 deadlines markets set for lifting it, from July 14 to November 2, passed with no end.
- KalshiWill the US agree to a new Iranian nuclear deal before Jan 1, 2027? (opens in a new tab)94% now · 37% at the call, Mar 1
Apr 202611 callApr 2026, 1 call. The odds moved 11 points toward the calls since they were made.
The AI revolution is still earlyPolymarket: 92 percent. Up 11 points since the call
Only about 3% of Americans pay $20 a month for an AI subscription. I’m already spending $5,000 a month. We’re still so early in this revolution. I’d say every industry is going through a reset right now.
From note 04, published on Apr 8, 2026.
Percentages are the chance each market gives the call.
- PolymarketAI bubble burst in 2026? (opens in a new tab)92% now · 81% at the call, Apr 8
- At call
- 81%
- Apr 8
- +1 mo
- 72%
- May 8
- +3 mo
- 84%
- Jul 8
- Low
- 72%
- May 8
- High
- 94%
- Oct 6
- Now
- 92%
- Nov 15
- PolymarketAI bubble burst by June 30, 2027? (opens in a new tab)82% now · 80% when it opened, Sep 22
- Opened
- 80%
- Sep 22
- +1 mo
- –
- +3 mo
- –
- Low
- 74%
- Nov 6
- High
- 85%
- Sep 28
- Now
- 82%
- Nov 15
A downturn here means at least three of six signs within 90 days, such as Nvidia’s shares falling 50% from their high.
- PolymarketAI bubble burst in 2026? (opens in a new tab)92% now · 81% at the call, Apr 8
May 20260Biggest movesMay 2026, no calls. Biggest moves: 3 markets.
No call this month. What moved most in politics, world events and AI:
- PolymarketUS announces new Iran agreement/ceasefire extension by May 31? (opens in a new tab)80% on May 23 → 1% on May 31
- PolymarketStarmer out by May 31, 2026? (opens in a new tab)58% on May 11 → 0% on May 31
- PolymarketWill Abelardo de la Espriella win the 1st round of the 2026 Colombian presidential election? (opens in a new tab)6% on May 1 → 100% on May 31
Jun 202622 callsJun 2026, 2 calls. The odds moved 68 points toward the calls since they were made.
The Strait of Hormuz stays hard to reopenKalshi: 96 percent, Polymarket: 96 percent. Up 68 points since the call
The Strait of Hormuz will be hard to open again the way it was before.
Percentages are the chance each market gives the call.
- KalshiWill the 7-day average of transits through the Strait of Hormuz (IMF PortWatch) be above 60 before Jan 1, 2027? (opens in a new tab)96% now · 33% at the call, Jun 1
- At call
- 33%
- Jun 1
- +1 mo
- 29%
- Jul 1
- +3 mo
- 77%
- Sep 1
- Low
- 14%
- Jun 15
- High
- 98%
- Nov 13
- Now
- 96%
- Nov 15
- PolymarketStrait of Hormuz traffic returns to normal by December 31, 2026? (opens in a new tab)96% now · 24% at the call, Jun 1
- At call
- 24%
- Jun 1
- +1 mo
- 18%
- Jul 1
- +3 mo
- 73%
- Sep 1
- Low
- 9%
- Jun 16
- High
- 96%
- Nov 14
- Now
- 96%
- Nov 15
- KalshiWill the 7-day average of transits through the Strait of Hormuz (IMF PortWatch) be above 60 before Jan 1, 2027? (opens in a new tab)96% now · 33% at the call, Jun 1
Israel invades the West Bank before 2030No market
Israel is going to invade the West Bank within a few years, probably before 2030.
No market prices this call yet.
Jul 202611 callJul 2026, 1 call. The odds moved 11 points toward the calls since they were made.
Meta wins back its losses and builds great modelsSince the market opened on Oct 12Kalshi: 83 percent. Meta shares: $641.69 at the Nov 13 close. Up 11 points since the call
Meta’s court case is just noise, and Meta will win back its losses quickly. Zuckerberg’s earlier bet on the metaverse was a really bad one, but I don’t think Mark is going to fall behind in this AI race.
Given his business experience and the huge audience on all of Meta’s platforms, I expect great models from Meta. They have everything they need: compute, researchers, money and infrastructure. Zuckerberg has also moved his own office to the AI team.
Percentages are the chance each market gives the call.
- KalshiWill Meta have a higher Arena Score than OpenAI at the end of 2026? (opens in a new tab)83% now · 72% when it opened, Oct 12
- Opened
- 72%
- Oct 12
- +1 mo
- –
- +3 mo
- –
- Low
- 72%
- Oct 12
- High
- 93%
- Oct 13
- Now
- 83%
- Nov 15
Meta shares · META- At call
- $612.91
- Jul 1
- +1 mo
- $590.24
- Aug 3
- +3 mo
- $725.93
- Oct 1
- Low
- $539.03
- Jul 30
- High
- $777.59
- Sep 24
- Now
- $641.69
- Nov 13
Closing prices from Nasdaq.
- KalshiWill Meta have a higher Arena Score than OpenAI at the end of 2026? (opens in a new tab)83% now · 72% when it opened, Oct 12
Aug 20260Biggest movesAug 2026, no calls. Biggest moves: 3 markets.
No call this month. What moved most in politics, world events and AI:
- PolymarketUS ceasefire against Iran continues through August 31? (opens in a new tab)82% on Aug 18 → No on Aug 30
- PolymarketWill Anthropic's valuation hit $2.0T by December 31? (opens in a new tab)36% on Aug 1 → 82% on Aug 31
- PolymarketWill Anthropic have the best AI model at the end of August 2026? (opens in a new tab)94% on Aug 1 → Yes on Aug 31
Sep 202622 callsSep 2026, 2 calls. The odds moved 31 points toward the calls since they were made.
Anthropic’s lead in AI is priced too highKalshi: 90 percent, Polymarket: 92 percent. Up 24 points since the call
There is no way Anthropic should be priced at 90% to have the best AI model. The competition is just a few weeks behind, and people are creating noise here.
Kalshi and Polymarket each run a market every month on who has the best AI model at the end of it.
Percentages are the chance each market gives the call.
- End of September · KalshiBest AI on Sep 30, 2026? Claude (opens in a new tab)Anthropic won · 6% at the call, Sep 1
- At call
- 6%
- Sep 1
- +1 mo
- –
- +3 mo
- –
- Low
- 0%
- Sep 30
- High
- 26%
- Sep 3
- Result
- 0%
- Sep 30
- End of September · PolymarketWill Anthropic have the best AI model at the end of September 2026? (opens in a new tab)Anthropic won · 8% at the call, Sep 1
- At call
- 8%
- Sep 1
- +1 mo
- –
- +3 mo
- –
- Low
- 0%
- Sep 26
- High
- 20%
- Sep 3
- Result
- 0%
- Sep 30
- End of October · KalshiBest AI on Oct 31, 2026? Claude (opens in a new tab)Google won · 21% when it opened, Sep 29
- Opened
- 21%
- Sep 29
- +1 mo
- 69%
- Oct 1
- +3 mo
- –
- Low
- 21%
- Sep 29
- High
- 100%
- Oct 31
- Result
- 100%
- Oct 31
- End of October · PolymarketWill Anthropic have the best AI model at the end of October 2026? (opens in a new tab)Google won · 14% at the call, Sep 1
- At call
- 14%
- Sep 1
- +1 mo
- 64%
- Oct 1
- +3 mo
- –
- Low
- 12%
- Sep 15
- High
- 100%
- Oct 30
- Result
- 100%
- Oct 31
- End of November · KalshiBest AI on Nov 30, 2026? Claude (opens in a new tab)90% now · 74% when it opened, Nov 1
- Opened
- 74%
- Nov 1
- +1 mo
- –
- +3 mo
- –
- Low
- 72%
- Nov 3
- High
- 90%
- Nov 15
- Now
- 90%
- Nov 15
- End of November · PolymarketWill Anthropic have the best AI model at the end of November 2026? (opens in a new tab)92% now · 22% when it opened, Sep 18
- Opened
- 22%
- Sep 18
- +1 mo
- 64%
- Oct 1
- +2 mo
- 73%
- Nov 1
- Low
- 13%
- Sep 22
- High
- 92%
- Nov 14
- Now
- 92%
- Nov 15
- End of September · KalshiBest AI on Sep 30, 2026? Claude (opens in a new tab)Anthropic won · 6% at the call, Sep 1
Google’s silence ends with a big releaseKalshi: 74 percent, Polymarket: 77 percent. Alphabet shares: $295.77 at the Nov 13 close. Up 38 points since the call
I don’t think Google’s silence will last longer than half a year, a year at most. A big release is coming. Google is the foundation of the current AI race, and we shouldn’t forget about it.
Percentages are the chance each market gives the call.
- KalshiBest AI on Dec 31, 2026? Gemini (opens in a new tab)74% now · 7% at the call, Sep 1
- At call
- 7%
- Sep 1
- +1 mo
- 38%
- Oct 1
- +2 mo
- 48%
- Nov 1
- Low
- 5%
- Sep 4
- High
- 74%
- Nov 15
- Now
- 74%
- Nov 15
- PolymarketWill Google have the best AI model at the end of December 2026? (opens in a new tab)77% now · 9% at the call, Sep 1
- At call
- 9%
- Sep 1
- +1 mo
- 40%
- Oct 1
- +2 mo
- 54%
- Nov 1
- Low
- 7%
- Sep 2
- High
- 77%
- Nov 14
- Now
- 77%
- Nov 15
- KalshiWill Google release Gemini 4 before Jan 1, 2027? (opens in a new tab)Released · 91% at the call, Sep 1
- At call
- 91%
- Sep 1
- +1 mo
- 96%
- Oct 1
- +3 mo
- –
- Low
- 58%
- Sep 9
- High
- 100%
- Oct 13
- Result
- 100%
- Oct 13
- PolymarketGemini 4.0 released by November 30, 2026? (opens in a new tab)Released · 92% at the call, Sep 1
- At call
- 92%
- Sep 1
- +1 mo
- 98%
- Oct 1
- +3 mo
- –
- Low
- 87%
- Sep 9
- High
- 100%
- Oct 13
- Result
- 100%
- Oct 13
- PolymarketGemini Carbon released by November 30, 2026? (opens in a new tab)Released · 79% when it opened, Oct 26
- Opened
- 79%
- Oct 26
- +1 mo
- –
- +2 mo
- 92%
- Nov 1
- Low
- 64%
- Oct 27
- High
- 100%
- Nov 10
- Result
- 100%
- Nov 10
- KalshiTop AI model on Nov 30, 2026? Gemini 4 Carbon (High) (opens in a new tab)87% now · 40% when it opened, Nov 1
- Opened
- 40%
- Nov 1
- +1 mo
- –
- +3 mo
- –
- Low
- 40%
- Nov 1
- High
- 87%
- Nov 15
- Now
- 87%
- Nov 15
Alphabet shares · GOOGL- At call
- $335.02
- Sep 1
- +1 mo
- $338.24
- Oct 1
- +2 mo
- $278.99
- Nov 2
- Low
- $278.99
- Nov 2
- High
- $354.97
- Sep 21
- Now
- $295.77
- Nov 13
Closing prices from Nasdaq.
Google released Gemini 4 in October 2026.
- KalshiBest AI on Dec 31, 2026? Gemini (opens in a new tab)74% now · 7% at the call, Sep 1
Oct 20260Biggest movesOct 2026, no calls. Biggest moves: 1 market.
No call this month. What moved most in politics, world events and AI:
- PolymarketWill Google have the best AI model at the end of October 2026? (opens in a new tab)64% on Oct 1 → Yes on Oct 31
Nov 202655 callsNov 2026, 5 calls. The odds moved 8 points toward the calls since they were made.
Oracle gets back above $200 before summer 2027Robinhood options: 23 percent. Oracle shares: $125.35 at the Nov 13 close
Oracle is going to get back above $200 before summer 2027. If I see that AI is not moving fast enough, I’ll say so here.
- Robinhood optionsOracle closes above $200 at least once by Jun 17, 202723%
An estimate from option prices, which give 12% for a close above $200 on Jun 17 itself.
Oracle shares · ORCL- At call
- $119.22
- Nov 2
- +1 mo
- –
- +3 mo
- –
- Low
- $115.16
- Nov 5
- High
- $128.38
- Nov 11
- Now
- $125.35
- Nov 13
Closing prices from Nasdaq.
- Robinhood optionsOracle closes above $200 at least once by Jun 17, 202723%
GTA 6 lifts Take-Two’s sharesKalshi: 98 percent, Robinhood options: 47 percent. Take-Two shares: $227.82 at the Nov 13 close. Up 8 points since the call
Take-Two’s investors are lowering their expectations for GTA 6. But in today’s world GTA 6 is not just a game: it is culture and a brand, and there is no way AI can compete with it in the next two to three years. That is enough to lift the shares, and Take-Two is the best company at making games, even using AI.
Take-Two’s CEO has also said that game sales are beating expectations, and I know he is a very humble person who doesn’t make noise.
Percentages are the chance each market gives the call.
- KalshiWill GTA 6 break the record for the highest-grossing video game in 24 hours? (opens in a new tab)98% now · 90% at the call, Nov 1
- At call
- 90%
- Nov 1
- +1 mo
- –
- +3 mo
- –
- Low
- 88%
- Nov 4
- High
- 99%
- Nov 13
- Now
- 98%
- Nov 15
- Robinhood optionsTake-Two closes above $227.82, its Nov 13 close, on Jun 17, 202747%
Take-Two shares · TTWO- At call
- $192.10
- Nov 2
- +1 mo
- –
- +3 mo
- –
- Low
- $192.10
- Nov 2
- High
- $231.31
- Nov 11
- Now
- $227.82
- Nov 13
Closing prices from Nasdaq.
- KalshiWill GTA 6 break the record for the highest-grossing video game in 24 hours? (opens in a new tab)98% now · 90% at the call, Nov 1
The US and Iran don’t meet again this yearPolymarket: 75 percent
The US and Iran won’t hold another senior diplomatic meeting before the end of 2026.
Called on Nov 15, 2026.
- PolymarketNext US-Iran senior diplomatic meeting by December 31, 2026? (opens in a new tab)Yes 25% · No 75%
The Israel–Iran ceasefire breaks before the year endsPolymarket: 32 percent
The ceasefire between Israel and Iran won’t hold through the end of 2026.
Called on Nov 15, 2026.
- PolymarketIsrael x Iran ceasefire continues through December 31? (opens in a new tab)Yes 68% · No 32%
Hegseth is out as Defense Secretary by year endPolymarket: 15 percent
Pete Hegseth will be out as Secretary of Defense by the end of 2026.
Called on Nov 15, 2026.