In the first half of the year 2025, EU natural gas imports from Russia was around 12.8 bcm (according to some Greenpeace research published today that I found being referenced in a news article.) Not only that, but "the EU seeks to phase out LNG imports from Russia by 2027," which is a strong signal that the trend will continue.
0.000037
Relative Brier Score
5
Forecasts
0
Upvotes
Forecasting Activity
Forecasting Calendar
| Past Week | Past Month | Past Year | This Season | All Time | |
|---|---|---|---|---|---|
| Forecasts | 0 | 0 | 5 | 5 | 5 |
| Comments | 0 | 0 | 5 | 5 | 5 |
| Questions Forecasted | 0 | 0 | 4 | 4 | 4 |
| Upvotes on Comments By This User | 0 | 0 | 0 | 0 | 0 |
| Definitions | |||||
New Prediction
New Prediction
Probability
Answer
1%
Yes
99%
No
Why do you think you're right?
Files
Why might you be wrong?
On just a policy level, even if there was a peace deal in Ukraine tomorrow, this would probably only lead to slow growth in the imports to the EU from Russia, not sudden ones which would impact Q4 2025. But there could be some unexpected event leading to the EU needing to import more natural gas all of a sudden, perhaps.
Files
New Prediction
Probability
Answer
30%
Less than or equal to 70
50%
Between 71 and 74, inclusive
15%
Between 75 and 78, inclusive
4%
Between 79 and 82, inclusive
1%
More than or equal to 83
Why do you think you're right?
The instructions say: "at the time of question launch, OpenAI’s highest scoring model is o4-mini (high) with a score of 70." But looking at the list on 24 Aug, o4-mini (high) has a score of 65, but the GPT-5 (medium/high) have scores of 68/69. Digging around the website a bit, I'm fairly confident that this is right (so I don't know what happened with the original statement of a score of 70).
OpenAI has just released the GPT-5 models a few weeks ago, and since there's only 4 months left to change the current situation, I don't expect great improvements to their best score during. I do however expect slight improvements, as the models keep being updated every so often. As a result, most (80%) of my probability mass goes on <=70 and the interval (71, 74), with slightly more to the higher end there, and the rest (20%) on anything above 75.
Files
Why might you be wrong?
If OpenAI are using the GPT-5 models internally to develop new models, methods for training etc., they may have somewhat significant jump (~5 points) in the next 4 months. They have undoubtedly already started with this development (I do believe that the zoomed-out view is that they are developing smarter models, period, not just focusing their attention on applications of agentic or other features).
We know that they keep the models for internal testing for at least a few months before they announce them publicly, so I do not think it is likely we would hear about it before the end of this year, but I'm not extremely confident in this.
Files
New Prediction
Probability
Answer
2%
Yes
98%
No
Why do you think you're right?
There is only a few days left for this to happen, so my prior is very low. I don't see any particular reason why the announcement should happen before 1 Aug 2025. Searching for sources for this, it seems the US and its allies have given *end of August* as the de facto deadline for agreeing to a deal.
Files
Why might you be wrong?
I haven't read into it *that much*, perhaps the end date signifies something that I have missed. And it could just be that the deal is announced a month earlier than many people expect.
Files
New Badge
Why do you think you're right?
I think permanently disabling a satellite is a very difficult task. So it would take a substantial amount of effort to actually accomplish that. (I think it has not happened in the past.) And the upside of such an attack is probably not so great anyway; having many satellites makes them resilient to malfunctions (and attacks).
Why might you be wrong?
Tensions between US and its adversaries could heat up in the coming year, and it could escalate to conflicts. I still think it's likely it would take this particular form, but it's certainly possible.