On February 24 2022 Russia has started a major invasion of territories in Ukraine.
This market will resolve to YES, if after 1 year of this war, Russia's estimated losses will be more than 5230 vehicles and fuel tanks.
Ukraine's Armed Forces estimate Russia's losses daily, which is published by Kyiv Independent.
This market will resolve based on these estimations as of January 1, 2023 - or the nearest available date.
https://www.instagram.com/kyivindependent_official/
recent estimations:
🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ127 | |
| 2 | Ṁ115 | |
| 3 | Ṁ74 | |
| 4 | Ṁ32 | |
| 5 | Ṁ20 |
@bingeworthy The estimates are in the morning. 2/25 morning estimates reflect Russian losses as of 2/24, not 2/24 morning estimates. All of TK's previous markets related to the Russian invasion closed on the stated final day, but resolved to the estimates in the morning on the next day for this reason.
@XComhghall If we took the last tomorrows estimates, we would be looking at the Russian losses of 1 year + 1 day.
@XComhghall These markets are different. When measuring the total losses in 2022 or February, you should take the estimates on January 1st or March 1st. The problem is, we are looking at a yearly cycle, and February 24th 2022 is already in the calculation, so if you were to include February 24th 2023 you would be calculating the total Russian losses after 366 days.
@PatMyron Probably copied from https://manifold.markets/itsTomekK/war-in-ukraine-will-russia-lose-100. Other markets did not resolve estimates as of 1/1.
@bingeworthy Good point. But I do not think that is how we typically use the phrase 'after 365 days' or 'after 1 year'.
No priming. No other exchanges besides what is shown. Two separate chats. You can try it yourself.
Chat 1
This is a math question. On 2/24/2022, Russia invaded Ukraine. Assuming no ceasefires or interruptions, to answer the question, 'How many UAVs will Russia lose after 365 days of war', I should use the statistics as of which day? In other words, which day is 'after 365 days' for 2/24/2022?
Should I use the statistics at the end of 2/24/2023, at the beginning of 2/24/2023, or at the end of 2/23/2023?
Chat 2
This is a math question. Russia invaded Ukraine in the morning on 2/24/2022. To answer the question, 'How many UAVs will Russia lose after 1 day of war', should I use the statistics at the end of 2/25/2022, at the beginning of 2/25/2022, or at the end of 2/24/2022?
To answer the question, 'How many UAVs will Russia lose after 365 days of war', should I use the statistics at the end of 2/25/2023, at the beginning of 2/25/2023, at the end of 2/24/2023, at the beginning of 2/24/2023, or at the end of 2/23/2023?
@XComhghall I do not value ChatGPT's input on this matter. I define "365 days" as exactly 365 days after the invasion. The chat is simply incorrect, as the closest you would get to accurate is taking the data from the morning of 2/24/2023, as that date is the closest to 1 year after the initial invasion. I am confused as to why you do not understand this.
@bingeworthy ChatGPT is a language model. I believe it accurately reflects the common usage of the English language, in accordance with which,
2/24/2022 is the first day of war. After 1 day is EOD 2/25/2022. 2/23/2023 is the 365th day of the war. After 365 days is EOD 2/24/2024.
@XComhghall It would only be the end of day 2/24 if the war started at the beginning of 2/25. It didn't, it started at the beginning of day 2/24.
@bingeworthy Yes. With the war beginning on 2/24/22 in the morning, 2/24/2022 is the first day of war. EOD 2/25/22 is after 1 day. 2/23/2023 is the 365th day of the war. EOD 2/24/2024 is after 365 days.
It would only be the end of day 2/23 if the war started at the end of 2/22. It didn't, it started at the beginning of day 2/24.
@XComhghall End of 2/23 and beginning of 2/24 are the same thing. EOD 2/25 is not 1 day, it would be 2 days because it would be counting both 2/24 and 2/25.
@bingeworthy 2/24/2022 is the first day of war. After 1 day is EOD 2/25/2022.
That is just the way the language works.
@bingeworthy It would only be the end of day 2/23 if the war started at the beginning of 2/23. It didn't, it started at the beginning of day 2/24.
@XComhghall If you were setting a timer for 1 day at 00:00, would the timer go off at 01:00 the next day? Because that is what you are implying here.
@bingeworthy After one day refers to end of day, the next day, in English.
Wow, you used 3 questions to make GPT say what you want it to say. Good for you!
Why would anyone want GPT to say the correct answer when you can make it say what you want it to say? Why would anyone show the entire chat when you can take three screenshots of exchanges after who knows how many attempts?
This is a math question. Russia invaded Ukraine in the morning on 2/24/2022. To answer the question, 'How many UAVs will Russia lose after 1 day of war', should I use the statistics at the end of 2/25/2022, at the beginning of 2/25/2022, or at the end of 2/24/2022?
To answer the question, 'How many UAVs will Russia lose after 365 days of war', should I use the statistics at the end of 2/25/2023, at the beginning of 2/25/2023, at the end of 2/24/2023, at the beginning of 2/24/2023, or at the end of 2/23/2023?
After one day refers to end of day, the next day, in English.
"After one day" does not mean end of day, the next day. It means after one day. Twenty-four hours.
Wow, you used 3 questions to make GPT say what you want it to say.
Yes, I used 3 questions because it required 3 questions to elaborate to the extent needed.
Why would anyone want GPT to say the correct answer when you can make it say what you want it to say?
If you want to say that it is priming, that is fine. It probably is. But it also reflects the actual situation that we are dealing with. If reciting the exact situation is priming, that may be an indication that it should be interpreted in that fashion.
Why would anyone show the entire chat when you can take three screenshots of exchanges after who knows how many attempts?
That was the entire chat, and that was my first and only attempt.
This is a math question. Russia invaded Ukraine in the morning on 2/24/2022. To answer the question, 'How many UAVs will Russia lose after 1 day of war', should I use the statistics at the end of 2/25/2022, at the beginning of 2/25/2022, or at the end of 2/24/2022?
To answer the question, 'How many UAVs will Russia lose after 365 days of war', should I use the statistics at the end of 2/25/2023, at the beginning of 2/25/2023, at the end of 2/24/2023, at the beginning of 2/24/2023, or at the end of 2/23/2023?
I gave ChatGPT this exact prompt and it agreed with you. The reason that I showed those screenshots was not to prove that I am right, but to prove a point that ChatGPT is not a reliable source for an answer to this problem.
@bingeworthy "After one day" does not mean "in one day". It means end of day, the next day.
@XComhghall I just do not know what to tell you. People do not use language in this way. When someone says "After 1 year" they use it synonymously to "In 1 year". Both logically and linguistically my argument makes more sense.
@bingeworthy I do wonder at your confidence that you know either more than ChatGPT, which does not represent one person's opinion, but is based on strict, pure logic, and millions of people's use of the language.
@XComhghall The problem is that ChatGPT does not agree with you. One of your tests agrees with you, and one of my tests agrees with me. This is why I said that ChatGPT is not reliable. The other issue is that this is not just a lingustic issue, but a logical issue. If we were to take your system, we would be counting 2/24 twice, which fundamentally changes the question.
@bingeworthy Your test showed none but priming. My test was a question of math and logic. Besides, the war did not start at 0 on 2/24, but in the morning. If instead you tell ChatGPT the fact that the war started in the morning, it will honestly and accurately tell you that regardless of the time that it started in the day, "after one day" means the end of the following day.
Your test showed none but priming.
So did yours. You intentionally left out the fact that the war started at the beginning in your original prompt.
Besides, the war did not start at 0 on 2/24, but in the morning.'
Yes, but the beginning of 2/24/2023 is still closer to that time compared to the beginning of 2/25/2023.
If instead you tell ChatGPT the fact that the war started in the morning, it will honestly and accurately tell you that regardless of the time that it started in the day, "after one day" means the end of the following day.
Again, no. ChatGPT is not reliable. I tried this experiment and it gave me the opposite result.
@bingeworthy No, I did not.
This was a separate new chat:
This is a math question. Russia invaded Ukraine in the morning on 2/24/2022. To answer the question, 'How many UAVs will Russia lose after 1 day of war', should I use the statistics at the end of 2/25/2022, at the beginning of 2/25/2022, or at the end of 2/24/2022?
The unreliability is caused by the use of the words 'beginning of 2/24', which are ambiguous, and can be interpreted to be 0 on 2/24 or 24 on 2/23. Anything else, for example, 'morning of 2/24', ChatGPT will reliably and consistently answer that after 1 day is the end of the following day.
When 'beginning of 2/24' is interpreted to be 0, the end of 2/24 is after 1 day. But anything besides 0 -- the case with the Russian invasion -- after 1 day is the end of 2/25.
When 'beginning of 2/24' is interpreted to be 0, the end of 2/24 is after 1 day. But anything besides 0 -- the case with the Russian invasion -- after 1 day is the end of 2/25.
False.
