AFP Confirms Ukraine Gains

• The Moscow Times: Russia Lost Ground in Ukraine in April, a First Since Mid-2023 – AFP Analysis

• SpaceWar.com: Russia lost ground in Ukraine in April, a first since mid-2023: AFP analysis

• Japan Today: Ukraine and Russia declare separate truces (includes AFP territory data)

ISW Confirms Ukraine Gains:

• Primary ISW Report: Russian Offensive Campaign Assessment, May 2, 2026

• Monthly Data Summary: The Russia-Ukraine War Report Card, May 6, 2026

• Weekly Data Snapshot: The Russia-Ukraine War Report Card, April 29, 2026

@ShaneBo thanks for bringing these to my attention. There are two thorny issues here:

Firstly, the ISW report you linked shows a net gain for Russia with infiltrations included and a net loss with infiltrations excluded. I've never seen ISW provide two different figures like this, so which should we use?

Previously, ISW seems to have included infiltrations in net territorial changes. The monthly figures for November to March in the March 31 report match the "including infiltrations" figures in the May 2 report, which leads me to think including infiltrations should be viewed as the default, and excluding them is a novelty. But I could be convinced otherwise.

The second issue is that the AFP analysis is "based on ISW data" and not an independent estimate. Should we treat these as two distinct estimates when they are, effectively, coming from the same source?

This is really my fault for not anticipating a foreseeable problem. With the way the resolution criteria are phrased, I think you have a strong argument that these should count as two instead of one, but resolving solely on ISW data does seem contrary to the clear intention of the market.

Btw, nothing I've said here represents a concrete decision and I'm keen to hear arguments (from you or anyone else) on both issues

@a_l_e_x you have the classic letter vs spirit conflict. Your original criteria were clearly built to have a solid confirmation of Ukrainian gains. Now you have a situation where they are failing. And we know that at the same time all other trackers are showing Russian gains, albeit smaller than before. The spirit of the question suggests there's no YES yet. Based on the exact wording, there's a YES if you choose to accept the ISW change.

If this were Metaculus, they'd most likely annulled the question. Here you're making the decision.

@MaxA @a_l_e_x A few points to my argument I'll lay out here:

1. @MaxA I disagree a bit on the letter vs. spirit conflict. I'll argue below that ISW identifying and removing the infiltration data is providing a source that more accurately tracks territorial control.

   If the market's spirit was to track which side is "winning" ground, Control is the most objective/standard metric. Including infiltrations introduces "noise" from transient movements that often disappear within 24–48 hours. The fact that ISW explicitly broke out these figures for the first time in May suggests they recognized that the "Total" figure was masking a significant shift in actual control dynamics.

   The inclusion of infiltrations in previous reports was a limitation of the data available at the time. By providing a breakout in the May 2 report, ISW is effectively 'cleaning' their data. To resolve on the 'Infiltration' figure is to treat a temporary scouting party as equal to an established defensive trench line. AFP's recognition of this suggest an additional credentialed acknoledgement that this is the most accurate data concerning actual territiorial control. Other sources not included in this collection are agreeing, and others are holding off their usual announcements as they consider this advancement. ALL to say that I would argue that this resolves yes both in spirit and by the letter

2. ON ISW and AFW sharing I get this point but:

   1. ISW provides the raw spatial data. AFP performed its own independent analysis of that data to reach a net square-kilometer conclusion.

   2. AFP is an independent entity with its own fact-checking and editorial standards. Their decision to headline the "Net Loss" figure constitutes a distinct, professional confirmation that the data supports a "YES" outcome.

3. I think this is of course also defensible by the letter: If the criteria say "If a major tracker/agency reports a net gain/loss," and both ISW and AFP have published headlines/reports stating exactly that, then I'd argue the "spirit" is irrelevant. Maybe particularly more so in this case because we have a very specific set of sources versus a broader range where you might be able to get more into the spirit debate. The criteria have been met. However, as I argued above I do think the spirit has been met

4. Summary of Thoughts: On Infiltrations: We should prioritize the 'Excluding Infiltrations' figure because 'Control' is the only metric that satisfies the common-sense definition of 'territorial gain.' Infiltrations represent presence in a gray zone, not control. The fact that ISW felt the need to break this out for the first time in May suggests they viewed the 'Control' metric as the more accurate representation of the month's reality. Resolving on 'Infiltrations' essentially credits 'scouting' as 'possession.' Some sources are beginning to shift to this and are actively debating it, but seems to me the most legit sources are begining to favor the infiltration exclusion

   On AFP: While they use ISW data, their analysis is independent. AFP didn't just retweet ISW; they performed a net calculation and applied their own editorial standards to verify the '120 sq km' figure. In OSINT, we rarely have independent primary data (most use the same satellites); independence is found in the analysis of that data.

   If the criteria require a major agency to note a gain, AFP has done exactly that in a headline. Following the literal text of your criteria leads to a 'YES' resolution.

Is there neutral ground, like a no go zone due to drones, that they both could be carving up? Or am I misunderstanding?

@BodeyBaker is this in reference to how the infiltrations work? are counted/not counted?

@BodeyBaker Yes. With the proliferation of drones, the whole concept of zones of control is beginning to lose its meaning. The line of contact these days is small teams sitting in trenches or basements, with both sides' drones flying overhead. The soldiers can't go above ground freely, so they can be living with an enemy's team living in another trench a thousand feet behind their backs, not knowing that and having no means to do anything about it.

@ShaneBo It's in response to Alex saying this:

Firstly, the ISW report you linked shows a net gain for Russia with infiltrations included and a net loss with infiltrations excluded. I've never seen ISW provide two different figures like this, so which should we use?

Sorry for not quoting

@MaxA @BodeyBaker @a_l_e_x thanks Max for the clarification and for including the point on drones. I think the drone induced change in control mechanics is a big part of ISWs change in how they view things, and more accurately captures the reality of the situation. I'll add what I hope brings some additional clarity to my previous argument that takes note of the drone reality:

1. While "including infiltrations" may have been the previous default, that was an artifact of lower-fidelity data. Now, ISW has better visibility. To use an outdated, noisier metric when a cleaner one is provided is I think a choice to prioritize inaccurate history over current reality. As Max notes drones have fundamently shifted what accounts for control and infilitrations aren't really controlling anything

2. AFP specifically chose to highlight the "Net Loss" figure. If a major world news agency (AFP) and the primary data provider (ISW) both agree that "Control" has decreased, that is a double confirmation by the letter of the law.

3. The spirit of a "territorial gain" market is to see if the map is moving. If Russia is "gaining" by sending people into basements where they are immediately pinned and unable to be resupplied, but losing their established defensive lines elsewhere, that is a net loss of power and territory.

I argue that ISW and AFP have simply caught up to this reality by excluding "infiltrations" from their confirmed control tallies. Many other sources are begining to do the same

@ShaneBo We're looking at the change from month to month. The Ukrainian gains only exist in ISW data if the absolute value of the area under control as of the end of March calculated using the old system is compared to the same as of the end of April calculated using the new system. So, basically, apples to oranges. And infiltrations didn't suddenly change qualitatively during April - the fuzziness of the LOC has been an issue for many months (I can give you multiple links to DeepState's Telegram and YouTube channels discussing that problem).

So, I think the optimal approach would be to take the delta of the "inclusive" ISW April data to the ISW March data (implicitly "inclusive"), and then use the deltas for "exclusive" statistics for April to May, May to June etc.

It should be noted that both Shane and I are arguing in favor of the positions we're holding here.

@a_l_e_x @MaxA I think I track the core of the debate here. Max, your point on drone ambiguity I think actually strengthens the case for excluding "infiltration" data: it supports the argument that the ISW and AFP sources are matching not only the letter but the spirit of the market by tracking physical control over transient presence. I'm curious if we agree on that, and are just debating how we assess the change in methodology as it relates to this markets specific resolution?

However, to address the methodological "apples to oranges" concern specifically, we can look at the data continuity. The net loss is not a statistical artifact of the May 2nd methodology shift; it is a physical reality that was already being recorded in the "old" system.

The March 31/April 29 ISW-Russia Matters Report Card used the standard reporting system and already showed a negative delta for Russia before the formal "Infiltration Layer" was even introduced:

"During the prior four-week period (March 3–31, 2026), Russia suffered a net loss of 12 square miles. In the shorter term, Russia lost 7 square miles (net) during the past week (April 21–28, 2026)... according to ISW data."

— Source: The Russia-Ukraine War Report Card (Russia Matters/ISW), published April 29, 2026.

This proves that the April net loss is a physical reality on the ground, not a result of "re-categorizing" old territory in May. The "Infiltration" layer didn't create the loss; it simply provided the technical explanation for the delta between what the Kremlin was claiming and what they actually held.

The market resolution criteria explicitly state:

"This market will resolve YES if, in any month of 2026, Ukraine’s net territorial gain... is greater than zero square kilometres according to two or more of the following sources."

We have exactly that:

1. ISW: Explicitly noted a "net loss of territory controlled" for April 2026.

2. AFP: Performed an independent net calculation and made the editorial decision to headline it as a "Russian Loss."

To reinforce the two source assessement. In OSINT, independence is found in the analysis, not the raw satellite data (which everyone shares). AFP's choice to validate and publish these specific figures fulfills the requirement for a secondary source.

Bottom line: I don't think we should discount the assessement because the primary tracker became more precise. The criteria have been met both by the literal text and the operational reality of the front line. I stand by this being a clear yes

Max, let me know what you see some faults or missing info here.

@ShaneBo @MaxA Adding to this: Max, I fully grant some challenges in assessment that might make some lean on opposite sides for the "spirit of the market" but @a_l_e_x I do think your criteria as written have very explicitly been met

@ShaneBo In this situation, AFP is not a primary source. They're just doing basic math on ISW data. Anyone following these trackers knows that every one of them invests a lot of effort into independent data collection, and when the data has been collected, its analysis becomes just basic middle-school geometry calculations. So, even if we ignore the change in methodology and whether it's a better reflection of the conditions on the ground, a YES based on AFP and ISW would be a resolution on pure technicality, just because the author incorrectly assumed that all listed sources would be independent.

I think your misunderstanding of the nature of this work most prominently shows when you apply the OSINT label to it. While a large portion of the sources are classical OSINT like the videos by Madyar, Rubikon, or other Telegram channels of AFU units or Z-influencers, there are many steps that need to be made before there's any data that's good enough to be analyzed, and sourcing for those steps comes from the places that are the opposite of open - and are definitely not satellite maps. I can't say much about the Russians, they're too icky for me to follow closely, but the same DeepState has done plenty of deep dives on their processes, and those talks make it obvious that these guys are not OSINT, at least, not anymore. I'm telling you this as someone who worked with one of the most prominent Ukrainian OSINT groups in the first years of the war in Ukraine. DeepState's or ISW's capabilities far exceed those available to the nerds who spent their days scrolling through Odnoklassniki in search of careless selfies. These are the teams doing the hard work, and AFP appears to have just glommed onto it.

@ShaneBo okay, I think your comments are partly AI-generated but there are some points worth discussing and I've written my thoughts below. I'm still not convinced either way, so if this continues to be disputed I might try to assemble a panel of trusted Manifolders (probably mods) to help me decide both issues

On issue 1:

The inclusion of infiltrations in previous reports was a limitation of the data available at the time. By providing a breakout in the May 2 report, ISW is effectively 'cleaning' their data.

No, ISW has clearly been able to include or exclude infiltrations since November at the latest and has chosen to include them as gains/losses. Is there any reason to think they've rejected that practice as inaccurate?

To resolve on the 'Infiltration' figure is to treat a temporary scouting party as equal to an established defensive trench line.

This is a stretch. It is entirely possible to control an area without trenches, fortifications, etc. But I think you have a fair point here.

AFP's recognition of this suggest an additional credentialed acknoledgement that this is the most accurate data concerning actual territiorial control.

In my view, AFP's use of the exclusionary figure suggests that it is more newsworthy, not necessarily more accurate for our purposes.

Other sources not included in this collection are agreeing, and others are holding off their usual announcements as they consider this advancement.

Who?

Infiltrations represent presence in a gray zone, not control. The fact that ISW felt the need to break this out for the first time in May suggests they viewed the 'Control' metric as the more accurate representation of the month's reality.

Ordinarily I would be inclined to agree that infiltration ≠ control (though I should point out that the word 'control' does not appear anywhere in the resolution criteria). But it's hard to ignore the fact that the standard practice of ISW itself has been to count infiltrations as gains. I don't really see how ISW providing both figures now suggests that one is more accurate than the other.

However, to address the methodological "apples to oranges" concern specifically, we can look at the data continuity. The net loss is not a statistical artifact of the May 2nd methodology shift; it is a physical reality that was already being recorded in the "old" system.

This is correct, but you've contradicted what you said earlier about ISW not having the data available until now.

On issue 2:

ISW provides the raw spatial data. AFP performed its own independent analysis of that data to reach a net square-kilometer conclusion.

ISW has done 99% of the work here by gathering evidence, making subjective decisions about who controls what (i.e. analysis) and displaying it all on a map. The full extent of AFP's "analysis" would have been to look at that map at the start and end of the month to calculate the net area that changed hands. This is why they reached virtually the same figure.

I could make a map of my backyard and shade the area covered by leaves, then send the map to someone on the other side of the world. We could both "independently" calculate the area covered by leaves, based on my map, and arrive at the same figure. But the other person hasn't seen my backyard and hasn't made independent decisions about what counts as covered.

In OSINT, we rarely have independent primary data (most use the same satellites); independence is found in the analysis of that data.

Satellites? ISW seems to rely mostly on Telegram posts and geolocated photos/videos from social media. But I fear I'm going off track.

Here's how I see it... ISW has done an analysis of the data. AFP appears to have done an "analysis" of ISW's analysis. I can't consider that a truly independent analysis. I think your stronger argument is that the resolution criteria don't explicitly require one.

On issue 1:

The inclusion of infiltrations in previous reports was a limitation of the data available at the time. By providing a breakout in the May 2 report, ISW is effectively 'cleaning' their data.

No, ISW has clearly been able to include or exclude infiltrations since November at the latest and has chosen to include them as gains/losses. Is there any reason to think they've rejected that practice as inaccurate?

My response:

I feel I made clear. They made a decision that this is actually the most accurate way to address the reality of control. Which I defined as the spirit of real territory gains or losses.

To resolve on the 'Infiltration' figure is to treat a temporary scouting party as equal to an established defensive trench line.

This is a stretch. It is entirely possible to control an area without trenches, fortifications, etc. But I think you have a fair point here.

My response: Sure, but I think I supported well enough in other analysis as to why this makes sense. But more importantly it is what ISW is suggesting makes sense

AFP's recognition of this suggest an additional credentialed acknoledgement that this is the most accurate data concerning actual territiorial control.

In my view, AFP's use of the exclusionary figure suggests that it is more newsworthy, not necessarily more accurate for our purposes.

My Response: An independent source putting their reputation behind it and agreeing to the logic seems to meet the criteria of a credenitialed validatation, but can cede that. In the end I just don't think it matters given your resolution criteria. They are one of your listed sources

Other sources not included in this collection are agreeing, and others are holding off their usual announcements as they consider this advancement.

Who?

Infiltrations represent presence in a gray zone, not control. The fact that ISW felt the need to break this out for the first time in May suggests they viewed the 'Control' metric as the more accurate representation of the month's reality.

Ordinarily I would be inclined to agree that infiltration ≠ control (though I should point out that the word 'control' does not appear anywhere in the resolution criteria). But it's hard to ignore the fact that the standard practice of ISW itself has been to count infiltrations as gains. I don't really see how ISW providing both figures now suggests that one is more accurate than the other.

My Response: I list the logic of this in my responses, and ISW also lists their logic and reasoning within their own reports. They are doing so for transparency and explicitly noting that the new way is more accurate. Max makes a great point regarding this with his drones input.

However, to address the methodological "apples to oranges" concern specifically, we can look at the data continuity. The net loss is not a statistical artifact of the May 2nd methodology shift; it is a physical reality that was already being recorded in the "old" system.

This is correct, but you've contradicted what you said earlier about ISW not having the data available until now.

My response: tI am referncing losses showing up in the old system

On issue 2:

ISW provides the raw spatial data. AFP performed its own independent analysis of that data to reach a net square-kilometer conclusion.

ISW has done 99% of the work here by gathering evidence, making subjective decisions about who controls what (i.e. analysis) and displaying it all on a map. The full extent of AFP's "analysis" would have been to look at that map at the start and end of the month to calculate the net area that changed hands. This is why they reached virtually the same figure.

I could make a map of my backyard and shade the area covered by leaves, then send the map to someone on the other side of the world. We could both "independently" calculate the area covered by leaves, based on my map, and arrive at the same figure. But the other person hasn't seen my backyard and hasn't made independent decisions about what counts as covered.

In OSINT, we rarely have independent primary data (most use the same satellites); independence is found in the analysis of that data.

Satellites? ISW seems to rely mostly on Telegram posts and geolocated photos/videos from social media. But I fear I'm going off track.

My response: this is an analogy not an analysis of methodolgy. It is supporting the ability of secondary sources to have their own independence not just primary sources.

Here's how I see it... ISW has done an analysis of the data. AFP appears to have done an "analysis" of ISW's analysis. I can't consider that a truly independent analysis. I think your stronger argument is that the resolution criteria don't explicitly require one.

My response: This is the main point of my argument, it matches your resolution criteria, but also has I think a fairly strong "spirit" argument that the way ISW is calculating losses/gains is the most accurate

@MaxA I can make a longer reply tomorrow, but I want to be clear that I am not suggesting AFP is a primary source. My argument isn't that AFP is "doing the hard work" of primary collection; it’s that for the purposes of a resolution contract, they are a designated analytical validator. The OSINIT mention is purely a reference to independence not rigor. They are listed as source and can make a decision using shared data. For the purposes of the market maybe it would be much better to only have primary sources, but that is not what is written.

@ShaneBo ON other sources the most notable is DeepStatesUA definition of gray zones which is perhaps looser than ISW on what constitutes control(territory gain/loss) as the infiltration will not count until some conditions have been met. Blackbird seems to be weighing on the fuzziness of the LOC as well,. Neither of these are fully in agreement with the specific assesment of control at this point, but their analyis seems clearly to support that infiltrations should be questioned.

@ShaneBo You are finally seeing my point:

"For the purposes of the market maybe it would be much better to only have primary sources, but that is not what is written."

Exactly what I've been saying all this time: spirit ("better to only have primary sources") vs letter ("but that is not what is written")

As for the gray zones and how each team defines it - irrelevant, as long as those definitions remain consistent month to month.

@MaxA I honestly think we mostly agree. Core contradictions is that I feel it’s possible to prove the spirit of the market by supporting just ISWs analysis and that though it may be best with primaries it doesn’t sufficiently weaken the core argument , and that at the end of the day for manifold the letter of the market is more important

To clarify: also noting that secondary sources have value and independence

@ShaneBo I tried and perhaps failed to make this clear in each of my comments. To make a spirit argument but prioritize how this market is written

The last two months have come quite close to triggering a YES resolution...

Black Bird Group reported a net territorial gain for Ukraine in February and a small net loss in March. ISW reported net gains of 152km² and 5km² for Russia in February and March respectively.

ISW shows net loss for the month, based on https://understandingwar.org/research/russia-ukraine/russian-offensive-campaign-assessment-february-15-2026/ and https://understandingwar.org/research/russia-ukraine/russian-offensive-campaign-assessment-march-3-2026/

Black Bird Group reports that Ukraine had a net gain of 37 square kilometres in February. If just one of ISW, Reuters, AFP and AP also independently determines that Ukraine had a net territorial gain in February, this market will resolve YES.

None of the resolution sources have reported a net territorial gain for Ukraine in January.

Ukraine's net territorial loss was:

• 560 sq km according to Creamy caprice

• 526 sq km according to LostArmour

• 470 sq km according to Divgen

• 389 sq km according to Black Bird Group

• 265 sq km* according to ISW (1 Jan - 27 Jan)

11 months to go...