Gran Turismo Sport: General Discussion

[chzsln485]

The Xbox one should definitely be able to handle weather/night rendering if the PS3 could in my opinion, looks like MS forced turn 10 to release the game too fast. It's best for PD to keep silent and hear out the complaints from the people and work diligently on the product. I have faith that PD will deliver massively. As a kid growing up to a man now see the evolution of GT1- GTSPORT is astonishing.

I wouldn't say the PS3 could handle it that well. The frame drops were pretty dire at times, but overall, I appreciate the inclusion of dynamic weather more than not.

 

[PzR Slim]

I wasn't saying they hadn't bothered, but I'm not sure I appreciate the 60fps framerate over the weather in PCARS. I'm normally all for 60fps, but in this case, I'm not sure 60fps is worth keeping for a weather system I find dull, simply because of the fact you know a wet track will be identical every time you race on it, and that the overall conditions cannot change mid-race.

I agree. I love the dynamic weather and time of PCars. However you were using it as an example to back up synon's claim that collaboration could hamper feature implementation. I was pointing out that you were wide of the mark with your example

 

[chzsln485]

I agree. I love the dynamic weather and time of PCars. However you were using it as an example to back up synon's claim that collaboration could hamper feature implementation. I was pointing out that you were wide of the mark with your example

Ah, OK, I see what you were getting at. Yes, it's rather more a design decision than a development or staff limitation. Either way, I still believe it to be the wrong decision.

 

[PzR Slim]

The Xbox one should definitely be able to handle weather/night rendering if the PS3 could in my opinion, looks like MS forced turn 10 to release the game too fast. It's best for PD to keep silent and hear out the complaints from the people and work diligently on the product. I have faith that PD will deliver massively. As a kid growing up to a man now see the evolution of GT1- GTSPORT is astonishing.

It does, in PCars so not sure what your point is. As I've already explained T10 made a design decision to not include it so they could maintain a solid 60FPS. This is known and has been known for some time.

 

[FoRiZon]

Why am i waiting for this.

 

[Imari]

The main issue PD had was being understaffed I doubt Turn 10 would ever been seen as a threat again since PD got a wake up call. Economics being the biggest factor. But that's my opinion. I can't wait to taste the new racers on the PS4.

Hard to say that they're not a threat when they'll have four games out to Polyphony's one, and they're also starting to move onto PC.

They may not be a big threat to a company as entrenched as Polyphony, but Polyphony would be fools not to pay attention.

The Xbox one should definitely be able to handle weather/night rendering if the PS3 could in my opinion, looks like MS forced turn 10 to release the game too fast.

But the PS3 couldn't handle it to Turn 10 standards. The required standard of graphics went up with next gen as well, as we can see with pCARS. pCARS can't hold a solid 60fps with time and weather either.

The PS4 would be able to handle a PS3 quality game with weather at 60fps, but that wouldn't look good beside true current gen titles.

You can see the Forza system operating at 30fps in Horizon 2 if you want. It looks pretty good. But for whatever reason it doesn't work at 60fps without presumably sacrificing more peripheral graphical features than they're willing to let go of.

For example, it's taken until FM6 to have weather and time, and it isn't dynamic or anything clever, it's just day dry track, day wet track, night dry track, night wet track.

Well, one might argue that the whole depth of puddles and hydroplaning thing is quite clever, compared to something like simply modifying the grip of the track surface.

I agree that the FM6 system is very simplistic compared to some, but such is life. I still think the best weather system on console is the Codies F1 system, where each patch of track has a wetness coefficient and the water flows downhill, builds in low lying areas of track and is cleared by cars running over it, as well as localised weather so that certain parts of the track can be wet while others are dry. Neither GT, pCARS or FM have anything on that.

Likewise, I think only pCARS and F1 on console have gotten into live track evolution, where the track rubbers in. Arguably it's a more important effect than weather or time, because while a few races may take place in changing conditions all races take place on an evolving track. The only races where it doesn't matter are those that are very, very short.

But weather and time look cool, so they get in first.

 

[Nexus]

I believe that Assetto Corsa simulates rubber depositing on tracks too.

 

[chzsln485]

Well, one might argue that the whole depth of puddles and hydroplaning thing is quite clever, compared to something like simply modifying the grip of the track surface.

I agree that the FM6 system is very simplistic compared to some, but such is life. I still think the best weather system on console is the Codies F1 system, where each patch of track has a wetness coefficient and the water flows downhill, builds in low lying areas of track and is cleared by cars running over it, as well as localised weather so that certain parts of the track can be wet while others are dry. Neither GT, pCARS or FM have anything on that.

Likewise, I think only pCARS and F1 on console have gotten into live track evolution, where the track rubbers in. Arguably it's a more important effect than weather or time, because while a few races may take place in changing conditions all races take place on an evolving track. The only races where it doesn't matter are those that are very, very short.

But weather and time look cool, so they get in first.

I don't have F1 2015, but that does sound very impressive as a weather system. I do think the FM6 puddles are impressive too, but only in small doses, as it were. The first time I did a wet race in FM6 I was blown away by the weather, but once I realised the puddles are the same shape/size and in the same location every race (plus the fact it never gets wetter or dryer) means a lot of the initial interest was eroded. I think the thing that makes a weather system really work is the fact that it's dynamic and unpredictable - it's not just about the simulation of the wet track itself. That's why I like the sound of the F1 system.

 

[synon]

I'm not sure that any of what I was talking about means that features are slower to be introduced. They may take more iterations to appear, but that's what you get for being on a two year cycle instead of 3+. If anything, sharing developmental technologies means that features can be introduced faster. You've got two teams working instead of one, and transferring tech only means adapting a feature instead of building it from scratch.


As far as I can see, the difference between the speed of feature introduction between T10 and Polyphony is more due to developmental and design philosophy. Turn 10 have a certain minimum level of quality for features before they'll put them in the game. Polyphony's minimum level of quality is simply much lower, and so they're able to put features in their game earlier.

Weather and time change are a perfect example, they never would have made it into a T10 game in the state they are in GT5 and 6 because they compromise frame rate. Polyphony were OK with it. That's fine, it's a valid choice and a lot of people are legitimately unable to detect frame drops.

Before anyone accuses me of picking on Polyphony, SMS did exactly the same thing with pCARS. Turn 10 is the odd one out choosing to hold features back for quality reasons. We know that they have time and weather tech, we've seen it in Horizon.

It was the fitting features to deadlines part, rather than than the collaboration part. Fitting the schedule to the features, to some degree, means that more features that don't fit the schedule exactly can be accommodated, introducing them earlier. I also imagine there has to be a fair amount of allowance in order to meet a fixed schedule, to allow for any unscheduled work that may be required, limiting it a bit further.

 

[DRLOVE]

Driveclub to me probably has the best weather effects I've seen in a video game but I found it weird that PD has a couple studios in the EU territories near evo studio.

 

[Johnnypenso]

I believe that Assetto Corsa simulates rubber depositing on tracks too.

Dynamic tracks are a part of the pc version of AC, hopefully the console version too. They simulate not only changing track conditions but differing starting conditions as well. Assuming this makes it to the console version as well, you could start a race weekend on a Friday and have a green track with say 92% grip, as it might have after a rain and without a recent race. At the end of the weekend you could be at 99% grip with lots of rubber laid down in the racing groove. You also have control over how quickly the rubber builds up during a race. You might have a 40 lapper going and start at 95% grip and build to 97% by race end...or 99%...or whatever you choose. You also have control over ambient temperatures which also affects grip levels. Driving the Nurburgring at a realistic 15C and 92% grip level on car approriate rubber is quite an enlightening and engrossing and humbling experience.

As a package, it really adds to the immersion and it is also another reason why a giant car count isn't really missed in a game like AC and won't be missed in a game like GTSport, so long as they make a heroic attempt to simulate real racing weekends and series. You can get so engrossed in one single event that it can take up multiple nights of racing. Completing a true racing series could take many nights or multiple weeks depending on your playing time. There simply isn't enough time to drive hundreds of cars unless all you do is short, hotlap sessions.

 

[ImaRobot]

Yes, Forza 6 appears to have sold between 1 to 2 million, I doubt that is enough to break even.

What is enough to break even?

For example, it's taken until FM6 to have weather and time, and it isn't dynamic or anything clever, it's just day dry track, day wet track, night dry track, night wet track. PCARS has a massively better implementation of weather, whereas FM6, despite having endurance races of sorts, doesn't actually have the time and weather change over the course of the race, which is one of the best thing about doing full length endurances in PCARS.

They choose to not take the route that everyone else took. The route everyone else took was to implement a feature that greatly hindered it's own performance when using it in any aspect, just to be able to say "Hey look, we have this feature!" it runs along the same line as GT's thousand-vehicle stable.

When I'm playing a game that takes as much precision as these games do, I prefer to be able to have my inputs register as fluidly and instantly as possible, in every area of the game. Would I like a full weather cycle? You bet. However, I'd prefer that the game functions at 100% first and foremost.

It's a calling card that a lot of people flock too, but what most don't realize, or accept, is that these features introduced in other games have presented many problems. It makes it not so much a problem for one game to not have a certain feature, if other games have it, but can't get it optimized correctly. It's function over form, that I prefer.

 

[Imari]

I don't have F1 2015, but that does sound very impressive as a weather system. I do think the FM6 puddles are impressive too, but only in small doses, as it were. The first time I did a wet race in FM6 I was blown away by the weather, but once I realised the puddles are the same shape/size and in the same location every race (plus the fact it never gets wetter or dryer) means a lot of the initial interest was eroded. I think the thing that makes a weather system really work is the fact that it's dynamic and unpredictable - it's not just about the simulation of the wet track itself. That's why I like the sound of the F1 system.

Real weather is dynamic but not entirely unpredictable. A track is always going to form rivers and puddles in the same places, because that's where the natural dips and hollows of the land are. You can see the weather coming in, and in real life you'd have someone on the radio helping you out, even if it's only a friend standing on top of the pit buildings watching the clouds.

It's a bit like learning where the bumps are on a track, once you know where they are then that's solid knowledge regardless of whether you're driving a car that's strongly affected or not. With weather, there's going to be some variation based on exactly how much water there is, but you're going to know which parts of the track to avoid and which to take care on. Because really, if it was fully unpredictable it would be impossible. You can't do anything if you start hydroplaning mid-corner when you're at ten tenths. You'd be into a tree and dead.

I think what makes dynamic weather interesting is that it requires adaptability and additional skills from racing in the dry.

Firstly, one needs to learn the new lines and cornering speeds suitable for wet conditions. This means learning where the puddles and rivers are and how to avoid them at maximum speed. Some areas that might have previously been passing opportunities may now not be, because of standing water. Arguably it also becomes much more important to drive smoothly.

Secondly, one has to be able to judge the extent of the rain. Being about to do it visually is important, at least as an approximation. This is a major failing of the GT5/6 system, and why they had to implement the rain gauge. One should also be able to feel how the car is behaving, and refine the estimate of the available grip and level of water accordingly.

Thirdly, one needs to be able to adapt to the ongoing changes. If the rain gets heavier on your windscreen, that means in a lap or so there's going to be a lot more water on the track. Push hard now while you've still got grip, then take extreme care.
It's another variable to pay attention to, and what separates the men from the boys is those that can see the signs and adapt appropriately before the conditions catch them out.

This is why I don't really have much of a problem with FM6's weather. It's extremely limited, but seems like a decent simulation of driving in a steady downpour. That's only one of a million different possible situations, but it's something. It fits the T10 philosophy that I was talking about before, a limited feature that is of extremely high quality within it's scope.

GT6's weather isn't great either, because while it's "dynamic" most of the skills that you would use in the real world to be fast don't apply. There's no puddles or rivers. You can't see the water on the track, you have to look at your water gauge. The way the tyres respond to water on the track is iffy at best, slicks are mostly fine instead of a death sentence. That also is consistent with the Polyphony philosophy.

To me, it seems like they're both not great and it's swings and roundabouts. One gives you a fairly excellent simulation of one specific situation, the other gives you a fairly average simulation of a range of situations. But that's just my take on it.

 

[DRLOVE]

Real weather is dynamic but not entirely unpredictable. A track is always going to form rivers and puddles in the same places, because that's where the natural dips and hollows of the land are. You can see the weather coming in, and in real life you'd have someone on the radio helping you out, even if it's only a friend standing on top of the pit buildings watching the clouds.

It's a bit like learning where the bumps are on a track, once you know where they are then that's solid knowledge regardless of whether you're driving a car that's strongly affected or not. With weather, there's going to be some variation based on exactly how much water there is, but you're going to know which parts of the track to avoid and which to take care on. Because really, if it was fully unpredictable it would be impossible. You can't do anything if you start hydroplaning mid-corner when you're at ten tenths. You'd be into a tree and dead.

I think what makes dynamic weather interesting is that it requires adaptability and additional skills from racing in the dry.

Firstly, one needs to learn the new lines and cornering speeds suitable for wet conditions. This means learning where the puddles and rivers are and how to avoid them at maximum speed. Some areas that might have previously been passing opportunities may now not be, because of standing water. Arguably it also becomes much more important to drive smoothly.

Secondly, one has to be able to judge the extent of the rain. Being about to do it visually is important, at least as an approximation. This is a major failing of the GT5/6 system, and why they had to implement the rain gauge. One should also be able to feel how the car is behaving, and refine the estimate of the available grip and level of water accordingly.

Thirdly, one needs to be able to adapt to the ongoing changes. If the rain gets heavier on your windscreen, that means in a lap or so there's going to be a lot more water on the track. Push hard now while you've still got grip, then take extreme care.
It's another variable to pay attention to, and what separates the men from the boys is those that can see the signs and adapt appropriately before the conditions catch them out.

This is why I don't really have much of a problem with FM6's weather. It's extremely limited, but seems like a decent simulation of driving in a steady downpour. That's only one of a million different possible situations, but it's something. It fits the T10 philosophy that I was talking about before, a limited feature that is of extremely high quality within it's scope.

GT6's weather isn't great either, because while it's "dynamic" most of the skills that you would use in the real world to be fast don't apply. There's no puddles or rivers. You can't see the water on the track, you have to look at your water gauge. The way the tyres respond to water on the track is iffy at best, slicks are mostly fine instead of a death sentence. That also is consistent with the Polyphony philosophy.

To me, it seems like they're both not great and it's swings and roundabouts. One gives you a fairly excellent simulation of one specific situation, the other gives you a fairly average simulation of a range of situations. But that's just my take on it.

GT6 is on hardware pushing 10years old and comparing it to current gen games is like a low blow if you ask me. New phase, new standards for current gen games if you ask moi. My problem with GT is lack of premiums/damage/physics and of courses how could we all forget the 90's funky jazz/techno menu music. GT needs to re discover it's roots as to what made it a special game.

 

[ImaRobot]

GT6 is on hardware pushing 10years old and comparing it to current gen games is like a low blow if you ask me. New phase, new standards for current gen games if you ask moi. My problem with GT is lack of premiums/damage/physics and of courses how could we all forget the 90's funky jazz/techno menu music. GT needs to re discover it's roots as to what made it a special game.

Producing desirable weather effects at 60FPS seems to be an issue even on this generation as well. Pcars has it, but like GTs, it produced some odd problems. Driveclub, however, has a beautiful system implemented, but that's at 30 FPS. What route are they going to take? We'll have to wait and see, but they like to push boundaries beyond what is actually possible with their hardware, so I don't really have high hope at this point. It's all a waiting game for now, though.

 

[Matteo05]

Producing desirable weather effects at 60FPS seems to be an issue even on this generation as well. Pcars has it, but like GTs, it produced some odd problems. Driveclub, however, has a beautiful system implemented, but that's at 30 FPS. What route are they going to take? We'll have to wait and see, but they like to push boundaries beyond what is actually possible with their hardware, so I don't really have high hope at this point. It's all a waiting game for now, though.

I think this was the main problem of the last two games. They pushed a little too much.

They should make sure that everything INSIDE the boundaries works before pushing.

 

[Tornado]

They should, and hopefully they will, but at this point there is no indication that they actually learned that lesson.

 

[chzsln485]

They should, and hopefully they will, but at this point there is no indication that they actually learned that lesson.

It's not impossible, as every GT prior to GT5 seemed to operate well within the boundaries of their consoles (GT3 and 4 in particular ran extremely well considering the visuals they put out at the time on PS2).

 

[synon]

What is enough to break even?

I think the concensus for GT6 was it would break even at around 2 million sales. T10 being a larger studio, Forza 6 probably requires a bit more over a two year cycle.

They should, and hopefully they will, but at this point there is no indication that they actually learned that lesson.

The previous generation was about a dynamic environment, this generation it is possibly VR, so the emphasis might have changed to reflect this.

 

[ImaRobot]

I think the concensus for GT6 was it would break even at around 2 million sales. T10 being a larger studio, Forza 6 probably requires about the same over a two year cycle.

So, what is enough to break even?

The previous generation was about a dynamic environment, this generation it is possibly VR, so the emphasis might have changed to reflect this.

This is not any indication that they realized that they have pushed their boundaries in the past. It actually seems like it might be the opposite, really. That's not to say that I have a problem with VR. Let's just wait to see how it all plays out.

 

[Tornado]

So you doubt an unknown amount of sales for a game was enough to break even based on baseless assumptions concensus for how much another game from a different studio with a different development period required to break even?

 

[synon]

So you doubt an unknown amount of sales for a game was enough to break even

No, I doubted the range I stated.

based on baseless assumptions concensus for how much another game from a different studio with a different development period required to break even?

Well, any comparison, albeit fairly facile, can be made, comparing studio size pro rata with the development period. Of course, it's not a definitive statement

 

[dipsetab]

it seems like the thirst level is pretty high for GTS info patience guys

 

[Imari]

GT6 is on hardware pushing 10years old and comparing it to current gen games is like a low blow if you ask me.

I'm comparing design, and I find both to lack in different areas. If you read back you'll note that I said that in my opinion the best weather system on console was Codies F1, and that was also on PS360.

New phase, new standards for current gen games if you ask moi.

Which brings us back to the point that started this little round, until Gran Turismo actually has a game on current gen that's going to be difficult.

I think the concensus for GT6 was it would break even at around 2 million sales. T10 being a larger studio, Forza 6 probably requires a bit more over a two year cycle.

I don't know what they are now, but in 2011 they were around 70 people according to MS. That's a bit after GT5, at which point Polyphony was somewhere around the 100-110 mark if I recall.

http://blogs.microsoft.com/next/2011/09/16/forza-motorsport-pt1-turn-10-studios/

The big quoted numbers that we're used to hearing (3-400) come from adding in contractors, which Polyphony doesn't do. These contractors might be there for the whole development cycle, or they might be pulled in to do one job and then bail. Realistically, T10 is probably around the same size as Polyphony in terms of people who are 100% full time T10 employees. It's just that they have no reservations about outsourcing and they quote the max team size they can justify, presumably because it's good marketing. 400 people working for two years sounds like it would make a much better game than 100 people and 300 Korean high school students.

It's much harder to provide a reasonable estimate of salary costs when contractors are such a big part of the team. With Polyphony one can make an educated guess at an average salary and multiply by the rough amount of employees. Add in some costs for running and maintenance, travel and equipment and you can have a decent stab. We also had rough ideas of the budget for GT5 from Kaz, which helped narrow things down.

With T10, I wouldn't even know where to begin guessing how much those contractors might cost. The range would be so huge that the estimate would be almost meaningless. The break even point could be anywhere from ~750,000 copies to several million.

For reference, pCARS cost $5 million to make. SMS was also around 80 people and took a little over three years.

http://www.redbull.com/au/en/games/stories/1331625870374/project-cars-the-ultimate-racing-game

 

[pikes_peak]

The lack of news is mind-boggling ... hell, even @Jordan is quiet. Last time he posted here, was around New Year's Eve. The facebook guy doesn't want to share some news as well
What's going on @ PD's headquarter? Course maker 2.0?

So far we had:
- announcement @ PGW
- some crushed hopes @ PSX
- some Xmas pictures
- a glimmer of hope for the new year ('roll on 2016')
- no news @ Tokyo Auto Salon and probably Geneva too
- Beta still mentioned for 'early 2016'
- ...

 

[synon]

I'm comparing design, and I find both to lack in different areas. If you read back you'll note that I said that in my opinion the best weather system on console was Codies F1, and that was also on PS360.



Which brings us back to the point that started this little round, until Gran Turismo actually has a game on current gen that's going to be difficult.



I don't know what they are now, but in 2011 they were around 70 people according to MS. That's a bit after GT5, at which point Polyphony was somewhere around the 100-110 mark if I recall.

http://blogs.microsoft.com/next/2011/09/16/forza-motorsport-pt1-turn-10-studios/

The big quoted numbers that we're used to hearing (3-400) come from adding in contractors, which Polyphony doesn't do. These contractors might be there for the whole development cycle, or they might be pulled in to do one job and then bail. Realistically, T10 is probably around the same size as Polyphony in terms of people who are 100% full time T10 employees. It's just that they have no reservations about outsourcing and they quote the max team size they can justify, presumably because it's good marketing. 400 people working for two years sounds like it would make a much better game than 100 people and 300 Korean high school students.

It's much harder to provide a reasonable estimate of salary costs when contractors are such a big part of the team. With Polyphony one can make an educated guess at an average salary and multiply by the rough amount of employees. Add in some costs for running and maintenance, travel and equipment and you can have a decent stab. We also had rough ideas of the budget for GT5 from Kaz, which helped narrow things down.

With T10, I wouldn't even know where to begin guessing how much those contractors might cost. The range would be so huge that the estimate would be almost meaningless. The break even point could be anywhere from ~750,000 copies to several million.

For reference, pCARS cost $5 million to make. SMS was also around 80 people and took a little over three years.

http://www.redbull.com/au/en/games/stories/1331625870374/project-cars-the-ultimate-racing-game

The figure I recall reading was T10 had 70 full-time staff and the rest contractors making up a work-force of 250.
http://www.seattletimes.com/busines...softs-turn-10-studio-drives-innovative-games/
Plus the number of people working on Motorsport/Horizon being over 500 in total, though that might, as you say, include outsourcing numbers.

The figure for the Pcars budget seems a little low, the figure I saw for the cost of the 40 staff still working on Pcars1 was 400k dollars per month.

 

[dipsetab]

The lack of news is mind-boggling ... hell, even jordan is quiet. Last time he posted here, was around New Year's Eve. The facebook guy doesn't want to share some news as well
What's going on @ PD's headquater? Course maker 2.0?

So far we had:
- announcement @ PGW
- some crushed hopes @ PSX
- some Xmas pictures
- a glimmer of hope for the new year ('roll on 2016')
- no news @ Tokyo Auto Salon and probably Geneva too
- Beta still mentioned for 'early 2016'
- ...

agreed XDD

 

[jm79]

The lack of news is mind-boggling ...
What's going on @ PD's headquater?

I know, been way too long now haha.

What's next, GDC?

And we (PS4 owners at least) don't even have AC to fill the gap either.

 

[Imari]

The figure I recall reading was T10 had 70 full-time staff and the rest contractors making up a work-force of 250.
http://www.seattletimes.com/busines...softs-turn-10-studio-drives-innovative-games/

So it looks like they've stuck with their 70 core members and just use contractors as necessary. Interesting that they haven't expanded the core team since 2011, but I guess there's no need when you can just contract someone to do it. Honestly, I think it's a really good way for a developer to work. There are some jobs that you just don't need someone working on for the whole dev cycle. And it makes them really flexible, and it's easy for them to bring in outside expertise if needed.

Plus the number of people working on Motorsport/Horizon being over 500 in total, though that might, as you say, include outsourcing numbers.

Almost certainly, I'd say. If you assume that Playground probably has a similar structure to T10, then it's about right.

The figure for the Pcars budget seems a little low, the figure I saw for the cost of the 40 staff still working on Pcars1 was 400k dollars per month.

Well, unless they pulled money out of their butts then that's all they had from what SMS announced that they personally put into the project and what they received from the "kickstarter" funds.

40 guys on $120k a year average each seems very, very high. Those are management salaries. Where did you hear that? I mean, at $400k a month over three years that would be more than $14 million. That's a massive discrepancy from how much the sources I can find say the game cost.

The $5 million (quoted as 3.5 million euro) is from here: http://www.eurogamer.net/articles/2013-04-18-driving-ambition-pcars-crowdfunding-and-the-fsa

You can also see it at the top of the WMD forum page as 3.75 million euro: http://forum.wmdportal.com/

Wiki said $5 million and linked to the Eurogamer article, but I'm not entirely sure how they got to $5 million (presumably USD) from 3.5-3.75 million euro. It's more like $4 million on current exchange rates.

As with all financial figures there's some wiggle room, and I'd probably believe it if someone told me it ended up being $6 or 7 million. But $14+ million? That's a lot.

 

[jm79]

As with all financial figures there's some wiggle room, and I'd probably believe it if someone told me it ended up being $6 or 7 million. But $14+ million? That's a lot.

Perhaps the rest was publisher money, as a top-up.

Didn't the same thing happen with Shenmue 3 recently?