Shiny claw machine mechanics

[Staybeasting]

After 3 hours I finally got shiny wasabi kitty. Me and My friend were able to move her in between the purple and green plushie and on the 8th try I grabbed her. Thankfully I’m done with that. Last thing I need is the penetrator award. My highest so far is 32680

[jargonaut]

39 minutes ago, AshyraVash said:

Congratulations! I've been testing several different timings, too, but none of them did the trick. The ones that reportedly worked for others were: w: 4300, d: 3900 and w: 4200, d: 3980, if that helps any.

Your timings may be a bit off; I'll PM you mine in a moment. You on PC (and or use AHK)?

Edited February 3, 2020 by jargonaut

[AshyraVash]

32 minutes ago, jargonaut said:

Your timings may be a bit off; I'll PM you mine in a moment. You on PC (and or use AHK)?

Ah, PC and AHK, yes. The first timing was right in the reddit post and the second one from MacroGamer. I tried some other ones as well, but 4300 seemed like the limit of any direction. I've been letting it run whenever afk, just to experiment, now if LJT would just stop calling me, maybe my character wouldn't be found playing space monkey a few hours later. xD

[jargonaut]

I'm still not 100% convinced it's purely about the physical location of Wasabi (when in non-glitched position) or the position of the claw related to the plushie, but simply a matter of getting both directional timings right. Something you could in theory do with your eyes not on screen, but a stop watch. Hell, I would even have accepted a longer timing than what is physically possible visually (read: holding a directional button for a bit longer even after the claw hits a wall) cause a thing like that could have just been another way R* makes us guessing and testing.

 

But do this: Edit your script and make it try the nine positions below and loop it. This will extended the testing area a bit to all directions (cause I still believe it's actually 35 milliseconds and not millimeters Fun found in the code). Then make sure you have set notifications to 60 mins (if you haven't yet done it for whatever reason). Then restart your game to make sure you have clean cache, start the game in single player mode and jump to invite only session to make sure no-one has been there to make it dirty yet.

 

{w: 4200, d: 3870}
{w: 4200, d: 3865} <- my hit on 200th try of this bunch
{w: 4200, d: 3875}
{w: 4205, d: 3870}
{w: 4195, d: 3870}
{w: 4205, d: 3865}
{w: 4205, d: 3875}
{w: 4195, d: 3865}
{w: 4195, d: 3875}

 

As I said above, me hitting it exactly on 200th attempt and with no extended data, it would still be impossible to say the exact timing by millisecond, but visually on my screen, the first one fits and sits perfectly while all others including the one that picked it up fit inside the 35ms tolerance area. So it may have been that all these positions are inside the right area and for whatever comical überuniversal reason I needed 200 x 1/200 = 1. 

 

But change your script to loop those nine. My script used 30 secs / attempt, and it took 200 attempts or 100 mins to get it, quite precisely.

 

Edit: I'm going for my second nap now, but before that: Even the above set of timings doesn't look to be hitting exactly the center on your screen, humor me and let it loop that bunch.

Edited February 3, 2020 by jargonaut

[AshyraVash]

49 minutes ago, jargonaut said:

I'm still not 100% convinced it's purely about the physical location of Wasabi (when in non-glitched position) or the position of the claw related to the plushie, but simply a matter of getting both directional timings right. Something you could in theory do with your eyes not on screen, but a stop watch. Hell, I would even have accepted a longer timing than what is physically possible visually (read: holding a directional button for a bit longer even after the claw hits a wall) cause a thing like that could have just been another way R* makes us guessing and testing.

 

But do this: Edit your script and make it try the nine positions below and loop it. This will extended the testing area a bit to all directions (cause I still believe it's actually 35 milliseconds and not millimeters Fun found in the code). Then make sure you have set notifications to 60 mins (if you haven't yet done it for whatever reason). Then restart your game to make sure you have clean cache, start the game in single player mode and jump to invite only session to make sure no-one has been there to make it dirty yet.

 

{w: 4200, d: 3870}
{w: 4200, d: 3865} <- my hit on 200th try of this bunch
{w: 4200, d: 3875}
{w: 4205, d: 3870}
{w: 4195, d: 3870}
{w: 4205, d: 3865}
{w: 4205, d: 3875}
{w: 4195, d: 3865}
{w: 4195, d: 3875}

 

As I said above, me hitting it exactly on 200th attempt and with no extended data, it would still be impossible to say the exact timing by millisecond, but visually on my screen, the first one fits and sits perfectly while all others including the one that picked it up fit inside the 35ms tolerance area. So it may have been that all these positions are inside the right area and for whatever comical überuniversal reason I needed 200 x 1/200 = 1. 

 

But change your script to loop those nine. My script used 30 secs / attempt, and it took 200 attempts or 100 mins to get it, quite precisely.

 

Edit: I'm going for my second nap now, but before that: Even the above set of timings doesn't look to be hitting exactly the center on your screen, humor me and let it loop that bunch.

Well, when it comes to the W timing, that can go on for quite a bit longer, but the D one just stops responding and offers the grab instead when it reaches the end. I've only started using AHK on this claw machine, specifically because I didn't trust myself to accurately grab at the same location each time with the keyboard and I tend to let nightclub and bunker run while afk anyway, so it seemed more productive than just watching security or TV screens. xD I haven't really looked into how to set notifications up, but I'll see if I can google it and I'll give it a few spins, thanks a lot for the findings! I also wondered about the 35mm and how this could even be measured, it must be milliseconds, surely.

[InfluxOG]

I made an easy to follow tutorial on how to push the Wasabi plushie out of place to make it easier to get:

 

 

Figured it might be useful to people who still have questions, tried to describe it best I could based on the info we gathered here, if I missed anything let me know.

 

[LeeC22]

1 hour ago, AshyraVash said:

I also wondered about the 35mm and how this could even be measured, it must be milliseconds, surely.

It isn't, it's millimetres. All GTA V measurements are based on real world metres, centimetres and millimetres... they have to be for the physics to work properly.

 

Shiny Wasabi Kitty is approximately 32cm tall, 23cm from front to back and 16cm wide. It sits in a bounding box .22971400 x .23845000 x .32432400 (that's X, Y & Z in metres), or 22.9cm x 23.8cm x 32.4cm. The position checking will probably be done by checking X & Y of the claw centre, to the X & Y of the plushie centre. Where this causes things to get tricky, is that the SWK plushie is longer from centre to back, so when it's aligned with the back of the cabinet (and the back of PRB), the centre is more than the claw's width away from the back. That means you have to stop the claw before it hits the back by about 28mm, to hit the centre of the checking area.

 

On the subject of time, you never check position/collision based on time, it's always distance. The reason for that is if you are working with small time segments, like 35ms, then anything less than 30fps (which gives a 33ms average frame time), can cause a time difference greater than you would be checking for, if you get a performance spike. You might get a 38ms frame update time and if you were 1ms away from that 35ms window on the previous frame, you're now 2ms past it on the next... it's too subject to fluctuations in performance for it to be reliable for checking. The only time you use time, is for things like: is a button held for more than 2 seconds? Because it doesn't matter how much you are over by.

 

Position checks are always position/distance based and what that 35mm probably means, is that the centre of the claw, can be within 17.5mm in any direction, centre to centre (-17.5 to +17.5). That's 20% of the SWK width, so it's quite wide.

 

That offset based on the distance from the back of the cabinet is probably what is throwing everyone off and that's the part that's going to be harder to hit.

 

Edit: Just a reference point, the distance from centre to back of PRB, is pretty much the same as the width of the claw... that's why you can get it easily by pushing the claw all the way to the back, you just have to line up the horizontal centre.

 

 

Edited February 3, 2020 by LeeC22

[AshyraVash]

9 minutes ago, LeeC22 said:

It isn't, it's millimetres. All GTA V measurements are based on real world metres, centimetres and millimetres... they have to be for the physics to work properly.

 

Shiny Wasabi Kitty is approximately 32cm tall, 23cm from front to back and 16cm wide. It sits in a bounding box .22971400 x .23845000 x .32432400 (that's X, Y & Z in metres), or 22.9cm x 23.8cm x 32.4cm. The position checking will probably be done by checking X & Y of the claw centre, to the X & Y of the plushie centre. Where this causes things to get tricky, is that the SWK plushie is longer from centre to back, so when it's aligned with the back of the cabinet (and the back of PRB), the centre is more than the claw's width away from the back. That means you have to stop the claw before it hits the back by about 28mm, to hit the centre of the checking area.

 

On the subject of time, you never check position/collision based on time, it's always distance. The reason for that is if you are working with small time segments, like 35ms, then anything less than 30fps (which gives a 33ms average frame time), can cause a time difference greater than you would be checking for, if you get a performance spike. You might get a 38ms frame update time and if you were 1ms away from that 35ms window on the previous frame, you're now 2ms past it on the next... it's too subject to fluctuations in performance for it to be reliable for checking. The only time you use time, is for things like: is a button held for more than 2 seconds? Because it doesn't matter how much you are over by.

 

Position checks are always position/distance based and what that 35mm probably means, is that the centre of the claw, can be within 17.5mm in any direction, centre to centre (-17.5 to +17.5). That's 20% of the SWK width, so it's quite wide.

 

That offset based on the distance from the back of the cabinet is probably what is throwing everyone off and that's the part that's going to be harder to hit.

 

 

Oh, that makes sense, thank you for clearing that up for us with such a detailed explanation!

 

So... if I understand it correctly, our duration on key-presses in scripts may not be accurate, either, because that determines how far the claw will move and due to differing FPS our claws may end up in different places? Fascinating! But how do you measure this, to find out if the claw even reached the optimal position? Although, 20% is a wide margin for error, generally.

[capton_cutclass]

Can someone help me do that Wasabi bug on PS4

[LeeC22]

4 hours ago, AshyraVash said:

Oh, that makes sense, thank you for clearing that up for us with such a detailed explanation!

 

So... if I understand it correctly, our duration on key-presses in scripts may not be accurate, either, because that determines how far the claw will move and due to differing FPS our claws may end up in different places? Fascinating! But how do you measure this, to find out if the claw even reached the optimal position? Although, 20% is a wide margin for error, generally.

The problem is, you can set a value in the key-press script that results in a waiting time that doesn't sync with the game update interval. Imagine an ideal world where 60fps gave you a constant 16ms, unless the time you entered isn't a multiple of 16, then it's either going to stop early or late. If you drop to 30fps, then it has to be a multiple of 32. So there's a risk you can spend ages tweaking values by 10ms, 5ms or even less and they won't be making any difference to what is happening in game, unless it shifts you from one frame update into the next. Real world time != in-game time, that's the important thing to remember, not until we get computers that are running at 1000fps anyway.

 

Edit: The only variable in that, is when you start the script relative to the game update... just to complicate things even further.

 

Measuring the distance is easy, thanks to the existing natives for checking distances. You simply check the X/Y distance from centre of claw to X/Y centre of plushie and if the distance is greater than 17.5, then you've missed. If it's within that, then you snap the plushie centre to the claw centre and pick it up. That's probably why it jumps when you get a successful hit.

 

Edit: Some more info... the internal dimensions of the cabinet seem to be 85cm. If you subtract the claw width from that and then time how long it takes the claw to travel from front to back, you can probably work out a ms to mm relationship, which might get you a bit closer to success. It could do with two people in the same session, with one taking a screengrab of the claw at the front and at the back while someone else is playing the machine, just to see what the relative positions are. I play solo, so I can't do that. That can then be used to gauge the model's positions in 3DS Max to get an accurate travel distance.

 

Claw bounding box is -.08583900 to .08583900 = .171m (17.1 CM)

 

Claw Max Y = .09911900

PRB Max Y = .09619200

SWK Max Y = .11922500

 

So you can see how close the claw and PRB distances are. Max Y = centre to back distance.

Edited February 4, 2020 by LeeC22

[AshyraVash]

13 minutes ago, LeeC22 said:

The problem is, you can set a value in the key-press script that results in a waiting time that doesn't sync with the game update interval. Imagine an ideal world where 60fps gave you a constant 16ms, unless the time you entered isn't a multiple of 16, then it's either going to stop early or late. If you drop to 30fps, then it has to be a multiple of 32. So there's a risk you can spend ages tweaking values by 10ms, 5ms or even less and they won't be making any difference to what is happening in game, unless it shifts you from one frame update into the next. Real world time != in-game time, that's the important thing to remember, not until we get computers that are running at 1000fps anyway.

 

Edit: The only variable in that, is when you start the script relative to the game update... just to complicate things even further.

 

Measuring the distance is easy, thanks to the existing natives for checking distances. You simply check the X/Y distance from centre of claw to X/Y centre of plushie and if Math.ABS(distance) is greater than 17.5, then you've missed. If it's within that, then you snap the plushie centre to the claw centre and pick it up. That's probably why it jumps when you get a successful hit.

That's a whole lot of variables to contend with then, in addition to the game update being tied to latency, too, I bet! At least now it makes sense why none of the timings seemed to fit with frametimes that might be anywhere between 6,9ms and 16ms.

 

One thing I'm not sure about anymore, is if the jumping even indicates a successful hit, since there was at least one person who showed a video of a different plushie jumping while another got grabbed and videos of grabs that must have been further than 20% off center. It's a very weird machine!

[LeeC22]

1 minute ago, AshyraVash said:

That's a whole lot of variables to contend with then, in addition to the game update being tied to latency, too, I bet! At least now it makes sense why none of the timings seemed to fit with frametimes that might be anywhere between 6,9ms and 16ms.

 

One thing I'm not sure about anymore, is if the jumping even indicates a successful hit, since there was at least one person who showed a video of a different plushie jumping while another got grabbed and videos of grabs that must have been further than 20% off center. It's a very weird machine!

Yeah, the problem is... Rockstar Logic. What I have explained is based on how I would do it and the variables I would take into account. Based on what I see in this game, they probably check how much rain has fallen since you started moving the claw and calculate a successful hit based on that. I've kinda run out of hard things to bang my head against with this game, it's way beyond common sense.

[LeeC22]

Just confirmation on this 35 value, I think I have found the code in the script that does this checking and the line says:

if (fVar8 < 0.035f)

That's a float, and its value is set as a result of GET_DISTANCE_BETWEEN_COORDS, so it's definitely a distance check. There's some GET_RANDOM_INT_IN_RANGE stuff going on as well, so I need to do some function following to see what's going on.

 

Before I go any further though... Am I getting into perilous waters with digging this far into online functionality? Does anyone finding some reliable values get classed as cheating online?

 

Edit: Ah.. okay... the random values are stored as Globals, that makes sense as it makes them impossible to get at from my perspective.

 

Edit: Just realised while I was lay in bed, that .035f check means the actual check radius is even bigger than I thought. That's a 7cm diameter checking circle around the centre of the plushie, which in the case of SWK, makes it about 40% of the width of the plushie... that's pretty large. So it looks like the RNG is the controlling factor and unless I have read the code wrong, there is only a single value in the range that counts as a valid hit and that's zero.

Edited February 4, 2020 by LeeC22

[jargonaut]

  

1 hour ago, LeeC22 said:

Just confirmation on this 35 value, I think I have found the code in the script that does this checking and the line says:

if (fVar8 < 0.035f)

That's a float, and its value is set as a result of GET_DISTANCE_BETWEEN_COORDS, so it's definitely a distance check. There's some GET_RANDOM_INT_IN_RANGE stuff going on as well, so I need to do some function following to see what's going on.

 

Before I go any further though... Am I getting into perilous waters with digging this far into online functionality? Does anyone finding some reliable values get classed as cheating online?

 

Edit: Ah.. okay... the random values are stored as Globals, that makes sense as it makes them impossible to get at from my perspective.

Thanks for this and data in the other posts!

 

Okay so it's distance and quite forgiving. But I'm still puzzled about why I with my CPU from -13 and adaptive V-sync could google the very first timing suggestion I found on google, set on loop and go for a nap and wake up with SWK while others struggle and fail to get it. 

 

Oh well, for what it's worth, here's a small gif from a slowed down video where you can see how it grabbed it for me. Note that it's quite in line with the highest point of her hair in the back and a bit to back from the hair thing she has in her hair.

 

https://i.imgur.com/rLwjae1.mp4

[lost-in-emotions]

HEEEYAA, i`ve got Wasabi finally, too!

With a big BUT: I try to aim for the location like in the GTA Series Video. IMPOSSIBRU!

The Plush is so silly in that corner, that the claw cant reach it with full W and full D.

 

So i tried the shake thing. At first the Plush hopped nearly on the same spot, but backwards.

So my Friend helped me again, and next the Plush hopped nearly in the middle of the machine.

 

 

And in this Location i could nearly hit the same spot like in the GTA Series Video and after ~ 20mins ....

 

 

[jamfamalam21]

I appreciate the traction and help we're getting on trying to figure this shid out, cheers my fellow sons. It seems like we've kind of gone in a circle though... SWK's diameter is 7cm and = to 40% of her (cheers for the info @LeeC22) - so although her head's f*ckin massive and supposedly easy to grab, then how the fubldyluck can @Fun 2's friend with a 100% grabbing rate still not wallop this bitch over the head?? What is the damn reasoning lol. I'm not a genius when it comes to the ins and outs of a game's coding but you can't just have unknown random sh*t that hasn't even been written to somehow appear from thin air and effect the actions of the claw game functioning..? Beats me

[jamfamalam21]

2 minutes ago, lost-in-emotions said:

HEEEYAA, i`ve got Wasabi finally, too!

With a big BUT: I try to aim for the location like in the GTA Series Video. IMPOSSIBRU!

The Plush is so silly in that corner, that the claw cant reach it with full W and full D.

 

So i tried the shake thing. At first the Plush hopped nearly on the same spot, but backwards.

So my Friend helped me again, and next the Plush hopped nearly in the middle of the machine.

 

 

And in this Location i could nearly hit the same spot like in the GTA Series Video and after ~ 20mins ....

 

 

epic style Alice, the Lord has truly blessed you my child. (ty @JuSuCri for finding the vid big up) I'm convinced that the easiest way to get this green haired slag is to glitch her the fr*ck out elsewhere in the machine to a point where, you know, I can actually PLAY THE GAME TO IT'S NORMAL FUNCTIONALITY without wanting to rip my left bollock off.

[LeeC22]

6 hours ago, jargonaut said:

Okay so it's distance and quite forgiving. But I'm still puzzled about why I with my CPU from -13 and adaptive V-sync could google the very first timing suggestion I found on google, set on loop and go for a nap and wake up with SWK while others struggle and fail to get it. 

Probably because the hit area is so large and movement in games is controlled by the frame update interval to keep the speed consistent across all machines. So 4000ms of input is going to equal virtually the same amount of claw movement on every machine, be it PC at 144fps or console at 20fps.

 

With movement of objects in games, you use the elapsed frame update time as a multiplier, so that the movement speed on every frame rate remains the same. Even if the game was to jump erratically from 10fps to 60fps, the speed would remain constant, that means the amount of input used remains constant. That's a core mechanic of keeping the game the same across all platforms and hardware.

[Seeker85]

Any PC player who wants to help me with the Glitch, please

[LeeC22]

4 hours ago, jamfamalam21 said:

I appreciate the traction and help we're getting on trying to figure this shid out, cheers my fellow sons. It seems like we've kind of gone in a circle though... SWK's diameter is 7cm and = to 40% of her (cheers for the info @LeeC22) - so although her head's f*ckin massive and supposedly easy to grab, then how the fubldyluck can @Fun 2's friend with a 100% grabbing rate still not wallop this bitch over the head?? What is the damn reasoning lol. I'm not a genius when it comes to the ins and outs of a game's coding but you can't just have unknown random sh*t that hasn't even been written to somehow appear from thin air and effect the actions of the claw game functioning..? Beats me

The random factor is a killer. Picking a random number in a range and then only taking 0 as a win, is like having a lucky wheel with 1000 prizes and only having one with the car on it. You've seen how hard it is to win the car with just 20, imagine how hard it would be with 1000. 

 

There seems to be 4 random ranges, one for PRB, one for SWK, one for Master and another for the rest. No idea what they are though, because they're globals... unless I have missed them in this thread somewhere. I think they have made the grab area so big, so that the random number can be the controlling factor. Hitting that grab-spot is easy, getting that random number... not so much.

 

Edit: Found those random ranges on an earlier post. I might try and recreate the machine as a mod for SP and get some display markers on there to try and see what is happening and what it is checking. Might make it easier when I can see things visually, to understand exactly what it is doing.

Edited February 4, 2020 by LeeC22

[Bardiamist]

2 hours ago, Seeker85 said:

Any PC player who wants to help me with the Glitch, please

Go. I have same nick in social club.

[Seeker85]

Hace 19 minutos, Bardiamist dijo:

Vamos Tengo el mismo nick en el club social.

Solicitud de amistad enviada 

Edited February 4, 2020 by Seeker85

[AshyraVash]

18 hours ago, jargonaut said:

I'm still not 100% convinced it's purely about the physical location of Wasabi (when in non-glitched position) or the position of the claw related to the plushie, but simply a matter of getting both directional timings right. Something you could in theory do with your eyes not on screen, but a stop watch. Hell, I would even have accepted a longer timing than what is physically possible visually (read: holding a directional button for a bit longer even after the claw hits a wall) cause a thing like that could have just been another way R* makes us guessing and testing.

 

But do this: Edit your script and make it try the nine positions below and loop it. This will extended the testing area a bit to all directions (cause I still believe it's actually 35 milliseconds and not millimeters Fun found in the code). Then make sure you have set notifications to 60 mins (if you haven't yet done it for whatever reason). Then restart your game to make sure you have clean cache, start the game in single player mode and jump to invite only session to make sure no-one has been there to make it dirty yet.

 

{w: 4200, d: 3870}
{w: 4200, d: 3865} <- my hit on 200th try of this bunch
{w: 4200, d: 3875}
{w: 4205, d: 3870}
{w: 4195, d: 3870}
{w: 4205, d: 3865}
{w: 4205, d: 3875}
{w: 4195, d: 3865}
{w: 4195, d: 3875}

 

As I said above, me hitting it exactly on 200th attempt and with no extended data, it would still be impossible to say the exact timing by millisecond, but visually on my screen, the first one fits and sits perfectly while all others including the one that picked it up fit inside the 35ms tolerance area. So it may have been that all these positions are inside the right area and for whatever comical überuniversal reason I needed 200 x 1/200 = 1. 

 

But change your script to loop those nine. My script used 30 secs / attempt, and it took 200 attempts or 100 mins to get it, quite precisely.

 

Edit: I'm going for my second nap now, but before that: Even the above set of timings doesn't look to be hitting exactly the center on your screen, humor me and let it loop that bunch.

I've let it run the loop for about two hours yesterday and about three hours today and nothing so far. While the coordinates seemed a bit too to the left initially yesterday, Wasabi did move into a centered position gradually and today it hasn't budged at all from the look of it. Something about it isn't adding up. Then again with the 40% grab location it shouldn't matter too much anyway if it's a bit off-center. Maybe I should just try to have it knocked out, too. xD I haven't had a whole lot of luck with the Lucky Wheel either, come to think of it, so it could just be bad luck of the roll.

Edited February 4, 2020 by AshyraVash
It appears I was mistaken, Wasabi has moved and is clipping into Master Hentai a little once more xD

[jargonaut]

50 minutes ago, AshyraVash said:

I've let it run the loop for about two hours yesterday and about three hours today and nothing so far. While the coordinates seemed a bit too to the left initially yesterday, Wasabi did move into a centered position gradually and today it hasn't budged at all from the look of it. Something about it isn't adding up. Then again with the 40% grab location it shouldn't matter too much anyway if it's a bit off-center. Maybe I should just try to have it knocked out, too. xD I haven't had a whole lot of luck with the Lucky Wheel either, come to think of it, so it could just be bad luck of the roll.

Thanks for update!

 

It still feels so crazy that others have so big issues with getting these things and for others it's just a matter of waiting a little/having a reasonable number of goes. 

[AshyraVash]

6 hours ago, Seeker85 said:

Any PC player who wants to help me with the Glitch, please

How did it go? I might be able to help in a bit, too.

[Seeker85]

would it work like that? I've been trying for an hour without results

 

https://ibb.co/KV8MXNx

 

[lost-in-emotions]

^ this looks like on my first try out. after ~ 30min. i gave up and asked my friend again.

after this the plush will fell more on the middle in the machine. then it worked for me

[Seeker85]

10 minutes ago, lost-in-emotions said:

^ esto parece en mi primera prueba. después de ~ 30min. Me di por vencido y le pregunté a mi amigo otra vez.

después de esto, la felpa caerá más en el medio de la máquina. entonces funcionó para mí

ah then it has to be more in the middle? OK thanks !!

[AshyraVash]

So I just did a bit of a test with jargonaut's little loop and went to find a new public session, ran the first coordinate/timing and the claw was landing too far to the left. I found another session immediately and tested again and this time the claw is actually centered. I didn't think to capture screens of it, but that proves Wasabi isn't spawning in the same location at least.

[LeeC22]

Okay, this video might be interesting to some. I created a mod to spawn the machine in single player, I added all the plushies at the stored locations, then I added a cylinder marker at another set of offsets that are used in the checking process. I also made the markers 7cm in diameter, so they should (if they're correct) show the exact grab area for each plushie.

 

Disclaimer: This is purely based on the info I have found and my interpretation on it based on the script contents. Like everything online, there's no guarantee that this is 100% of the story, there could be other things going on in other functions. There is another function that takes Vector magnitudes and does lots of processing with global variables, so I have no idea how much that affects things.

 

 

Edited February 4, 2020 by LeeC22