Tricking Scatt with Fake Targets

Your highest shooting achievement
10m pistol France championship qualifications
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
May 13, 2020
99 Posts
64 Helpful Votes
0 Best Q&A Answers
Scatt calibrates on itself. Literally.
Here is a target "black center" on which my MX-W2 calibrated without the slightest problem:
TrickTarget-Scatt.JPG
Of course there's a story behind this trick. 3 days ago I posted the Q&A tread "Question about reduced target size", in which I wondered how the Scatt -- at least the wireless MX-W2 -- handled the real-distance 10m pistol shooting although the sensor is in fact around 9m from the target with your arm and pistol straight in shooting position.
As I tried to solve the issue I came to one hypothesis, that the Scatt does not measure, nor even use at all, the black round shape of the target center, meaning all it needs is a roughly dark shape with some minimum contrast with the surrounding white area. Then of course I had to prove this hypothesis. Well, it checks out perfectly.

Here are the other trick targets that I tested. The surrounding square is added for comparison, it's side is 8cm.
TrickTarget-8meter.JPG TrickTarget-5meter.JPG TrickTarget-FlatStar.JPG TrickTarget-Square.JPG TrickTarget-PeakStar.JPG TrickTarget-Scatt.JPG
- A scatt-printed target for 8 meters, titled "10m Air Pistol Target. To be placed at 8m from shooter". The black circle is 4.8 cm in diameter
- The same printed for 5 meters. The black circle is 3.0 cm in diameter
- A star with outside diameter 6cm and 1cm peaks
- A 4cm square
- A star with outside diameter 9cm and 3cm peaks
- At last the Scatt name.
The first two black rounds were printed with the Scatt target printing tool and the others with a drawing software. Just the black shape on a blank white page.

Now what about the scores? Well in a single word: Identical. How did I check it? Using one of Scatt's most interesting tools, the trace speed, which measures the moves and is an indicator of stability. If the Scatt sensor & software did use the black shape size at all, my speeds would be very different with the 5m target vs. the 8m, and considering that 8m is actually the distance I'm shooting the 5m should have given way out of range values. But id dit not. In fact the average values are extremely close, with a 4% standard deviation for S1 and 5.5% for S2 (which by the way also shows that I was shooting very consistently).
There is now way you can compare these shapes in terms of size. Which measure is meaningful? outside? inside? average radius? What is the size of the Scatt logo? This just shows one thing: the shape does not matter very much.

Anything special about my target setup? Not really, in fact it was only average in lighting and poor in contrast. You can barely see the white sheet of paper on the bookshelf side, with the scatt logo in the middle. The sensor calibrated immediately on it.
IMG_20200601_175928.jpg

That's it for today. I had a lot of fun doing this trick, and debunking a tiny piece of Scatt's secrecy was a delight.
What about a "Weird tricky target" competition in the forum?

David
 
Last edited:
OP
D
Your highest shooting achievement
10m pistol France championship qualifications
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
May 13, 2020
99 Posts
64 Helpful Votes
0 Best Q&A Answers
By the way there is an interesting information from this test, at least with the MX-W2 (and probably MX-02 too):
Since you can print your target the size you want without any effect on the scores, closer target should better be scaled relative to the eye distance, not the sensor. By doing this way the size of the target black area would stay consistent relative to the sights.

Otherwise if you measure from the sensor the target looks smaller. For instance if you shoot 50m rifle (22LR), a custom 10m setup will show you a -9% target size, and a custom 5m setup will show you a -19% visually smaller target, which is simply impossible to aim at correctly with the tunnel sight of a rifle.

Now why does Scatt recommend measuring target distances from the sensor? Well I don't really know. Not sure they will want to disclose an answer. An hypothesis is that older sensor models worked differently and needed this constraint.
 
Your highest shooting achievement
GB Team
Skill Level
Amateur/Hobby Shooter
Primary Discipline
High-power Rifle
SCATT Experience
Over 10 years
Joined
Mar 10, 2020
73 Posts
78 Helpful Votes
0 Best Q&A Answers
Good test! This shows that MX-nn camera sensor Scatt is simply looking for the centre of the shape as its reference point and everything is calculated based on the angular displacement of the sensor from that centre, using the known angular dimensions of the scoring rings on the target selected.
As you say, this allows you to print a scaled target that matches the angular size of the real target, as seen by the eye, so the distance of the target from the eye is the distance required when printing the target.
 
Your highest shooting achievement
ramon@olympicpistol.com
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
Mar 10, 2020
63 Posts
29 Helpful Votes
1 Best Q&A Answers
As I understand it, we have to measure from the sensor to set the distance ring on the sensors of the MX devices and from the shooting line for the target resizing so that we have a similar apparent target size. Different shooters, different body lengths but the shooting line remains at the same distance (just like at the range).

The SCATT USB device always automatically measured less than the distance you were shooting at (I had to check the lanes at my range to make sure that they were at 10m after being puzzled by the distance measured by the USB). SCATT confirmed that the device was measuring its own distance to the target frame and that it was not very precise.
 
OP
D
Your highest shooting achievement
10m pistol France championship qualifications
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
May 13, 2020
99 Posts
64 Helpful Votes
0 Best Q&A Answers
The SCATT USB device always automatically measured less (...)
SCATT confirmed that the device was measuring its own distance to the target frame and that it was not very precise.
I'm not a Scatt employee and don't actually know how various devices work. From my test I would say the Scatt MX-02 and MX-W2 don't measure the distance, and additionally that they don't need to. But for the USB device I have no idea.

we have to measure from the sensor to set the distance ring on the sensors of the MX devices and from the shooting line for the target resizing
Not exactly, and by the way the shooting line is irrelevant.
- From the sensor to the target for the focus ring
- Preferably from the eye to the target for target resizing.
 
OP
D
Your highest shooting achievement
10m pistol France championship qualifications
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
May 13, 2020
99 Posts
64 Helpful Votes
0 Best Q&A Answers
Any comment from Scatt? Peter?
 
Your highest shooting achievement
ramon@olympicpistol.com
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
Mar 10, 2020
63 Posts
29 Helpful Votes
1 Best Q&A Answers
Not exactly, and by the way the shooting line is irrelevant.
The shooting line is relevant because you are simulating shooting at the range and you can not step over it. Every shooter positions themselves a bit different depending on the line and the range's table. This is the variable I was not able to take into account when making the target calculator.

The software takes into account movement of the weapon, where it is aiming towards, and distance to calculate the trace and impact adapted to the shooting distance.

The sensor sending all this info is attached to the weapon, so the distance to it is needed to calculate. The dial is not very precise though. I wonder how they do their magic.

Maybe I misunderstand the objective of your experiment.

Have you considered clamping your pistol or sensor to run this tests and eliminate variables that could bias it?

Can you please explain the statistic part so that even I can understand it?
 
OP
D
Your highest shooting achievement
10m pistol France championship qualifications
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
May 13, 2020
99 Posts
64 Helpful Votes
0 Best Q&A Answers
Maybe I misunderstand the objective of your experiment.
The objective of my experiment was precisely to get a hint as to how the MX-W2 finds the distance to the target, if it uses at all a visual measure of the black area it "sees", and if yes, how.
It rose from the fact that when you shoot with a pistol at a 10m range the sensor is in fact roughly at 9m at the far end of your arm and pistol. And from the second fact, that if you use Scatt instructions to set up a custom range and the measure from the sensor happens to be 9m, meaning you are at exactly the same distance as the first situation, you should use a custom-printed target 90% the original size. Hence the target size question.

In my interpretation the answer is that the Scatt does not "measure" that distance at all. It does not even have to compute it. And that the size of the black does not matter at all. Once again this is is for the MX-W2 and probably the MX-02 and all models using the same sensor.

Going on with the trick targets was just a funny mean of confirmation, which turns out positive.

There are rather easy explanations of how that is achieved. I also suppose examination of Scatt's patents would confirm. If I have time I may discuss that in the "Lobby" forum for those interested and for the fun of it [personal opinion here]
 
Your highest shooting achievement
ramon@olympicpistol.com
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
Mar 10, 2020
63 Posts
29 Helpful Votes
1 Best Q&A Answers
The objective of my experiment was precisely to get a hint as to how the MX-W2 finds the distance to the target, if it uses at all a visual measure of the black area it "sees", and if yes, how.
Thanks for the explanation David. The MX sensors don't guess the distance, you set it up with the dial. Only the sensors with a frame measure themselves the distance to the target. At least this is what I've understood from back and forths with SCATT support :)
 
OP
D
Your highest shooting achievement
10m pistol France championship qualifications
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
May 13, 2020
99 Posts
64 Helpful Votes
0 Best Q&A Answers
Can you please explain the statistic part so that even I can understand it?
That's gonna be a post quite longer than usual, I hope it will be worth it.

I shot six targets, one for each of the experiments. The first one was scaled according to Scatt instructions for the setup I use in my living room (see picture) and the following 5 are contradictory tests.

The S1 average values for each of my 6 targets are the following (please excuse the terrible values, that was a bad day):
195.7 / 216.3 / 204.5 / 196.5 / 207.4 / 203.9, with an average of 204.1 and a standard deviation of 7.6
The S2 average values are:
183.2 / 214.7 / 210.0 / 208.1 / 204.9 / 206.5 with an average of 204.6 and a standard deviation of 11.0
See standard deviation (statistics) on Wikipedia. The ratios of the deviations to the averages are 3.7% and 5.3%. They give a "common sense" idea of the dispersion of the values around their averages. In the present case this dispersion is extremely small.
The target scores range from 87 to 91 and are absolutely consistant with the S1 and S2 values for one of my bad day trainings, and I have almost 100 Scatt trainings that confirm these values.

Most interesting is the maximum spread between values, which are 20.6 = 10% for S1 and 31.5 = 15% for S2.
Now consider the first two targets: The first one is the "correct" reduced target as to Scatt instructions at 8m, my home setup distance, and the second is the "correct" for 5m although it is still set up at 8m. If there was a relation in the Scatt's computations between the black area and the distance, the first target would give correct results and the second way incorrect results, with values 8/5 = 60% higher. It is obviously not the case. In addition the values and scores would supposedly be out of any reasonable scale for the other 4 test targets which use purposely quite improbable shapes of largely varying sizes (as shown by the 8cm square common to all). There again the results remain consistant will the "real" target.

My interpretation is that the sum of these results suffices to conclude that the size of the black center is not used at all in any kind of score or distance estimate in the Scatt hardware/software system. As i wrote just above there are very reliable ways to explain who that would still work so well.

In the present case statistics were an interesting way of getting a high confidence proof estimation, because the values turn out to be so close to each other and especially the standard deviations are so small.

Does that answer your question?
 
Last edited:
OP
D
Your highest shooting achievement
10m pistol France championship qualifications
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
May 13, 2020
99 Posts
64 Helpful Votes
0 Best Q&A Answers
The MX sensors don't guess the distance, you set it up with the dial.
Once again I think this not correct. The dial on the sensor is only a focus dial, as on a camera lens. It is necessary so that the image sensor gets a focused image of the target, exactly like in you eye or a camera. Nothing more.

The distance of the target of standard size is given by your choice of target, like 10m air pistol, 50m rifle or whatever you practice. A good example can be found when starting a new session with the "25m precision and 50m pistol" target. It is the same target and same size (called C50, black is 20cm in diameter) but in addition you have to indicate your shooting distance at the range, 25m or 50m. If you try both at home you should get scores with a 1/2 ratio between the two options. I'll try it since I have such targets.
 
Last edited:
Your highest shooting achievement
GB Team
Skill Level
Amateur/Hobby Shooter
Primary Discipline
High-power Rifle
SCATT Experience
Over 10 years
Joined
Mar 10, 2020
73 Posts
78 Helpful Votes
0 Best Q&A Answers
Once again I think this not correct. The dial on the sensor is only a focus dial, as on a camera lens. It is necessary so that the image sensor gets a focused image of the target, exactly like in you eye or a camera. Nothing more.

The distance of the target of standard size is given by your choice of target, like 10m air pistol, 50m rifle or whatever you practice. A good example can be found when starting a new session with the "25m precision and 50m pistol" target. It is the same target and same size (called C50, black is 20cm in diameter) but in addition you have to indicate your shooting distance at the range, 25m or 50m. If you try both at home you should get scores with a 1/2 ratio between the two options. I'll try it since I have such targets.
I agree with Davidf92 - it appears that the dial on the camera sensor is only for focusing the camera (so the distance from target to sensor is the critical measurement). His trial with different aiming mark sizes and shapes strongly suggests that the camera doesn't mind what it sees, provided there is a dark shape that it can consistently differentiate from a light background, and that it can consistently work out the centre of the dark shape, which becomes its reference point (datum). If this is the case, Scatt then works everything out from the angle the sensor is pointing away from this central reference point, and the direction (as on a clock face). This is commonly known in maths as a polar coordinate system (as opposed to standard x-y cartesian coordinates). The scaling is determined purely by the target that has been selected, for which the scoring rings subtend known angles from the centre.
With USB Scatt, the system is completely different, in that the sensor measures the position of the array of IR 'lights' on the target frame and determines where they are within the sensor field of view relative to the 'centre' (determined by the calibration shot (coarse) and subsequent (fine) zeroing from the sighter shots). Scatt almost certainly calculates the distance of the sensor from the target by the angular size of the target array that it sees (which is larger in the sensor's field of view the closer the target frame is). The accuracy of this distance calculation will increase the closer the target is, but there are other reasons why it is better to have the target at a longer distance.
 
Your highest shooting achievement
GB Team
Skill Level
Amateur/Hobby Shooter
Primary Discipline
High-power Rifle
SCATT Experience
Over 10 years
Joined
Mar 10, 2020
73 Posts
78 Helpful Votes
0 Best Q&A Answers
The distance of the target of standard size is given by your choice of target, like 10m air pistol, 50m rifle or whatever you practice. A good example can be found when starting a new session with the "25m precision and 50m pistol" target. It is the same target and same size (called C50, black is 20cm in diameter) but in addition you have to indicate your shooting distance at the range, 25m or 50m. If you try both at home you should get scores with a 1/2 ratio between the two options. I'll try it since I have such targets.
Davidf92 - presumably pistol shooters find zeroing for different targets/distances to be more difficult than rifle shooters. For a rifle shooter using a ring foresight the aim is always at the centre of the target, whatever the size or distance of the aiming mark. For a pistol shooter, adopting a 6 o'clock aim, the zero will depend on the size of the angle between the aim point (below the black) and the centre of the target. At 50m the "25m precision and 50m pistol" target appears to be half the size of the same target at 25m, so the angle between the aim point and the centre of the target will be roughly twice as big at 25m as it is at 50m (depending on how far below the black your aim point is for each distance).
 
Your highest shooting achievement
Double Distinguished
Skill Level
Amateur/Hobby Shooter
Primary Discipline
High-power Rifle
SCATT Experience
Over 10 years
Joined
Apr 9, 2020
17 Posts
9 Helpful Votes
0 Best Q&A Answers
HHHHHuuuuummmmmm, very interesting. i was having trouble getting my MX-2 to work, and emailing scatt, i did not have the distance set properly. i had measured to the line. when i changed it by about a meter, target and focus, it started working, BUT like you said, the target looked to small. I didnt realize it was just the focus setting that fixed it. I will try this again, with my own experiment. ill set the sensor to target distance from it, and print the target from eye distance. should take care if it. ive been using my WS-1 set at sensor distance, with a 10m size target and it works fine. you have solved a my problem without knowing it!!! Thanks
And i have to add, great piece of equipment, and support is awesome!!
 

Peter

Administrator
Skill Level
Amateur/Hobby Shooter
Joined
Sep 9, 2019
287 Posts
190 Helpful Votes
1 Best Q&A Answers
Hello everyone!
I hope that this will cover pretty much everything, but don't hesitate with your follow-up questions ;)

1) Regarding the shapes and the experiment:
Your experiment is correct in regard to concluding that a round shape of the target is not mandatory for the sensor to operate. BUT, the cleaner, more symmetrical, and sharper in terms of contrast the target on that piece of paper is – the more precise coordinate calculations are going to be. You can dive deeper and actually conduct an experiment comparing visual qualities of the trace and the S1 parameter data across different shapes – I’m sure you’ll arrive at the same conclusion that the round shape is the optimal one.

P.S. A target too small or too big may not be recognizable by the camera.
P.P.S. As a rule of thumb, asymmetrical targets are going to have their logical centers that don’t match its geometrical center.

2) Scaling down targets:
Printing out a target for 9.1 meters for outdoors in the open when you essentially have all the conditions to practice at real 10 meters. In this case, feel free to print out the target for 10 meters, since this is not a genuine reduced distance situation.

It’s crucial to understand that any other case the distance between the shooter’s eye and the lens doesn’t proportionally change with the distance. Therefore, there is no way to get the perfect textbook match due to how physics/geometry works. On top of that, the less is the physical distance we are dealing with (2.5-5 meters being the worst offenders) – the bigger discrepancy we get.

If we are looking at the issue solely from the perspective of the shooter’s comfort and his/her visual perception of the target during the practice session, then it would make sense to print out the target in regards to the distance measured from the shooter’s eye. Here’s an example: let’s say you are used to standing with your eyes 10.1m from the target at the range, right? Then transferring the same setup indoors (let’s say we have only ~5 meters to work with at home) would imply that the shooter stands with his eyes 5.05m away from the target (while the target itself is printed for exactly 5 though). At the same time, the distance between the lens and the target is just 4.1 and that is fine, for the sensor should have no issues getting precise results with simple round black targets (albeit score rings make it worse, therefore we usually recommend printing the target without them).

The eye will perceive the target as totally normal in this scenario. However, the trainer’s readings are going to differ from the standard 10m session, which is especially true when it comes to AirPistol (due to angular oscillations caused by the wrist that implies the barrel’s axis wiggling and shifting against the “sights-eyes” axis). Rifle shooters got it easier in that regard.

The perfect identical match and balance between the shooter's perception and sensor's readings are impossible to get due to how the world works so to speak. Think of reduced distance training with SCATT like if it was a stationary exercise bike – yes, it enables you to do the same thing you would do in the open, namely work the pedals – but some nuance of a real race is missing no matter how high-quality this simulation is. You are rocking that exercise bike, because there's a hurricane outside and as soon as it blows over it would make the most sense to get on a real bike in actual racing environment.

As you can see I'm awful with analogies, but at the end of the day: the best, the most precise and beneficial way to practice with SCATT is at the actual distance the event/exercised is designed for. Thus, if you have a choice to practice AP 10m at 10m, you should pick this option 100% of the time.

From the thread on lateral movements started by Tarta86 quoting Davidf92
of course this is not an issue concerning the Scatt itself, no malfunction there. It is just a side-effect and limitation of simulation at a closer distance. Obviously training the 10m at 2m is close to impossible even if the Scatt is optically able to do it. Contrarily, training 300yd at 10m should not be too much affected since you are not likely to sway your body enough for the pb to arise”

Actually, we encourage you guys (and everyone else interested) to feed us your “settings” so to speak, including the distance, size of the printed targets, distances between the eyes vs. sensor, eyes vs. the target, etc. Once we get enough data, a big enough sample – we might implement a better, a more customizable “smart print-out” feature.

3) Measuring the distance:
MX-series doesn’t measure the distance to the target, since that is not needed whatsoever. The device is working with visual patterns and the distance you set before going into the exercise. We are focusing on angular deviations first and foremost.

4) The lens itself:
The lens is needed only for the device to focus on the target. It’s not involved in actual calculations.

5) The way Infra-red and Camera-based models operate is totally different, there is literally no point in comparing the two.
 
OP
D
Your highest shooting achievement
10m pistol France championship qualifications
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
May 13, 2020
99 Posts
64 Helpful Votes
0 Best Q&A Answers
Thank you @Peter for the explanation and many details.

One bizarre thing is that your pattern of answer is like "yes you can do this but it would be better if you did that". Somehow I get the feeling that you are missing a point here: for those who are not a professional or a trainer or high-level competitor, but just the guy next door who likes shooting sports, there is a huge difference between simply understanding by the user's guide that the Scatt can be used at shorter distances at the condition of using software-printed targets, or conversely being able to pick-up a sheet of paper and draw as target any roughly round shapes of approximate size. The first is a constraint while the second is a liberty. The second means that I can carry my pistol and set up a custom training place anytime anywhere without needing a printer. All I need is my tablet PC and a lazy black marker in my pistol box and not even a ruler. And whether I draw a nice round or a square doesn't even matter. I'll just draw it a bit small at first and widen it if it's looks smaller than usual in my sights. I can train in motel rooms every evening along a business trip the week before a competition whatever the available space, or during a one-month vacation trip without even knowing where I'll sleep the next night. In fact my only limit becomes the local right to carry my gun. This is huge. I don't even understand that you don't make it a marketing advantage.

I will definitely carry on with experiments when I have time, in even the most unfavorable conditions, to test the limits with my MX-W2 and share the results here for those like me who would like to enjoy the pleasure of practicing and training more often than they can go to a shooting range.
And since I am only a very average shooter, it would be great that better ones take on at their level to test the scoring precision issue, and still others with a variety of riffles and pistols. We'll see.

Of course the best and most obvious would be that Scatt people themselves did the job. You manufacture great piece of hardware and don't even tell us what we can do with it ;)
 
Last edited:
OP
D
Your highest shooting achievement
10m pistol France championship qualifications
Skill Level
Amateur/Hobby Shooter
Primary Discipline
Air Pistol
SCATT Experience
Less than 1 year
Joined
May 13, 2020
99 Posts
64 Helpful Votes
0 Best Q&A Answers
The way Infra-red and Camera-based models operate is totally different, there is literally no point in comparing the two.
Could you please clarify which product in your catalog uses which technology, hence the extent of customary use for each of them?
 

Peter

Administrator
Skill Level
Amateur/Hobby Shooter
Joined
Sep 9, 2019
287 Posts
190 Helpful Votes
1 Best Q&A Answers
Top