Favero on the New Timings

[Ripost]

http://www.favero.com/pdf/miss-hit%20en.pdf

[Alan]

"The hit given on hard surfaces (chest, b , etc) tends to bounce and
the impact time is shorter than a hit on soft surfaces. Too fast actions on these surfaces can not be signaled for their insufficient duration."


What is too fast ? Should we decree a speed limit in Foil ?

[esskreemr]

The speed limit has been declared. That's part of the issue.

So does this affect primarily German or French style tips?

[t('.'t)]

favero touched me once, and he said he would kill me if i told anyone...

[DHCJr]

The old style Prieur and Sport 7 would have an advantage. Remember the plastic shaft.

[oiuyt]

Originally Posted by DHCJr

The old style Prieur and Sport 7 would have an advantage. Remember the plastic shaft.

Care to further explain? Are you suggesting that the shaft of the tip would absorb some of the impact and debounce a bit slower? Something else?

-B :)

[edew]

I'm guessing that Donald's claim is that the plastic sleeve will absorb some of the banging so that the point won't bounce off as quickly.

I personally don't think that will affect much more than several hundredths of a mill-second.

Or, he could imply that the plastic sleeve will slide against the inside of the barrel with more resistance, hence the point would stay depressed longer. Time to bang my barrel up a bit for a nice locked hit, I guess.

But in either case, the length of time that the tip is depressed is irrelevant, because it's not how long the tip is depressed, but how long the tip is depressed while touching the lame. The hard shell underneath makes the whole tip bounce off quicker than usual and so the tip no longer maintains contact with the lame, independent of whether the tip is depressed or not.

[DHCJr]

Did either of you read the article? This has nothing to do with sliding or slowing the tip. What the article talked about was the touching of the tip and the piece that is touching the barrel. You have taped up the barrel and the blade, but there is a metal piece that is touching the inside of the barrel. Since it is touching the inside of the barrel and you touch the tip and anything that is electrically touching the barrel you will short out the tip causing no touch.

The prieur shaft is plastic, i.e. non-conductive, so the problem in the article could not occur. If they had poor taping on their barrel, there would be a problem, but that is not what the article is about.

Please re-read the article.

[keith]

well there are three issues;

1) microbreaks - which they claim to have fixed by programing.

2) the debounce on hard surfaces (manplates etc) no technical fix.

the figures concern point 1 and DHCJr is right that the old style LP or Prieur tips would be the best choice.

Also for the rules folks - point one of the document suggests a software change to help deal with microbreaks, is that legal? (remembering the issue around timing for on vs off target)

[mrbiggs]

The microbreaks thing is VERY interesting...maybe another explanation for why some of these straight attacks aren't registering.

[jBirch]

Originally Posted by keith

well there are three issues;

1) microbreaks - which they claim to have fixed by programing.

2) the debounce on hard surfaces (manplates etc) no technical fix.

the figures concern point 1 and DHCJr is right that the old style LP or Prieur tips would be the best choice.

Also for the rules folks - point one of the document suggests a software change to help deal with microbreaks, is that legal? (remembering the issue around timing for on vs off target)

I was wondering about that too. Basically the software would be an algorithm that would reject the break as an outlier and keep the counter going. My question is around the legislated length of the break. Essentially, the rule would need to read, "debounce of 14ms with no microbreak greater then 1ms" or something.

I'd be interested to know stochasitically which of these three problems is causing the majority of misregistered hits in foil.

James.

[keith]

I'd be curious to know how they have set up the new contact timing, and more interestingly how long shorts are allowed to be within the 15msec window.

This fix should not have any effect on making flicks easier (they just don't contact long enough) or the hard straight hit not registering but the scittering across the lame hits may be back in play with these modifications.

...but I am feeling foolish for switching to the LP GT tips.

[keith]

Originally Posted by jBirch

I'd be interested to know stochasitically which of these three problems is causing the majority of misregistered hits in foil.

James.

well this is the interesting bit. If you just sample by asking after each event if the signal is present 15ms later then you could get some very odd results.

or you do as you suggest an say that in each 15ms period no more than a 3ms (for example) break is allowed.

the second way is a better sampling technique but I have no idea which of the two techniques would give a better result for foil

[DieterS]

As a manufacturer of a popular scoring machine, I can easily imagine how a competitor could creatively interpret the new rules to mitigate "micro-breaks." Here's how.

Lessee, the rule for foil debounce/dwell time is now up to 15ms and the interruption tolerance is 0ms (the tip must be depressed continually).

To smooth over "micro-breaks," simply increase the time required for any interruption to register! How about a 2ms "micro-break" interruption lockout? A 5ms lockout of "micro-break" interruptions is even better. Not satisfied? Well, I'll tell you what. How about going all the way? With a 14ms "micro-break" lockout, we can all essentially return to the historic 1-5ms debounce time!

Dieter
Eigertek

[oiuyt]

Originally Posted by DHCJr

Did either of you read the article?

How do you expect me to EVER get to 2400 posts if I need to spend time reading articles before asking questions about them? :)

Hrm, okay, off to read the article...

-B :)

[mrbiggs]

Originally Posted by oiuyt

How do you expect me to EVER get to 2400 posts if I need to spend time reading articles before asking questions about them?

Hrm, okay, off to read the article...

-B

Personally, I just get the gist of it from the link. In this link, I can see a "Favero", so I know it has something to do with Favero,for example.

[DanInMI]

Originally Posted by DieterS

As a manufacturer of a popular scoring machine, I can easily imagine how a competitor could creatively interpret the new rules to mitigate "micro-breaks." Here's how.

Lessee, the rule for foil debounce/dwell time is now up to 15ms and the interruption tolerance is 0ms (the tip must be depressed continually).

To smooth over "micro-breaks," simply increase the time required for any interruption to register! How about a 2ms "micro-break" interruption lockout? A 5ms lockout of "micro-break" interruptions is even better. Not satisfied? Well, I'll tell you what. How about going all the way? With a 14ms "micro-break" lockout, we can all essentially return to the historic 1-5ms debounce time!

Dieter
Eigertek

Cmon. As "As a manufacturer of a popular scoring machine" you know that microbreaks are not about how long the tip is depressed. It is about interuptions in the perfect electrical connection.

Without allowing for microbreaks you must have a perfect connection for the entire duration of the touch. If you were going to design an efficient switch it certainly would not involve taking a metal button that you depress by pressing it against a sweaty piece of fabric that is on a moving, uneven surface, and sliding it, would you? It makes sense that there might be brief interuptions in the connection, or drops in the OHMs level below the appropriate levels, even though the tip has stayed continually in contact and depressed.

If some machines are programmed to make allowances for that and some are not that would explain why there may be valid touches not registering on some machines, wouldn't it?

[keith]

Originally Posted by DanInMI

Without allowing for microbreaks you must have a perfect connection for the entire duration of the touch. If you were going to design an efficient switch it certainly would not involve taking a metal button that you depress by pressing it against a sweaty piece of fabric that is on a moving, uneven surface, and sliding it, would you? It makes sense that there might be brief interuptions in the connection, or drops in the OHMs level below the appropriate levels, even though the tip has stayed continually in contact and depressed.

The point DieterS raises is actually very important - Applying a filter over a set of events of variable duration can lead to exactly the problem he raises, short duration events interpretted as long duration events by the algorithm.

Unless you actually know the type of events you are trying to filter out noise from you have no real idea how your post filter signal relates to your input signal. One thing that does appear to be quite clear from the new timings is that no one hooked a box up to a high speed recording device to determine what the actual signals looked like during a bout. Now maybe they decided that they didn't have to but the problems with debounce and microbreaks indicates they should have.

Depending on your choice of filter you can get all sorts of interesting interpretations of your input signal - and quite possibly all of them can be wrong if you don't have a proper appreciation of the input .

[Craig]

Another question: does the spec issued by the FIE even allow for micro-breaks? My guess is no.

Craig

[DanInMI]

Originally Posted by Craig

Another question: does the spec issued by the FIE even allow for micro-breaks? My guess is no.

Craig

My guess is no also. Perhaps they should.

It seems to me that if each company that is manufacturing boxes is creating thier own code that there can be drastic differences in how they behave. The FIE should own the code and make it available to those that would create the boxes so that they are all alike. So, for example, if Favero creates a code that works better than all the others, the FIE would select it and buy or license it from Favero and distribute it to other manufacturers. OR they could require Favero to directly licence it to others so that Favero does not hold a monopoly on scoring machines.