What can Canada do to nurture its chess talents?

[Carl Bilodeau]

Re: What can Canada do to nurture its chess talents?

...

Anyway, Jean seemed to not want to pursue the 'eliminate class prizes' issue further, which is a main thing he referred to, besides the sponsorship issue and his less specific guidelines, and wishes, on this message board, as far as I recall.

....

On this week HPE we could read that Jean was leaving today for europe for a trip of two majors tournaments where he will have two opportunities to have GM norms.

The first tournament start on saturday. There will be 600 players and 100 GMI. It is called: 26ième Open International d'échecs de Cappelle-la-Grande. It will be from february 13 to 20.

See the website:
http://www.cappelle-chess.fr/fr2/default.php



Good luck Jean!

Carl

[Carl Bilodeau]

Re: What can Canada do to nurture its chess talents?

A more likely scenario is that eventually all computer chess programs beyond a certain strength and playing at a long enough time control will invariable play each other to a draw, because they will be good enough to avoid any mistake large enough to cause forced loss of the game.

I believe this will happen, and perhaps not too far off. I do wonder what the time control would have to be.

If and when it does, we will finally be able to say that best play leads to a draw, and that anything but a draw is not best play.

In fact I don't agree with you once more. I studied in Computer and Operationnal Research in the University of Montreal in the Mathematic department. A bachelor that was half math and half computer systems.

In Operational Research we learn in the basic that some problems are exponential and can not be solve by a linear solution.

Chess combinaisons are exponential on a graph. It is exponential essentially because every move you add in the game this move has to be multiply by all the other moves. So on a graph it goes right to the sky in a matter of a few moves.

A computer is linear. If you have one processor on your computer and add a second processor you only double the capacity and this is called linear. If you build a computer 1 millions times faster it is still linear since it is only a linear multiplication of the capacity.

On this graph I show you two curbs:

The red line is exponential and the green one is linear. You can see that both lines meet at 9 on the graph. But for the green line to go up to number 16 would need so much more computers. Then let say the next point of the red line would be 50 which would be so high on this graph that it would require bilions of computers with each one having bilions of processors and take still some billions years to solve one game. This is why you have the nalimov tables for the 3,4,5 and 6 pieces and will never get the 9 or 10 pieces nalimov since it is exponential. Everybody would like to have a 64 pieces Nalimov table but nobody will ever have one.

Iif you home computer can compute 15 moves ahead, a computer a millions times faster could probably take a century to compute a few additionnal more moves. And it would take this super computer a billion years to add only one more move.

To acheive what you say would require different type of computers that could do exponential computations and I doubt this is possible even with quantum computers in a century.

So with a linear computer we have to cheat when we want to solve an exponetial problem and this is what Fritz does. This means that we give the computer some rules so that he computes only some major threads. Fritz would loose to any exponetial computers if it ever existed.

Here is the linear computer processor increase graph over time:


Carl

[Paul Bonham]

Re: What can Canada do to nurture its chess talents?

In fact I don't agree with you once more. I studied in Computer and Operationnal Research in the University of Montreal in the Mathematic department. A bachelor that was half math and half computer systems.

In Operational Research we learn in the basic that some problems are exponential and can not be solve by a linear solution.

Chess combinaisons are exponential on a graph. It is exponential essentially because every move you add in the game this move has to be multiply by all the other moves. So on a graph it goes right to the sky in a matter of a few moves.

A computer is linear. If you have one processor on your computer and add a second processor you only double the capacity and this is called linear. If you build a computer 1 millions times faster it is still linear since it is only a linear multiplication of the capacity.

On this graph I show you two curbs:

The red line is exponential and the green one is linear. You can see that both lines meet at 9 on the graph. But for the green line to go up to number 16 would need so much more computers. Then let say the next point of the red line would be 50 which would be so high on this graph that it would require bilions of computers with each one having bilions of processors and take still some billions years to solve one game. This is why you have the nalimov tables for the 3,4,5 and 6 pieces and will never get the 9 or 10 pieces nalimov since it is exponential. Everybody would like to have a 64 pieces Nalimov table but nobody will ever have one.

Iif you home computer can compute 15 moves ahead, a computer a millions times faster could probably take a century to compute a few additionnal more moves. And it would take this super computer a billion years to add only one more move.

To acheive what you say would require different type of computers that could do exponential computations and I doubt this is possible even with quantum computers in a century.

So with a linear computer we have to cheat when we want to solve an exponetial problem and this is what Fritz does. This means that we give the computer some rules so that he computes only some major threads. Fritz would loose to any exponetial computers if it ever existed.

Here is the linear computer processor increase graph over time:


Carl

Nice pictures, Carl, I'm glad being in the Army didn't make you forget all your math. It's good that you remember the difference between exponential and linear, which most everybody has learned in high school.

I'm familiar with everything you describe. But it appears you misunderstood what I am claiming will happen at some point in the fairly near future.

I did not claim that computers would solve the entire chess search tree. The Milky Way Galaxy couldn't hold a computer big enough to do that, even if time wasn't an issue, because every position would require storage. There are more positions than atoms in the entire Milky Way.

You have to ask yourself, in the late 1990's humans were still able to beat the best computers even though those computers could examine millions upon millions of positions, something humans have no capability for, so how did the humans win? They won by understanding positional strengths and weaknesses and long term strategies better than the programmer could write into the chess program.

Now it appears the best humans cannot beat the best computers any more in normal time controls. How have the computers improved so dramatically while humans haven't? To a certain degree, it is because the programs can search further down the tree, but the limitations on that growth are being met now, as you pointed out. Yet computer chess programs are still getting stronger, so what is happening? The programmers are coming up with better algorithms to score a position.

Let's say we take the best 8 computer software programs in the world today and play them against each other in a neverending double round robin. Let's put them each on identical hardware, say a quad core Intel, doesn't really matter. Let's make the first double round robin, they play each other with blitz time controls. Obviously there's going to be lots of wins and maybe a few draws.

Now the next double round robin, let's make the time control game in 10 minutes, twice as long as blitz. You may still see lots of wins, but more draws.

Next double round robin, double the time control again, now it's game in 20 minutes. More draws should appear.

With each new double round robin, double the time control. Each one should see more draws than the previous one, within some statistical deviation.

My theory is that eventually you will reach a time control at which none of the programs can defeat any of the other programs. All games will be draws. At that time control, you can continue to hold double round robins ad finitum ad nauseum, and you will see nothing but draws, ad finitum ad nauseum.

If you tried this with humans, the results would initially look the same. But eventually with humans, the time controls get so long that their concentration and calculation can't be maintained, and some of them would lose due to mental fatigue. Computers don't have this problem.

It would be great if someone could actually try this experiment with the programs and document the increasing number of draws, and perhaps draw a pretty graph that would point to a time control at which the infinite draws should occur. Or perhaps they don't have to draw the graph, perhaps the time control is reasonably short enough that it can actually be observed.

Within a few decades at most, assuming Mayan prophecies do not come true and the world does not end in 2012, the time control at which this occurs will be short enough to be observed and documented. That is my opinion.

It has little to do with searching more nodes of the chess tree. It has everything to do with mathematically formulating all the essential strengths and weaknesses in ANY random chess position, to such a fine point that you never make a mistake big enough to lose the game.

Actually, this might happen a lot sooner. I currently work with someone who has a Ph.D in Chemistry and a Ph.D in physics. He's only in his mid-30s. He has done research work on quantum computers, a whole different ball of wax that if one ever becomes reality, everything we know will change overnight.

I'm skeptical. Quantum computers seem paradoxical, like travelling backwards in time (something I believe they are supposed to be able to do). My branch of interest has more to do with neural networks, something we know can be done because we all have an incredible example of one inside our head.

[Paul Bonham]

Re: What can Canada do to nurture its chess talents?

You're putting too much emphasis on knowledge of opening theory, in my opinion.

Then please give us your explanation of why, at the same time that we have exploding growth of chess opening databases online, or on DVDs, easily updateable and browseable for anyone with even an old clunker of a PC, we are also seeing explosive growth of teenage chess superstars and even GMs? Kasparov, whatever you may think of him, got at least this one right. The young mind is most capable of learning and memorizing all this information, just as it is with learning new languages. Older minds are more hard-wired and can't rearrange themselves nearly as well.

Sorry, but chess, like a sport, is a game of winners and losers. There may be no way to sugar-coat it without changing the game beyond recognition, and that might easily become like a mad experiment gone horribly wrong.

Of course there will always be winners and losers, but when these two groups are very static, the losers go away, and a new bunch of losers must be recruited. We need to make the two groups more dynamic, not as dynamic as in poker, but more dynamic then they are now. My idea of using brilliancy and strategy prizes could do this, albeit yes, it would be an experiment. If it does go horribly wrong, should be easy to go back to the old system.

Oh, and as far as players deliberately making bad moves to prevent their opponent winning a brilliancy prize, the judges should be expert enough to see through that. The brilliancy is awarded based not just on the specific line that was played, but on whether a particular move in what seems an even position is irrefutable even against best play.

These type of prizes, and incorporating parallel chess960 events, don't change the game beyond recognition IMO.

Retaining newbies is nice, but sometimes there's no stopping quitters. Replace them at a rate faster than they leave, and the CFC will finally have success.

Hmmm, maybe Toyota could use this philosophy. Don't worry about all those customers selling all their Toyotas because we screwed up the accelerators. Just find some new customers to replace them! When those new customers encounter quality problems, they'll stop buying Toyotas too, but replace them as well, faster than they abandon ship!

The key IMO is to focus on minimizing quitters, while also recruiting new people. Give the quitters a reason to stay. Give them a prize once in a while even if they can't buy a winning streak. Show their good moves or their good strategic plans for everyone to see, make them feel they did something once in a blue moon.

I'm going to end this discussion with this: at some point in the future, I may come back to Canada (I'm in the U.S. right now) with enough money to start trying some of my ideas myself, not as an organizer, but as a sponsor to events with my ideas being incorporated. I've got plenty more ideas besides what I've written about here, and I'm still coming up with more, and I would prefer it to be Canada where I try them than the U.S., although maybe I should consider that the U.S. might be more open to them.

[Carl Bilodeau]

Re: What can Canada do to nurture its chess talents?

....

It's good that you remember the difference between exponential and linear, which most everybody has learned in high school.

....

Difference between exponetial and linear is not the goal of Operational Research. The goal is to recognize, in a problem, the type of algorithm, linear, exponential, etc,. required. Yes a doctor knows what a virus is (like you and me) but his science is to recognize one.

In a 10 minutes blitz the two computers will have less draw since they can not compute enough to reduce the probability of big mistakes. The same happen for humans so your example doesn't work.

If two Kasparov plays together very long game they will have more and more draws since it is the same system playing against himself. So two computer systems having the same CPU power should have more draws.

But take two computer systems completly differents like Fritz against Rybka and Rybka will win more often.

Take two Rybka computer systems, one with 300 cpu and another with 10000. The first one will loose since the stronger computer will bring him in lines that hide tricks like a GMI would do with you and me.

Carl

[Egidijus Zeromskis]

Re: What can Canada do to nurture its chess talents?

OFFTOP

It would be great if someone could actually try this experiment with the programs and document the increasing number of draws, and perhaps draw a pretty graph that would point to a time control at which the infinite draws should occur.

Similar things were calculated by Kirr at kasparovchess.crestbook.com (RUS) http://kasparovchess.crestbook.com/v...322836#p322836

Results

(0 = 2000)
(Average rating = (Ra+Rb)/2 )
(data point ~ 32 games match)

[Tom O'Donnell]

Re: What can Canada do to nurture its chess talents?

Paul,

1) The explosion of very strong teenagers certainly is due in large part to the computer revolution. I think mainly though for most of them it is because they can literally play thousands of games as kids against a wide spectrum of opponents, certainly a lot more than most players who developed before computers could ever have played. Sure, for the GMs it is important that they have access to opening books, DVDs, ChessBase, etc. but I think Kevin is right that for the majority of players (i.e. those 99%+ who are not GMs) should spend more time studying tactics and learning how to analyse and less time studying openings.

2) Brilliancy prizes are nice, but to use them as the method for awarding the majority of prizes in a tournament will literally lead to scandal. Since the Winter Olympics are about to start, think figure skating competitions. Those are judged, and almost every event there is some sort of scandal.

Also, you produce games where players are trying to discover what the judges are looking for, producing farcical results. I remember reading Joel Benjamin's book "American Grandmaster" and he talks about Isaac Kashdan awarding brilliancy prizes at Lone Pine. Now Kashdan was a very strong player, but he liked flashy sacrifices and in one game one of the players had a choice between two winning sacrifices. He chose to sacrifice his Queen because he was pretty sure it would get awarded the prize. That's just a silly way to play, imo.

Your comment about "...The brilliancy is awarded based not just on the specific line that was played, but on whether a particular move in what seems an even position is irrefutable even against best play..." is fantasy stuff, for the most part. If the position looks even it probably is close to even, no? How often to players really have positions where they think they are okay and suddenly they are lost? Not very often (i.e. closer to never), imo.

If you want to randomize the prizes in chess, randomize the prize giving structure. For example, have some sort of graph based on score and award players a certain number of tickets based on their score. Then have a random draw to award prizes. Theoretically in that case anyone who scored at least one draw would have a chance to win a prize. ;-)

[Paul Bonham]

Re: What can Canada do to nurture its chess talents?

Difference between exponetial and linear is not the goal of Operational Research. The goal is to recognize, in a problem, the type of algorithm, linear, exponential, etc,. required. Yes a doctor knows what a virus is (like you and me) but his science is to recognize one.

In a 10 minutes blitz the two computers will have less draw since they can not compute enough to reduce the probability of big mistakes. The same happen for humans so your example doesn't work.

If two Kasparov plays together very long game they will have more and more draws since it is the same system playing against himself. So two computer systems having the same CPU power should have more draws.

But take two computer systems completly differents like Fritz against Rybka and Rybka will win more often.

Take two Rybka computer systems, one with 300 cpu and another with 10000. The first one will loose since the stronger computer will bring him in lines that hide tricks like a GMI would do with you and me.

Carl

As usual, Carl, you read my post but didn't get it.

I didn't write that we would play two Kasparovs against each other or two Rybkas against each other, or that we would have them run on different CPUs of different speeds.

I wrote that we would play 8 DIFFERENT computer chess programs, preferably the top 8 in the world, against each other on identical hardware (say 8 separate desktop systems with quad core CPUs).

I did write that the same effect of increasing draws with increasing time controls would happen with humans, up to a certain point. You wrote that "the same happen for humans so your example doesn't work". HUH? HELLO! EARTH CALLING! Read my previous post again, you aren't understanding what I'm writing.

[Carl Bilodeau]

Re: What can Canada do to nurture its chess talents?

Paul,

1) The explosion of very strong teenagers certainly is due in large part to the computer revolution. I think mainly though for most of them it is because they can literally play thousands of games as kids against a wide spectrum of opponents, certainly a lot more than most players who developed before computers could ever have played. Sure, for the GMs it is important that they have access to opening books, DVDs, ChessBase, etc. but I think Kevin is right that for the majority of players (i.e. those 99%+ who are not GMs) should spend more time studying tactics and learning how to analyse and less time studying openings.

2) Brilliancy prizes are nice, but to use them as the method for awarding the majority of prizes in a tournament will literally lead to scandal. Since the Winter Olympics are about to start, think figure skating competitions. Those are judged, and almost every event there is some sort of scandal.

Also, you produce games where players are trying to discover what the judges are looking for, producing farcical results. I remember reading Joel Benjamin's book "American Grandmaster" and he talks about Isaac Kashdan awarding brilliancy prizes at Lone Pine. Now Kashdan was a very strong player, but he liked flashy sacrifices and in one game one of the players had a choice between two winning sacrifices. He chose to sacrifice his Queen because he was pretty sure it would get awarded the prize. That's just a silly way to play, imo.

Your comment about "...The brilliancy is awarded based not just on the specific line that was played, but on whether a particular move in what seems an even position is irrefutable even against best play..." is fantasy stuff, for the most part. If the position looks even it probably is close to even, no? How often to players really have positions where they think they are okay and suddenly they are lost? Not very often (i.e. closer to never), imo.

If you want to randomize the prizes in chess, randomize the prize giving structure. For example, have some sort of graph based on score and award players a certain number of tickets based on their score. Then have a random draw to award prizes. Theoretically in that case anyone who scored at least one draw would have a chance to win a prize. ;-)

You are right. Brillancy prizes would lead to scandals. For me the Nobel prize for Al Gore and Obama were more about Brillancy than actual measured results and both lead to scandals.

In a kid chess club, the organizer who is also a parent will more likely vote for his own kid for a Brillancy prize as I have seen in the past. If you have 4 parents in this situation in a club, chances are 80% in my opinion that the four kids of these parents will have a Brillancy prize. This is what I have seen.

Carl

[Carl Bilodeau]

Re: What can Canada do to nurture its chess talents?

OFFTOP



Similar things were calculated by Kirr at kasparovchess.crestbook.com (RUS) http://kasparovchess.crestbook.com/v...322836#p322836

Results

(0 = 2000)
(Average rating = (Ra+Rb)/2 )
(data point ~ 32 games match)

In my opinion the line you could also draw would rather be that way:

A rapid increase of draws in the beginning of the graph which would eventually stabilize.

[Egidijus Zeromskis]

Re: What can Canada do to nurture its chess talents?

In my opinion the line you could also draw would rather be that way:

A rapid increase of draws in the beginning of the graph which would eventually stabilize.

What is "a rapid increase" in your case? (in rating points).

[Kerry Liles]

Re: What can Canada do to nurture its chess talents?

In my opinion the line you could also draw would rather be that way:

A rapid increase of draws in the beginning of the graph which would eventually stabilize.

Hey, that looks like a hockey stick... wait, wrong thread... sorry.

[Paul Bonham]

Re: What can Canada do to nurture its chess talents?

Paul,

1) The explosion of very strong teenagers certainly is due in large part to the computer revolution. I think mainly though for most of them it is because they can literally play thousands of games as kids against a wide spectrum of opponents, certainly a lot more than most players who developed before computers could ever have played. Sure, for the GMs it is important that they have access to opening books, DVDs, ChessBase, etc. but I think Kevin is right that for the majority of players (i.e. those 99%+ who are not GMs) should spend more time studying tactics and learning how to analyse and less time studying openings.

Hi Tom, thanks for your input. I can only say that everything I'm suggesting is designed to get chess players spending more time on tactics and less time studying openings :) . The thousands of games the kids are playing are refining their tactical knowledge, yes, but at least as much (or even more so IMO, FWIW, which might be nothing), they are refining their opening lines. But regardless of which they're doing "more" of, they are certainly more knowledgeable of opening lines, especially their favorite ones, than any generation of kids before them. And so when Joe or Jill Newbie comes into a tournament for the first time and plays Timmy Wunderkid, it's all over in the opening, even if Timmy toys with them for 30 or 40 moves. And by the time the weekend is over, Joe or Jill has been beaten by at least a couple of little Timmys and maybe a little Sally to boot and skulks away never to be seen again. Of course, this is exaggeration to a point, some Joes or Jills will determine to study the game more and come back. But we don't want any of them to leave if we're talking about growing the base, so even if they lose in the opening, give them credit for something they might have done that was worth noting.

2) Brilliancy prizes are nice, but to use them as the method for awarding the majority of prizes in a tournament will literally lead to scandal. Since the Winter Olympics are about to start, think figure skating competitions. Those are judged, and almost every event there is some sort of scandal.

True, but what I recommended means the judges never know who they are judging, and we also have to make sure they were not at the event so that they don't recognize the game. Maybe they don't even get to see the opening in case that gives them a clue who was playing. It's a lot of work and logistics to maintain this, I agree. Will it be worth it? I happen to think so, but only trying it will give us a hint.

Also, you produce games where players are trying to discover what the judges are looking for, producing farcical results. I remember reading Joel Benjamin's book "American Grandmaster" and he talks about Isaac Kashdan awarding brilliancy prizes at Lone Pine. Now Kashdan was a very strong player, but he liked flashy sacrifices and in one game one of the players had a choice between two winning sacrifices. He chose to sacrifice his Queen because he was pretty sure it would get awarded the prize. That's just a silly way to play, imo.

I get your point, although that seems an extreme example. Would you say that in your OTB games, you never look for tactics? I'm sure you would say you almost always look for tactics first before anything else, n'est-ce pas? And if you don't see anything immediately tactical and winning, then you start looking for a way to create a tactical opportunity, say setting up a discovered check. If it's there, you would probably weigh it against any weakness in your own position that gets created as a result, and what counterplay might be available, and if it seems sound, you might then weigh it against more positional, long-term strategic type moves. So in a way, we might say that everybody has the same methodology or way of playing a move? The whole reason I would give both brilliancy prizes AND strategy prizes is so that any player could feel that they could still win a prize without forcing themselves to play a style that isn't fitting for them.

Again, I do see your point, some players might value brilliancy prizes more than strategy ones and say to hell with winning or losing, to hell with my rating, just go for a brilliancy right here, right now. I don't know that that would be a bad thing: if they did it often and their brilliancies turned out to be duds, they would either work to improve their tactical vision or they would start trying the more strategical route.

My concern is that just awarding pure wins doesn't seem to be working at expanding the chess base, so try something else.

Your comment about "...The brilliancy is awarded based not just on the specific line that was played, but on whether a particular move in what seems an even position is irrefutable even against best play..." is fantasy stuff, for the most part. If the position looks even it probably is close to even, no? How often to players really have positions where they think they are okay and suddenly they are lost? Not very often (i.e. closer to never), imo.

Agreed, true brilliancies will be rare. I guess I should qualify the term "irrefutable". I didn't mean it to be in the sense of "immediately winning", although it could be that. More usually, it will be in the sense of a move that begins to weaken the opponent's position by prompting a needed reply from the opponent, where all other replies would induce even greater weakness. The game doesn't necessarily end just a few moves after the brilliancy.

In fact, there could be a very fine line between brilliancy and strategy prizes, and in some cases, a move could be nominated for both prizes.

If you want to randomize the prizes in chess, randomize the prize giving structure. For example, have some sort of graph based on score and award players a certain number of tickets based on their score. Then have a random draw to award prizes. Theoretically in that case anyone who scored at least one draw would have a chance to win a prize. ;-)

Has that ever been tried? I don't know, it seems maybe TOO random. Winning top prize when you only achieved a single draw would probably be embarrassing to most people. Kind of like rubbing sugar in the wounds? haha!

But if you won top brilliancy prize, well, you got that because you DID something notable.


In conclusion, I agree that my ideas are a little out there, experimental. But as the saying goes, nothing ventured, nothing gained.

[Carl Bilodeau]

Re: What can Canada do to nurture its chess talents?

...
In conclusion, I agree that my ideas are a little out there, experimental. But as the saying goes, nothing ventured, nothing gained.

Experimental? You wrote it at 10:42 PM and with the strange "logics" of this, I feel there was probably a few beers involved in this experimental experience. :)

[Kevin Pacey]

Re: What can Canada do to nurture its chess talents?

Then please give us your explanation of why, at the same time that we have exploding growth of chess opening databases online, or on DVDs, easily updateable and browseable for anyone with even an old clunker of a PC, we are also seeing explosive growth of teenage chess superstars and even GMs? Kasparov, whatever you may think of him, got at least this one right. The young mind is most capable of learning and memorizing all this information, just as it is with learning new languages. Older minds are more hard-wired and can't rearrange themselves nearly as well.

I think Tom has given a response to this line of reasoning that I can concur with. Not sure you and Tom and I will ever agree on exactly how significant knowing lots of 'theory' is for class player/newbie level. Certainly they can make more work for themselves in learning it than is necessary.

Of course there will always be winners and losers, but when these two groups are very static, the losers go away, and a new bunch of losers must be recruited. We need to make the two groups more dynamic, not as dynamic as in poker, but more dynamic then they are now. My idea of using brilliancy and strategy prizes could do this, albeit yes, it would be an experiment. If it does go horribly wrong, should be easy to go back to the old system.

I can't claim to properly estimate the % of class players/newbies who quit, at least in Canada, because of becoming discouraged or due to thinking they don't know nearly enough theory (and not wanting to study theory, or chess as a whole, because they think it'll take too long). If the % is small, and more are quitting because they think the CFC doesn't offer enough services, or because of some other aspect(s) of [tournament] chess (e.g. not enough weekends free in busy modern/hard times to justify having a membership), than your argument would certainly have no weight.

The CFC has tried surveys of ex-members in the past (they've tried to minimize quitters, tried to reduce the annual turnover rate of members of ~33%, but never succeeded, nor has Bridge, and I wouldn't be surprised if even poker has such a problem, though the influx of newbies is far greater, I imagine). I have no idea what questions were asked, but maybe asking whether the ex-member had no time on most weekends would have been very critical, for example. One Ex-CFC prez told me that, at least in his time before the late part of the last decade, ex-members gave up not because of anything the CFC did (that may have changed later on). Like I said, I don't know if most ex-members admitted they quit because they were weak, disliked knowing theory, or any number of reasons. They might be reluctant to admit quitting due to discouragement too.

The trouble with experiments gone wrong is that you may not be able to undo the damage so easily, and certainly you have caused a major setback and lost revenue (and, at least temporarily, members of the Federation(s)).
When the need to experiment/try something new has not been clearly established, radical ideas are less attractive since the risk:reward ratio is not so favourable. Wait for FIDE to declare Chess960 to be the new standard due to chess being exhausted, for example, before having just Canada do so. Then the whole world is on the same page.

Hmmm, maybe Toyota could use this philosophy [of not worrying so much about regaining quitters]. Don't worry about all those customers selling all their Toyotas because we screwed up the accelerators. Just find some new customers to replace them! When those new customers encounter quality problems, they'll stop buying Toyotas too, but replace them as well, faster than they abandon ship!

If the CFC was competing for customers who could get a better product then this reasoning would be stronger. Remember also, most everyone needs a car, so an influx of new customers for Toyota is harder to come by than an influx from the public of people who haven't yet known the urge to get into organized chess events, perhaps because the CFC/local organizers/clubs simply don't advertise/membership drive. In addition, the CFC is not clearly offering a badly defective product, as I've reasoned above.

The key IMO is to focus on minimizing quitters, while also recruiting new people. Give the quitters a reason to stay. Give them a prize once in a while even if they can't buy a winning streak. Show their good moves or their good strategic plans for everyone to see, make them feel they did something once in a blue moon.

If you can get lots of potential quitters to stay as an organizer who offers these prizes, and can make it work successfully, fine. You may drive away may people who are used to place/class prizes.

The thing about quitters (and we're talking about those who hoped to take chess by storm, and learned it would take time, if ever) is I've played some I've known in offhand games. They refuse to take pawn or piece odds, out of pride. Tell them they might win a brilliance/strategy prize, they don't care since they want first and foremost to win the event, or as you put it, kick butt. In Reagan's words, for them, there is no substitute for victory :).