Hi all,
very interesting replies thus far. Slight warning: This will be a longer post 
Personally, I understand the idea behind this proposed change. My opinion is the following:
1. Sufficient publicity of the discussion / information of athletes
Let me first say that a sidethread in a forum that not too many people regularly read seems not the right place to start such a discussion. If changes are really intended, then as many people as possible must be heard on that to give their input. For very important topics, e-mails should be sent around at least to all current League players in all divisions, with a short summary and with the link to the Art of Memory forum thread. Sending bulk mails is not too hard, Mailchimp works nicely and is easy to use. The disadvantage of just posting it here, without any kind of further notification, is that this will heavily skew the discussion towards the opinions of those who are active here.
2. People are attached to what they know
I am sure many players like the format as it currently is, they have trained for it and invested time into that. And for some Words and Names are their better events, so the point may also have some emotional aspect (“something is being taken away”). Humans are like that. So, it is clear that these players will opt to keep things as they like it.
3. Where is the beef?
In general, if someone wants to change something and argues that a change would improve things, the onus should be on them to prove that with hard data.
a) The problem of comparability
The example of a 35 words average player competing against a 45 words average player has been given. The question that must be answered is whether, in a match between two Numbers or Cards or Images players with the same strength difference, the weaker player has a better chance to score. This cannot be answered on the basis of some “hunch” or by citing “evident structural differences”. Either it is the case or not. The main difficulty I see in that regard is how to determine what constitutes a comparable difference in strength in another discipline. The formula ratings do not seem apt for that because they are no pure Elo ratings, only taking matches into account, but they also count scores etc… And it is not too hard to inflate one’s score if one finds a weaker training partner against who you can always go full risk for good scores.
b) Comparability based on an example guess
That means we would need to make an educated guess. 45 to 35 words seems, to me, about the same as someone who can do 25s Numbers safely (>90% full accuracy) competing against someone who can do 40s Numbers safely (>90% accuracy). And in these cases, the 25s player will have a huge advantage. They can simply go for a 30-35s time and win almost every time because usually accuracy heavily decreases below the safe threshold, here for the 40s player.
c) The comparability problem for more closely matched opponents
Of course, 45 words to 35 words is an extreme example. What about closer differences, like a 45 words player against a 47 words player? In that case, there are of course enough variances in score that the 45 words average player will with her or his share of matches. The same will apply for International Names. Variance leads to wins for the weaker player.
So, we see that the 45 vs 35 Words example is not really typical. Due to the League format, it is much more common that people of comparable strength will compete against each other. But the interesting question is: How does the win/loss ratio of a game between two rather closely matched Words players (with a small advantage for one side) compare to the win/loss of a match between similarly closely matched Numbers or Cards players?
I feel that for Cards it might well be that even a small advantage in safe times, let us say of a safe 21s player vs a safe 24s player, is a huge advantage, surely comparable to 47 vs 45 words, probably even higher due to the fact that you can close gaps at Cards much more easily. A similar dynamic should apply for Images.
But the question is: What happens to the main argument for a change if it is shown that for a minor differences in strength, Cards or Images is an even better option than Words or Names? In fact, it might make that argument completely moot.
And even for Numbers, I do not see it as a given that a game between closely matched players might not yield a better winning percentage for the stronger one than the 47 words player might expect vs the 45 Words player. That is because, as mentioned, usually accuracy decreases rather rapidly after leaving one’s “safe score zone”. So, in a match between a safe 21s Numbers player vs a safe 24s Numbers player, the best strategy might be in fact for the 24s Numbers guy to go for 24s and not for sub 21s. And then it might well be that a win/loss ratio happens that is more favorable for the 21s Numbers player than for the 47 Words player.
d) There is no avoiding the nitty-gritty details
What I am saying: Changing something that profoundly, basically doing away and erasing existing, functional and well-loved events requires to really go into the nitty-gritty of the matter and unearth whether it is really true what your intuition tells you. Science is littered with disproven hypotheses that supposedly “made sense” (Earth as the center of the universe, anybody?) after somebody worked hard and took a closer look at all the complexities and how they interact. As far as I can see, that has not been done. I understand that some people feel it might be like that, but feeling alone does not seem sufficient.
4. An alternative proposal: Three different amounts of data
Apart from the above, let me make an alternative proposal (that I had also sent to Simon a few months ago, by the way). It tracks rather closely with what has already been proposed here by some posters:
Three different amounts of data for each discipline:
a) Amounts:
Words:
30
50
70
Words v2 (because more than 50 Words might prove tricky for some with a 4 min recall time):
30
40
50
International Names:
20
30
40
Names:
20
30
40
Numbers:
60
80
200
Numbers v2 (since some people say 80 is already being memorized fast enough):
80
120
200
Cards:
52
78 (1.5 decks)
104 (2 decks)
Images:
16
30
60
Images v2:
30
60
90
b) Recall time
It stays at 4 minutes for each choice. That means if you really want to look at 200 Numbers or two decks of cards, you will need to recall quickly. And spectators will see more recall action in the same amount of time, making it more interesting.
c) How to choose
The one who chooses the event also chooses the amount of data (“I choose Words 40”). The opponent might have one or two vetoes (as in the tiebreaker rules for the tournaments) regarding the amount (not the event itself), meaning they can say: “Veto, I choose Words 30.”. They can choose a higher or a lower one if they think that gives them an advantage.
d) No penalties
I feel the lack of penalties is a huge part of Memory League’s appeal, making it less stressful with less of a “tightrope walk” appeal than the classical events. You just enter your stuff and see what happens.
e) Advantages of that proposal:
For the spectators: They would be able to see three times the variety of matches and each amount of data would have its own appeal and its own strategical intrigue, with some players being stronger at shorter amounts than at longer ones. Basically, no two matches would ever be the same.
Also, it might be interesting whether someone saves their veto for exactly the right time to get the maximum impact. And, as mentioned, for the larger amounts spectators would have more visual activity on their screens, would get more of a wow effect (“a score of 163 in 1 minute - fantastic!”) and the action might also be a bit more frantic because people need to recall fast.
For the players: Yes, Memory League is cool. But it is also cool to expand your horizons, to try something new that does not dismiss what you already know well but that adds something to it. People would train the different amounts and some would surely find new favorite disciplines with the increased amount of variety and option.
For the commentators: More variety in the disciplines means more strategic depth of the matches. Matchups with shorter amounts than usual and with longer amounts might be equally spectacular, the former by way of raw speed and the race appeal, the latter by way of the huge mass of data memorized and recalled, to the wire.
For the ML organizers: The above changes would address the things @Simon has written, granting players the option to choose a smaller amount if they want that, so they can memorize all the data. It would also not take away beloved disciplines or unilaterally shift the balance between players without an in-depth and time-consuming examination of win/loss ratios (and the “comparability conundrum”) to back up whether that makes even sense.
Also, while the recall time would still be 4 min and not 3 min, it would be more action-filled minutes and thus better for the spectators. And if it is really wished by the ML team, of course reducing the recall time for the shorter amounts of data to 3 minutes is still an option, while leaving it at 4 minutes for the other events.
Thank you for reading, this was a long post. I tried to compress it as much as possible but there were many points I wanted to bring across. And I think when the discussion is about such an important topic, it is worth it to invest a bit more time to write something that really sums up one’s thoughts properly in order to convey them to others.
I hope that this proposal might find at least a few people with similar ideas. Personally, I would be thrilled to try it out while leaving the old system in place and then, based on experience, as @Konsti put it so nicely, people could decide whether they want a permanent change or not.
Huge thanks to @Simon, Issa and Josh for their work and to everybody in the community for making it what it is :).
