First, it is easier to explain why peaches are more likely to have good fruit.
Peaches are commonly self-pollenated, so undesirable recessive genes have been weeded out. That mostly happened before humans started grafting them. But even then, there was selection for good growth from seeds, because many of us know, and some have always known, that we can get away with it.
And the best peach cultivars are not so many generations from when peaches were mostly grown from seeds. A lot of time, but not so many peach generations.
Apples are more self-incompatible, and therefore mostly crossing. So a recessive gene isn't weeded out of apples. Not even before grafting started.
If a gene is carried by 5 % of a population, and it is intercrossing at random in each generation, then 0.05 x 0.05 or 0.0025 or 0.25% of the offspring of the population will show the recessive gene. This isn't often enough to weeed it out.
Corn is like apples, in that a plant seldom pollinates itself. So undesirable recessive genes are common.
Now to why apple seedlings are generally inferior to their parents.
First, what you say about having to make hundreds of crosses to get a decent fruit is an exageration. You have to do that to get a superior fruit, comparable to the best on the market. Not better, just comparable. but as Dale Zinn, told me, he has yet to get a seedling that didn't help the cider. Dale lived 80 years on a farm in northeast Kansas. His apple orchard was a mix of antique varieties and seedlings. The average seedlling wasn't anything special. And his antique varieties were each something special.
What happens when you cross apples, or other plants, is called regression on the mean. The average of the seedlings will be something between the average of the parents, and the averge of the populations the parents came from. The average of the populations the parents came from is always lower than the best selections from the population. Otherwise, they would have selected some other vareity from that population.
So what to do? give up? Not me.
Some varieties give pretty good apples. Unselected Johnathan x Yellow Delicious seedlings used to be sold mail-order as cheap specials in garden catologs. They weren't as good as the named varieties, generally. But the customers were usually saticefied. They make a nice OK apple. Generally. And many newer varieties are from this cross.
I have read, and Dale agrees, that Antanovka seedlings make OK apples. And a variety named Alexander or Alexadria, or something like that is said to give pretty uniform seedlings.
Of course, if you are going for superior apples, you don't want to get seeds from a cross that gives uniform apples. Those have already been selected from.
I say, if you have the room, grow some seedlings. See for yourself.
And before you go investing a bunch of years in growing seedling apples, you might contact some profesional apple breeders and see if you could check out their breeding orchards at harvest time. See for yourself how a seedling population does.
Walter

Thanks, Walter. I am in fact growing some apple seedlings, about 40 right now just to see what I get. They are OP seeds from antique apples grown in an orchard of mostly antiques, so I should have a lot of variety in the crosses. I now understand how the recessives won't get weeded out and I sort of understand the regression from the mean. Suppose I lived long enough to do 50 generations of apple crosses, selecting say the 1% superior specimens each time to cross for the next generation. Would I have a better mean after 50 generations? I'm not quite sure what the mean is of. You also have a comment on peaches about how they are not so many generations from when they were grown from seeds. I guess the longer it has been since they were grown from seeds the more the mean has slipped due to the huge amount of selection applied to each generation?

What I am really after is to understand enough to plot an approach to breeding beyond one based on crossing two apples I like and selecting some good progeny. For example, if it turns out I could improve the mean in say 10 generations, I could probably squeeze that in before I die so maybe its worth trying. One other thing I don't see how fits is if wide crosses will be any better or worse than narrow crosses (by wide I mean less related). I decided to grow out antique crosses because they will on average be a lot wider than the standard apple breeding program today. Thinking about it, this may not improve the regression from the mean issue, but should give me a greater chance of getting something more unusual.

Scott

"Would I have a better mean after 50 generations?"

Yes. And may you live so long. This would be what is called recurrant selection. It is the most powerful breeding tool around. The mean of the population will shift some in every generation. The amount of the shift will depend on the ratio of the genetic variaton divided by the total variation. So if you can reduce non-genetic variation by growing the seedlings under the most nearly identical conditions possible, you will get best results.
The mean is an average. There are at least 3 ways to leagally get something called an average. One is the exact middle between the extremes, called the median by mathematicions. A second average is the number that is most common in a group. Mathemticians call this the mode. The mean is found by adding up all the numbers and dividing by the number of numbers. This is most often called the average, but it is important to be specific about which average one means, because they can be quite different.
What I said about peaches being closer to where they were usually grown from seeds, I now question some. Breeders who plan to rely opn grafting can make crosses with great variation in the offspring, knowing that any one excellent seedling can be grafted forever. But a farmer who is going to be growing from seeds will not usually make such wide crosses, because he wants the seedlings to be good uniform fruit of predictable quality.
For example, a farmer near where I now live grew an orchard of Elberta seedlings a century ago. They were said to be as good as Elbertas. This is important if you aren't going tograft them. But apples have been grafted since the time of the Roman Empire, at least. Peaches grown at the same time were seedlings. So apples were not selected to come true from seeds. Peaches were.
But the reason they were treated differently is that already in the wild, peaches were inbred and teneded to come true from seeds. Wild apples, like most domestic ones, are outcrossers. They would not come as true from seed from the beginning.
Also, there are many speices of apples that intercross, so there is much variation. There are not many species that cross redily with peaches. Almonds do, but the fruit is not useful.
As for trying to select for 10 generations, I am trying that too. Good luck to both of us.
Starting from intercrossed antique apples may be an excellent way to start. You are avoiding inbreeding depression. That is the deteriation of the aveage due to inbreeding. In apples, it is not as bad as some fruits, but still it is a factor.
One thing to think about is that precosious fruiting is something that yeilds to selection. And we have better chances of getting 10 generations if a generation is 5 years than if a generation is 8 years. And some varieties are even slower than that.
I read that the time from grafting to fruit is related to the time from seeds to fruit in the next generation. So we want to start with varieties that bear young. Beyond that, it is a matter of taste and the climate we breed for.
Wide vs. narrow crosses. The wider crosses will have more genetic diversity, so the the ratio genetic divided by total variation will be higher because the numerater (top number in a fraction) will be higher.
The narrow cross will be more predictible, as there will be less genetic variation.
And you might want to concider tree shape, etc. in your selection. Spur type vs. not spur type. Overall vigor is important. Disease resistance is nothing to sneer at.
Have fun with this.
Walter

The trick of grafting seedling wood onto a mature tree works with Prunus...I'm not sure if it does with apples or not.

Wow, thanks for more great information! Things are making a lot more sense. My breeding goal is taste, but not at the expense of a plant that is unusually bad in any other respect. I don't want to limit things too much by requiring X Y and Z other things, and even for taste there is no specific kind of taste, just something that is special. By not having that many other limits I will be able to be a bit more picky on selection for taste. Of course this is all in theory, right now I only have a couple tiny seedlings!

Re: grafting to speed things up, in a plant breeding book I have (popular one written by an English author, name escapes me now), he mentions that technique for all fruits. I asked about it at some point in the fruit and orchards forum and someone replied that the commercial fruit breeders don't use that method because of the work involved. But for home breeders and esp. if I aim for multiple generations, it seems like the way to go. I have a bunch of M9 rootstocks I was going to graft the apples to, but was thinking of getting M27's instead, they are the most precocious from what I have heard. What the book author says is it is possible to bud from a seedling even in the first year if you give it a good growing situation. If your rootstock then produces in only a couple years, thats 3-4 years between generations which is a lot better than 10.

Scott

PS to above: I just Googled on precociousness of apple rootstocks, and M9EMLA, what I have, is in the top group and M27 isn't. So I will stick with M9..

Apple trees, seem fun. Have you tried bonsai in or penjing, in that, it is my next task. Since bonsai seems to have a chance to alter genetic traits a bit, I don't know that it does, if it does it seems like it might lead to a new hybridization, eventually. I am pretty sure it has already been tried, but can't find any info, yet. Weaker version of same hybrid or cultivar, when bonsai is crossed, stronger or new? I guess eventually I will find out. Lol

Oh thanks for the info in this thread, encorages me to get a low grade microscope and text book on plant cloning to further enjoy my bonsai penjing hobby. Apologies for the interruption. Thanks again.

There is an apple variety named " Rescue " that can survive minus 75 degrees below zero Fahrenheit , There are giant apples ( over 2 pounds ) , Unusual tasting apples long keeping apples why not ad these to the mix ? Also if you graft you seedlings to existing grown trees you might get a generation a year . Do you plan to be alive 100 years from now ? what do you know that I don't?

There is an apple variety named " Rescue " that can survive minus 75 degrees below zero Fahrenheit , There are giant apples ( over 2 pounds ) , Unusual tasting apples long keeping apples why not ad these to the mix ? Also if you graft you seedlings to existing grown trees you might get a generation a year . Do you plan to be alive 100 years from now ? what do you know that I don't?

scottsmith perhaps you know something I don't?
but I have found B9 to be a better choice than M9
newer info about the dreaded FB favors B9
but it is as ever conflicting & therefore confusing!