No, they cannot be perfect villagers, because eventually they will unlock the gold trade. And that would mean, that if you would want to unlock the previous trades you would have to use gold for it (expensive). If you have paper trade as the last one. Then you can trade paper indefinitely, because it would not get locked (because it would be the last trade). The same situation is with iron and blacksmith - if you have the iron trade as the last one, you can use iron to unlock all of the previous trades. And iron is very cheap (especially with an iron golem farm). It would be much worse if you have for example the diamond trade as the last one - then you would have to use diamonds to unlock trades.
The situation doc is in, is not that bad, cause he could get a decent emerald trade as the last one (eg. Relatively cheap pickaxe trade) and it would work good enough then. Yes, that's rare, but shouldn't be absolutely unheard of. 0.266% is 1/376 (not one out of every 376 villagers, but one out of every 376 unlocked trades should be gold). By my calculations, by the time you've executed 260 trades, there is a 50% chance you should have unlocked a gold trade. So I suppose it is still possible to unlock one, but I'm fairly certain I executed several hundred trades with librarians in my own testing, and never unlocked a gold trade. Similarly, for Docm to get over 320 emeralds (at 24:30 in this episode), even ignoring all the trades with the same villagers in prior episodes, he has to already have several hundred distinct trade events.
I'm not saying that my testing proves beyond a shadow of a doubt that a gold trade on a librarian is impossible, the only way to definitively rule that out would be by examining the source code. But at the same time I hope you would concede that its entirely possible that the numbers on the wiki could be either incorrect or simply based on old code that has since changed. And 400 to 1000 trials, in this case, is a pretty statistically significant amount. By my calculations, with the wiki's probability of 0.26666%, after 400 trials you should see at least one gold trade 65.63% of the time. After 1000 trades, you should see at least one gold trade 93.08% of the time.
After 2000 trials, the chance of NOT seeing a gold trade would be less than 0.5%. So I understand your skepticism, but trust me that I haven't just dinked around with trading for a few minutes.
My original comment that 'I've done quite a bit of trade testing', is probably a gross understatement, check out my contributions to villager trading strategies in Minecraft Forum threads and. In addition to my most recent testing on this issue specifically, I just went back and opened up 5 separate save games from Minecraft 1.5 and 1.6, each of them with very large villager farms (70-100 villagers each, very actively traded).
Mindcrack Smp Season 4 Final World Updates
I didn't find a single librarian with an unlocked gold trade in any slot. That all to say: I wouldn't be surprised if Doc's 2 'nearly-perfect' villagers will never unlock a gold trade. I'd also love to see someone examine the current source code and see if the percentages given on the wiki are actually still correct.
Oh, i can easily concede that the info on the wiki is wrong. I have no problem with that:) and i dont doubt what you said. Just i still think the number in wiki can be correct.
And i too would not be surprised if the doc's villager will not open gold trades. About the source code - if i dont forget and will be able to find some time in the evening i will have a look at it. But that is a big if just one point. You are dealing with probabilities generated by a computer algorithm and with a pseudo-random generator - not with the actual randomness (usually there is no such thing as a really random numbers in computer calculations - it needs too much work and in most cases is not worth it). Sure, but aside from the fact that it is deterministic, any good pseudo-random number generator should have the same numeric distribution properties as a true random number generator.
Derana Dream Star Season 4 Final
And if a specific algorithm doesn't generate a value in one specific 0.26% about as often as every other arbitrary 0.26%, then its not a very good random number generator. I don't have any idea whether Mojang is using built-in Java functions for randomization, or whether they wrote their own Mersenne Twister routine or something, so I don't know if there is any reason to question their implementation. I don't recall that I've heard anyone raise questions on this point.