Records versus bots

[Marc Meakin]

FWIW I think records against bots should be separate from records against humans.
Rather like the Pro Ranks. (Pro Records maybe?)

[Kirk Bevins]

FWIW I think records against bots should be separate from records against humans.
Rather like the Pro Ranks. (Pro Records maybe?)

Any reason why?

[Marc Meakin]

FWIW I think records against bots should be separate from records against humans.
Rather like the Pro Ranks. (Pro Records maybe?)

Any reason why?

I just think that records created against fellow humans are more praiseworthy on the whole.
I also think that any half decent player can keep playing the bots until the right letter/number combinations show up in order to break the record.
EVEN I came close to holding the Omelette letters attack record.
Just my opinion.
Anyone else agree?

[Kirk Bevins]

I also think that any half decent player can keep playing the bots until the right letter/number combinations show up in order to break the record.

And why can't you do this against humans?

[Marc Meakin]

I also think that any half decent player can keep playing the bots until the right letter/number combinations show up in order to break the record.

And why can't you do this against humans?

I agree, but many players play one or two rounds against prune, fail and start again.
So it is 'easier' to achieve the top score against a bot.
Do you get more satisfaction breaking records or getting max games against bots or humans?

[Kirk Bevins]

Do you get more satisfaction breaking records or getting max games against bots or humans?

Breaking records? Depends what the record is. If I get a record by scoring 20 points in Russian then that's a bit shit. If I get a high score record in 15 rounder, then that's a bit different.

A max game is a max game, whether I'm playing Chris Davies or Apterous Shit - both maxes are equally fulfilling. I tend to put more effort in when playing humans and tend to use the ERE button too much against Prune.

[Charlie Reams]

I agree, but many players play one or two rounds against prune, fail and start again.

Except that this is now impossible, so your wish has been granted.

[Marc Meakin]

I agree, but many players play one or two rounds against prune, fail and start again.

Except that this is now impossible, so your wish has been granted.

Thanks, although a clear distinction between records against humans and records against bots would have been sufficient..
Yes I know, I'm an ungrateful cunt.

[Kirk Bevins]

Thanks, although a clear distinction between records against humans and records against bots would have been sufficient..
Yes I know, I'm an ungrateful cunt.

Why? Whether I play Apterous Prime or Innis Carson - who cares?

[Andrew Feist]

Thanks, although a clear distinction between records against humans and records against bots would have been sufficient..
Yes I know, I'm an ungrateful cunt.

Why? Whether I play Apterous Prime or Innis Carson - who cares?

I don't think anyone would care too much about the difference between Prime and Carson -- those are at least both games. Playing Prune is not a game, it's a test of patience, luck and/or skill, in some order. (This is the same issue that came up in the Game of the Week deal -- a max game vs. Prune would get a lot of ink, but a see-saw battle between two good players (but where both players necessarily missed a word here and there as the battle wore on) wouldn't.)

This is going to sound more sarcastic than it is supposed to, but I'm too tired to try and fix it: In the rarefied air of Kirk Bevins-land, where max games are expected, it's not going to matter much who you get that max game against, because you will by definition win. For the rest of us, playing Prune is rather a different affair than playing a person. I've played some games against Nude recently to work on the multi-talented leaderboard, but it feels really grubby to do so and if those scores disappeared I wouldn't be sad about it.

[Michael Wallace]

I think this might need splitting into a separate thread (hint hint), but some thoughts:

I really don't see why records against Prune are really any different to records against a human. It's as much a matter of "patience, luck and/or skill" as playing a person is, it's just you need slightly more luck (your opponent doesn't spot a blinder) and plenty more patience (because your chances of scoring in every round are necessarily reduced).

Indeed, where do you draw the line? What's the difference between playing Prune and playing someone so much worse for you that it has exactly the same effect? How is that inherently 'better' (less grubby?) than playing a bot?

I do kinda agree with Andrew on the GotW thing, but can appreciate it's a matter of perspective. I personally find exciting tussles between two players (of any ability) much more interesting than Yet Another Max Game (I also think that less good players putting in a particularly good performance for them is often overlooked). I can also appreciate that I'm presumably in a minority on this, because it's voted for by the users anyway, so you can't really blame the system either.

(edited to fix a capital letter, and then again)

[Charlie Reams]

I do kinda agree with Andrew on the GotW thing, but can appreciate it's a matter of perspective. I personally find exciting tussles between two players (of any ability) much more interesting than Yet Another Max Game (I also think that less good players putting in a particularly good performance for them is often overlooked). I can also appreciate that I'm presumably in a minority on this, because it's voted for by the users anyway, so you can't really blame the system either.

I've always said that GotW is whatever people make of it. That said, there are certain features of the site that make it easier to spot max games and high scores when they occur, whereas games which are notable for more subtle reasons are harder to flag automatically. Another factor, which I know affects me, is that voting in GotW requires reading through quite a lot of recaps (often 10 or more) and it's much easier to give a high score to a sea of bold than to follow the "story" of a game and decide on that basis. In either case I'm not really sure what can be done about it, although I'd welcome suggestions.

[Ian Volante]

For me, there's a large amount of difference in setting a high score against Prune and doing it against a person. I couldn't give a monkey's about beating a bot, even one of the good ones, whereas there's a whole extra edge in beating a real person, as you're pretty sure that they care too, and are trying their hardest.

It certainly holds more cachet for me that my standard format high score (or is it second highest now?) was obtained against one of the best octochamps in history.

[Michael Wallace]

For me, there's a large amount of difference in setting a high score against Prune and doing it against a person. I couldn't give a monkey's about beating a bot, even one of the good ones, whereas there's a whole extra edge in beating a real person, as you're pretty sure that they care too, and are trying their hardest.

So if you found out that Prune was a real person, who was just Really Shit at apterous, it's suddenly different?

[Ian Volante]

For me, there's a large amount of difference in setting a high score against Prune and doing it against a person. I couldn't give a monkey's about beating a bot, even one of the good ones, whereas there's a whole extra edge in beating a real person, as you're pretty sure that they care too, and are trying their hardest.

So if you found out that Prune was a real person, who was just Really Shit at apterous, it's suddenly different?

Well it's no fun beating a real person that's really shit as well, so no, but with the likes of Plum and Velvet, who are more my level, then if that was the case then yes.

[Michael Wallace]

Well it's no fun beating a real person that's really shit as well, so no, but with the likes of Plum and Velvet, who are more my level, then if that was the case then yes.

Ok, so setting a high score against Prune is the same as setting a high score against someone who was just that shit. So how do you draw the line, especially when you're implying that it depends on your own ability?

[Ian Volante]

Well it's no fun beating a real person that's really shit as well, so no, but with the likes of Plum and Velvet, who are more my level, then if that was the case then yes.

Ok, so setting a high score against Prune is the same as setting a high score against someone who was just that shit. So how do you draw the line, especially when you're implying that it depends on your own ability?

If it's not a challenge, it's not as much fun, simple as that. A lot of the challenge for me isn't just beating a score, it's beating the other person too. Obviously this area is as fuzzy as a brand new teddy bear, and I enjoy shooting at high scores too even if I am playing Prune.

Simply put, if I beat a record while playing a real person, I'll value it much more than if I beat a record while playing a bot.

[Michael Wallace]

The fun thing I can totally get behind - it's not as much fun beating a player (human or bot) who is really really shit, that's cool.

What I find hard to understand is the idea that a high score against a bot is inherently worth less than a high score against a human (which might not be what Ian is saying, but has been implied by others in this thread).

The only thing I can see is that it's 'better' because against a human you have a lower chance of actually managing it, but that seems a peculiar line to draw because the issue is clearly that it's a bot, not that it's incompetent. I don't think incompetence is the issue people are having - would anyone really subscribe to "records against rubbish people should be separate to records against competent people"?

[Jon Corby]

I guess another point is that when you're against a person and you get juicy letters - they're getting them too. So there's two people being given the chance to break a record. If bots can't hold the records (Rex would clean up) then maybe records shouldn't be achieved against them either - or, at least, be discernable from those achieved in 'proper' games.

[Jon O'Neill]

The fun thing I can totally get behind - it's not as much fun beating a player (human or bot) who is really really shit, that's cool.

What I find hard to understand is the idea that a high score against a bot is inherently worth less than a high score against a human (which might not be what Ian is saying, but has been implied by others in this thread).

The only thing I can see is that it's 'better' because against a human you have a lower chance of actually managing it, but that seems a peculiar line to draw because the issue is clearly that it's a bot, not that it's incompetent. I don't think incompetence is the issue people are having - would anyone really subscribe to "records against rubbish people should be separate to records against competent people"?

I think the key issue here is that aiming to break a record by repeatedly playing a bot is a bit like masturbation. You can jizz all night on your own but if you have sex and it lasts ten seconds, at least you've had sex with someone.

But what about hand jobs?

[Michael Wallace]

I think the key issue here is that aiming to break a record by repeatedly playing a bot is a bit like masturbation. You can jizz all night on your own but if you have sex and it lasts ten seconds, at least you've had sex with someone.

Apterous is sufficiently populated that you could have sex with different people all day, which amounts to the same thing, surely?

[Marc Meakin]

The fun thing I can totally get behind - it's not as much fun beating a player (human or bot) who is really really shit, that's cool.

What I find hard to understand is the idea that a high score against a bot is inherently worth less than a high score against a human (which might not be what Ian is saying, but has been implied by others in this thread).

The only thing I can see is that it's 'better' because against a human you have a lower chance of actually managing it, but that seems a peculiar line to draw because the issue is clearly that it's a bot, not that it's incompetent. I don't think incompetence is the issue people are having - would anyone really subscribe to "records against rubbish people should be separate to records against competent people"?

I think the key issue here is that aiming to break a record by repeatedly playing a bot is a bit like masturbation. You can jizz all night on your own but if you have sex and it lasts ten seconds, at least you've had sex with someone.

But what about hand jobs?

The equivalent of an Apterous hand job is when your human opponent lets you get the conundrum unopposed, in order for you to max the game

[Kirk Bevins]

What I find hard to understand is the idea that a high score against a bot is inherently worth less than a high score against a human (which might not be what Ian is saying, but has been implied by others in this thread).

I don't think Dale's 15-round record of 170 should stand as he was playing a human. If he was playing Rex he'd only have 156 and he wouldn't have the record. It works both ways.

[Marc Meakin]

What I find hard to understand is the idea that a high score against a bot is inherently worth less than a high score against a human (which might not be what Ian is saying, but has been implied by others in this thread).

I don't think Dale's 15-round record of 170 should stand as he was playing a human. If he was playing Rex he'd only have 156 and he wouldn't have the record. It works both ways.

Nobody gains records playing Rex, do they?

[Jon Corby]

Ultimately though all records are a bit woolly, in that so much is down to the luck of the letters. When the standard 15 round record gets broken, chances are it will be because of a huge max being available that occurs so rarely that most people never get the opportunity. It's all a bit meh really.

[Ryan Taylor]

Is it possible to get 11 nine's in one game? I don't know how many tiles there are of each letter in a pack so wondered if 11 nine's is achievable. Probably a stupid question to which the answer will be yes.

[Marc Meakin]

Is it possible to get 11 nine's in one game? I don't know how many tiles there are of each letter in a pack so wondered if 11 nine's is achievable. Probably a stupid question to which the answer will be yes.

Definitely. (if you include numbers rounds )
Well you certainly got a stupid answer.

[Jon Corby]

Is it possible to get 11 nine's in one game? I don't know how many tiles there are of each letter in a pack so wondered if 11 nine's is achievable. Probably a stupid question to which the answer will be yes.

Without even bothering to check the distribution it's undoubtedly gonna be a yes.

[Ryan Taylor]

Ok, was just nice to be confirmed.

[Jon Corby]

Ok, was just nice to be confirmed.

Well, it hasn't been really as I haven't bothered to check

[Marc Meakin]

Time for one of you boffins to create a scenario whereby 11 nines are acheivable. (with the correct letter distribution)

[Michael Wallace]

Time for one of you boffins to create a scenario whereby 11 nines are acheivable. (with the correct letter distribution)

There was a page somewhere that detailed arrangement of letters in the tiles so you could guarantee 9s regardless of what the pick was. I can't find it now, and can't remember how far you could get (although obviously not very), hopefully someone else will be able to point at it though.

[Jon Corby]

Time for one of you boffins to create a scenario whereby 11 nines are acheivable. (with the correct letter distribution)

Easy. I think actually we had a challenge once to set the decks to see how many rounds you could guarantee nines in, regardless of the vowel/consonant selections made by the players. I think some smartypants got it up to 5.

Damn you coonboy

[Jon O'Neill]

Time for one of you boffins to create a scenario whereby 11 nines are acheivable. (with the correct letter distribution)

Easy. I think actually we had a challenge once to set the decks to see how many rounds you could guarantee nines in, regardless of the vowel/consonant selections made by the players. I think some smartypants got it up to 5.

Damn you coonboy

I'm pretty sure Charlie did it so you could be guaranteed eleven 9s?

[Jon O'Neill]

11 nines

eleven 9s

[Jon Corby]

I'm pretty sure Charlie did it so you could be guaranteed eleven 9s?

Wow, that would be seriously impressive.

[Charlie Reams]

I only remember getting guaranteed nines up to 3 rounds, although 5 was possible if you assume people don't pick 5 vowels. I dunno where those pages have gone now, they were on the Countdown Corral and never got transferred to the wiki. Now I'm a bit more good at stuff, I might have another go and see if I can get it up. There were several games that people had devised where 11 nines came up (using the right letters), but they weren't guaranteed.

Back on topic, yes, high scores are basically a curiosity in most formats. Still it would have to be said that a quick glance through the current charts shows that they are pretty well correlated with the top players in those formats, so it's not completely vacuous. In some formats the max is fixed, which is fairer, but there's still a lot of variation in terms of round difficulty, obscurity of the words etc.

I'm not seeing any convincing argument that a record against a bot is inherently any different. Why does it matter how your opponent is generating their answers? If the bots were more "realistic", would it be okay? What if they were so realistic that you couldn't even tell they were bots? Would a record against a really crap human be different? Now there's an unfinished game cap, people can't just have endless restarts until they hit a bunch of lucky nines. And if you personally find high score hunting against bots tedious and unrewarding then that's fine, there are plenty of other things to do. I try to cater to all the different ways that people like to play.

[Matt Morrison]

The issue I had with bots vs. humans before was that bots were including in some inappropriate stats - for example "Shutouts" is completely wrong for bots, as people play Prune all the time. Charlie kindly and correctly excluded bots from "Shutouts" but as far as highscores go, yeah the distinction is fairly irrelevant.

[Ian Volante]

I might have another go and see if I can get it up

I do love rejection.

You do love an erection? Oh.

I resemble that remark.

I see an opportunity here.

[Ian Volante]

Would a separation of bot records lead to an upsurge in human v human matches in obscure formats?

[Jon Corby]

Now there's an unfinished game cap, people can't just have endless restarts until they hit a bunch of lucky nines.

Actually yeah, I guess you're right that this pretty much take care of my "objections" as they were. What I do wonder though is if the pre-cap state of affairs has left the high-scores dominated by ludicrous Human v Prune restart type games, which still makes the distinction worthwhile on this older data even if there's no (or significantly less of an) issue going forward. I have no idea as I haven't really logged onto apterous in months. I really should pop back in.

[Charlie Reams]

Would a separation of bot records lead to an upsurge in human v human matches in obscure formats?

Probably not, it's just too difficult to find an opponent for e.g. a Hebrew conundrum attack unless one of you is just a patzer, in which case you might as well just play Prune. Still it is nice to stimulate human v human competition, so if the "top joint score" page reappears any day (which would be human v human only) that might do it.

[Charlie Reams]

What I do wonder though is if the pre-cap state of affairs has left the high-scores dominated by ludicrous Human v Prune restart type games, which still makes the distinction worthwhile on this older data even if there's no (or significantly less of an) issue going forward.

I worried about this too, but I think it would be harsh to strip people of their records retrospectively when they were given no inclination at the time that a game against Prune wasn't a "proper" one. Looking at the major formats, a lot of current records were set against humans anyway: 15 rounder, 9 rounder, Letters Attack, Speed 9, Junior 15, Touchdown 15 etc. The only variant where playing Prune is a big boon (teehee) is Goatdown, where you get perfect picking in at least half the rounds. On the less played formats, a lot more of the records are against bots (not always Prune) and TBH I don't see that as a problem -- in fact, it's exactly what bots are good for.

[Jon Corby]

What I do wonder though is if the pre-cap state of affairs has left the high-scores dominated by ludicrous Human v Prune restart type games, which still makes the distinction worthwhile on this older data even if there's no (or significantly less of an) issue going forward.

I worried about this too, but I think it would be harsh to strip people of their records retrospectively when they were given no inclination at the time that a game against Prune wasn't a "proper" one. Looking at the major formats, a lot of current records were set against humans anyway: 15 rounder, 9 rounder, Letters Attack, Speed 9, Junior 15, Touchdown 15 etc. The only variant where playing Prune is a big boon (teehee) is Goatdown, where you get perfect picking in at least half the rounds. On the less played formats, a lot more of the records are against bots (not always Prune) and TBH I don't see that as a problem -- in fact, it's a good example of bots doing what they do best.

Then I can sleep well tonight.

[Ben Wilson]

If people are complaining that games against prune shouldn't count, the solution is obvious. KILL APTEROUS PRUNE.

Anyways I can't really see wher the problem is. You're playing the same game against prune as you would if you were playing, say, Bevins or someone like that. The fact that your opponent is slightly weaker shouldn't have an impact on what is ultimately a great individual performance- which is what the records pages are supposed to highlight, after all.

[Andrew Feist]

Now that I've had more sleep, I'll try to be more coherent about what I'm trying to say (not that anybody who disagreed before will agree with me now, but not the point): To me, playing against Prune is a lot like playing the duel which is a lot like playing Flat scoring -- it's all about what you can do b/c what's going on in the other chair is completely irrelevant to your score. There's a completely different mindset (at least for me, people who are competent can chime in in reply) involved -- playing Velvet is a game, playing Prune (or the duel) is a puzzle [can I 'solve' these nine/fifteen/twenty isolated rounds]. I fully acknowledge that means the games against Prune are a purer test of your abilities. And for standard games, as many people have said, it's a lot of luck just getting a game where 170 points are on tap. But getting 200 in a conundrum attack against Prune vs. getting 200 in a conundrum attack against, well, just about anybody, feels like two very different things to me.

[Charlie Reams]

But getting 200 in a conundrum attack against Prune vs. getting 200 in a conundrum attack against, well, just about anybody, feels like two very different things to me.

It is. Just not in a way that makes any difference to high scores.

[Kirk Bevins]

But getting 200 in a conundrum attack against Prune vs. getting 200 in a conundrum attack against, well, just about anybody, feels like two very different things to me.

It is. Just not in a way that makes any difference to high scores.

This is an interesting point in that I'm trying to break a record at the moment and I play Prune. I played a human and I thought "had you not buzzed in just before me then I'd have broken the record". Kinda annoying as against a human it's sometimes a reaction test but against Prune it's just a "can you get this in 30 seconds" test.

[JimBentley]

I don't think it's worth making a distinction at all. To break any of the hotly-contested records, you'll pretty much have to max the game (and have a high maximum available to start with), and if you're going to max the game, then who your opponent is doesn't matter at all.

If you wanted to be slightly facile about it, you could even say that it's harder playing against Prune as he offers no information. Rex on the other hand offers the most information (on the rounds he picks, you know that whatever he's declared is the max, which could help you choose between your safe 7 or dodgy 8, for instance).

[Matt Morrison]

Interesting point about Rex vs. Prune, Jim.

I was just having a think and I wonder if the more important distinction between human and bot games is for personal bests, rather than high scores.

[Charlie Reams]

I don't think it's worth making a distinction at all. To break any of the hotly-contested records, you'll pretty much have to max the game (and have a high maximum available to start with), and if you're going to max the game, then who your opponent is doesn't matter at all.

Apart from conundrums...

[Paul Howe]

Interesting thread. As mentioned, it would actually be more accurate to only allow records against Prune, as this removes the confounding effect of opponent skill level. As mentioned, the only real issue is with the goat formats, where prune does most of the hard work for you, and this does strike me as somewhat unfair (although I've certainly indulged in some goat/hypergoat pruning myself so won't complain too vociferously). As mentioned, a lot of the high scores are untouchable in a large percentage of games, so gunning for these against bots must be incredibly tedious. For some of the more obscure formats though (I'm talking about stuff like hypertouch, not ridiculous arcana like dyslexic wingdings hyper unlimited) the record is there to be had every or almost every game, should you play skillfully enough, and hunting the records is good fun.

I didn't start off with all those "as mentioned" btw, but half way through I realised I'd said naff all that was original and thought I'd better add them in.