The Great Roster Cull

[burgun824]

Just how much better do you think these teams are at low or high TV? Is this based off of data and if so from where?

2) Why the rudeness? I don't mind for a moment my opinions being challenged, but I should appreciate a little more courtesy in your tone

That's just dode74. He's not being rude, he's just....analytically minded.

I picture him to look something like this:

[Digger Goreman]

Disclaimer revisited: Ain't none of this going to happen... ever... so get yer panties out from yer cracks and chill out.... If you do want to say something intelligent by way of constructive criticism... sure, we can talk... but if you only want to hear yerself talkin an anti-position... there are plenty of walls and mirrors in the world....

Happy-happy... joy-joy....

Bless ya Sir, made me smile!...don't touch my Flings!!

You're too kind, Pug Seriously, though, as much as I despise Ag4, I would actually like to see a pair on the Fling Team... makes them a real threat in practice (one season the great and powerful Aahz played Flings)... though three (yes... *palm to forehead*... Aahz rolled three Ag upgrades) makes them AMAZINGLY powerful!!!

[dode74]

That's just dode74. He's not being rude, he's just....analytically minded.

Hey! I apologised and everything!

[Shteve0]

Sorry, I simply meant that this is a thread for just chucking out your irrational ideas and pet hates, not talking sense. And no apology needed, please accept my own if I was a little thin-skinned there.

doubleskulls' site is at naf.doubleskulls.net (or even naf.talkfantasyfootball.org). It's based on naf tournaments worldwide and offers up some interesting data, with Humans somewhere on a par with Vampires and Undead/Wood Elves very much high performance teams at low TV - 58% rating in one case. (I'm leaving out zons because I don't think I've ever met anyone who doesn't want to see them re-written. Maybe sundevil?).

The black box data is more troublesome, partly because of the filters and partly because of the nature of the format. For example, Pro Elf teams were involved in only 22 games between TVs of 2200-2400 compared to Chaos' 459. Nevertheless, here's a misleading and generally nonsensical trend snapshot, broken into groups of teams with low natural stats and demarcation in skill access; teams with high attributes with good attrition but few starting skills; and teams with high attributes and decent starting skills but low attrition:

Code: Select all

	        1000	 1200	 1400	Average	  2200	 2400	Average	  Change
Undead	  56.58	56.51	56.68	56.59		41.93	28.57	35.25		-21.34
Human	   48.35	50.62	49.79	49.59		48.98	52.56	50.77		+ 1.18
Skaven	  53.35	53.24	53.41	53.33		54.36	50.00	52.18		- 1.15
										
Chaos	   51.69	47.82	46.49	48.67		53.91	52.84	53.38		+ 4.71
Dark Elf	51.53	49.07	50.73	50.44		63.03	59.02	61.03		+10.58
High Elf	44.80	44.38	46.47	45.22		57.47	47.92	52.70		+ 7.48
										
Wood Elf	54.16	49.77	50.20	51.38		64.49	56.58	60.54		+ 9.16
Pro Elf	 46.94	48.58	48.24	47.92		59.88	59.09	59.49		+11.57

Undead are still ridiculous at low TV and shocking at high TV, humans and skaven remain relatively stable, and all the Elves and Chaos improve markedly over time. I'd suggest that all the top teams at this level need either high natural stats or exceptional skill access (GA linemen, for example, or S access spam). The figures, such that they are, appear to support that argument.

[dode74]

Given the dataset sizes and margins of error involved, and assuming that all we are really aiming for is that teams be within the Tier 1 range, they all look pretty good to me. Still, that is a matter of interpretation.

I do agree with your final statement, that top teams at high TV have "either high natural stats or exceptional skill access". It's something I've been discussing with Nelphine over at FUMBBL. This is totally off-topic, so feel free to ignore it, but one solution we agreed might work was to introduce a need to "open" skill categories for 20TV or on a double. For example, a GA access player who chose block as skill 1 would only be able to choose Dodge either on a double (should he wish to forego the other skill options) for 30TV or on a single for 40TV. Subsequent A or G skills would be at the normal cost. Initial skilling would be irrelevant to the scheme (e.g. a DE blitzer who chose dodge first would skill have to "open" the General category if he wanted to use it). The effect would be to increase the TV of those teams with good skill access - a Block/CPOMBer would be between 20TV (2 doubles) and 40TV more expensive, blodgers would be 10-20TV more etc - particularly when they start getting multiple linemen with multiple access. There might be issues with recording which were doubles or not, but I don't see that as any more difficult than recording how many SPP a player has.
Just a thought, so feel free to ignore me

[Digger Goreman]

Without controlling most variables and listing pages of assumptions, numbers are merely "interesting"....

"Beware of Geeks bearing statistics!" <not directed at anyone in particular....

[Shteve0]

The effect would be to increase the TV of those teams with good skill access

The better way to increase the TV of teams with good skill access is to cost it properly in the first place. Some position costs on some teams recognise this (human and amazon blitzers) some don't (orc blitzers). The common argument is that human blitzers are overpriced; I'd say the opposite, that orc blitzers are underpriced. When you get into Chaos (in particular) being able to field 16 S-access players at a time, it's problematic.

I'd favour a review of rosters, strategies and designs myself. I agree with a previous poster that some of the more samey rosters play subtly differently, but is 'just different enough' different enough when there's so much design space left to be explored? Surely those subtleties would be better discovered through individual teams' development paths than by rote.

[dode74]

Increasing cost of a player for skill access charges people for potential. That has an effect on low-TV and tournament games, which aren't the issue. Giving people the potential for free, as they are now, but charging them for realising it when they take the skills, means that TV will be a better representation of on-pitch capability. There is a long term difference between a 6338G player and a 6338GA player, but that difference is not really in effect on-pitch until the latter takes a skill from each category.

[Shteve0]

Whether you're aware of it or not, that potential is costed into certain players already, and (I would suggest) rightly so.

If you were to construct pricing for teams only for a format where skill options are not considered and teams regenerate endlessly, then yes, you have a point. But I'd argue that since development options are a consideration in all three formats (fixed league, perpetual league, reset tourneys), then it should be priced in. Both long term (perpetual) and fixed term (1-3 season) leagues are potential-matters environments, and the majority of tournaments I see these days have an allotment of skills based on Normals/Doubles choices, and S-access (and P-access in high RR cost teams) players are consequently at a considerable premium.

I also honestly feel that we would improve this game of ours if more teams were pitched at the level of Humans and some of the more recent additions - Slann, Underworld, the Khemri reboot and, yes, even your Khorne roster (though for the purposes of my ALT3RD rules experiment, I'm going to be labelling them "Norse") - but that may be a separate rant entirely.

[dode74]

This guide to creating teams, written by Galak, says that they're not costed in at all. What makes you say that they are?

I'm not suggesting that you shouldn't pay to have the option, but that you should pay to take it. There's a world of difference: potential paid for when realised as opposed to taking an up-front penalty. What extra, on-pitch, does a rookie 6338GA player bring over a rookie 6338G player? Assuming that you're going to say "nothing" then why should the GA player cost more for that single game?

[Shteve0]

Common sense. If they're not costed in, then I've been crediting the latter output of the rules committee with common sense they've apparently not deserved. And you too, by the sounds of it. If the entire pricing structure of the teams is truly based on planning for the teams playing only one single game - ONE GAME - as opposed to any format in which the game is actually - you know - played, then it's astonishing we've not ended up with a total car crash of a rule set.

3) Any player's final calculated price may be modified by 10k more or less from the calculated
price to achieve overall team balance. This is pretty important to balancing a roster in the long run.

If this is not a tool for controlling the relationship between potential and price, I don't know what is.

The comparison of two players in the same game is pretty feeble too, to be honest. Is a 6338 Frenzy player worth more than 6337 Block player? It wholly depends on the team context.

Edit: I wish that "written by Galak" or "approved by Galak" stamp wasn't thrown around in the way it is. I'm sure it's a very nice piece of work, and I like and respect Tom both for what he's contributed and the way he handles himself on here and in business, but I don't think it's appropriate to represent him as some sort of god-like, unquestionable force of nature. I've never seen him present his credentials in this manner, it seriously annoys me that so many others do. Learn to judge ideas and concepts on their own merit, please.

[dode74]

"Common sense" is hardly a standard: in this instance it merely means it's what you would do. Given that the metagame is based around the single game, and that vastly more games are played at low TV than at high in almost all environments, then it makes as much "sense" that on-pitch value be represented by the cost of the players at the rookie stage as anything. Now I'm not saying that guide is what has been used, but you can see that almost all the teams (bar Lizards, iirc) fit within the template set there.

I don't see that guideline you quoted as a tool for controlling the relationship between potential and price, but as a tool for balancing teams overall.

Edit: I wish that "written by Galak" or "approved by Galak" stamp wasn't thrown around in the way it is. I'm sure it's a very nice piece of work, and I like and respect Tom both for what he's contributed and the way he handles himself on here and in business, but I don't think it's appropriate to represent him as some sort of god-like, unquestionable force of nature. I've never seen him present his credentials in this manner, it seriously annoys me that so many others do. Learn to judge ideas and concepts on their own merit, please.

I didn't offer any opinion on what I thought of it as a piece of work. I used "written by Galak" as evidence of the provenance of the piece of work, nothing more nor less. Don't mix up the issues, please.
If you do want my opinion on it then I'll happily give it, but what we were actually talking about was whether skill access was accounted for within current roster creation: I provided evidence that it may well not be, which you took to be some sort of opinion on the item I linked.

Going back to "judging ideas on their own merits", I personally think that skill access should not be a penalty paid up front (and it appears not to be), but giving good access for free is a long-term error. Hence my suggestion above.

[Shteve0]

Right, okay, sorry - I'll wind my neck in a bit.

What that link (a link to one of your posts, btw, hence the leap) apparently demonstrates is the costing applied, but not the logic or rationale used to generate those costings. Rosters are not intractibly linked to a one off game, but form part of the league rules. As players we regularly decide to invest in brittle or (perceived) overcosted players precisely because of their development potential and their context within the team, not because it is perceived better or worse value than the players on an opposing team.

If you ask me, the guidelines also seem to have been developed - rather conveniently - at least partially in retrospect and in order to explain the common ground between prior design decisions. Pact teams break them, too. An equally applicable reverse engineered model would seem to involve mainstream skills at 10k and limited use skills on BOGOF, with G, A and S skill access costed at 10k, M and P access at 5k each and MA and AV bound in a -/+ 10k relationship. Obviously there are loads of teams or players that break these rules too, but the +/-10k rule will fix most and one could make all the others named exceptions, in this same way this list has... and this is much the model I believed had been applied in the first place.

Edit: even the skaven lineman breaks those guidelines you linked to without a 'clause 3' style intervention

[dode74]

It's only a link to one of my posts because I moderate there and had a lot of backed up stuff when the forums were deleted a year or two ago, so I reposted the original. I get why it might be taken that way, I just wanted to be clear that it wasn't my intent to say "Galak said, therefore it's right".

"Clause 3" is part of the rules. The notes at the bottom talk about the teams not possible to create using those guidelines:

Dwarf, Goblin, Chaos Pact and Lizardman teams. The Goblin and Dwarf teams because they have weapons on the roster, Chaos Pact because they have too many slots (7) and the Lizardman team because it has 70k of team discount (10k discount each on the 0-6 Saurus and the 0-1 Kroxigor).

As players we regularly decide to invest in brittle or (perceived) overcosted players precisely because of their development potential and their context within the team, not because it is perceived better or worse value than the players on an opposing team.

That may be your thinking, but I'm sure it's not necessarily the thinking of everyone. Keeping our discussion to tier 1 teams only, "brittle" players tend to be either fast, or agile, or both, or have good support. Those brittle players are better at (i.e. have better odds of) scoring in less time, in general. The tradeoff, for me, looks more like whether to try to play a ball game with those fast, agile teams, or try to play a pitch control (either through positioning or cas) game with the slower but more sturdy ones.

I believe those guidelines were developed after a good number of the original teams were created, (I'm not sure when they were first used) but the fact remains that the guidelines can be used to make 20 of the teams (bar those above), and only one of those teams which cannot be made is due to calculation of TV. Despite the fact that there are, I'm sure, other guidelines which could be used to create all the rosters, the fact is that these are the ones we have been presented with, which suggests that skill access was not used in calculation of prices.

All that is by the by, really. The main point we're discussing is whether skill access should be an up-front cost or not. I agree it should be paid for eventually (hence my suggestion above), but can you give me a logical reason why an unskilled 6338G player should cost less than an unskilled 6338GSPM player? What is the on-pitch difference between a rookie Pact marauder and a rookie Human lineman, and why should the Pact player pay more for it?
The "potential" argument doesn't hold water, for me. A single game needs to be "balanced" as well as possible as it is the building block on which the environment in which we play is built. As the environment develops and the higher potential player does too then he should pay more for his greater options, either in terms of cash or in terms of TV.

[Darkson]

The whole balance of the game, including inducements, is meant to be about the game, not looking at the future - that's why there are no inducements that affect the team beyond the game (like bonus MVPs or extra winnings).

For what it's worth, I also disagree with making players pay for Skill access. Making them pay more in the long run, once they've taken skills, sure, but not on rookie players.
That's not to say that I think Dode's way is right, but if one of the two options discussed here had to be used, I'd have less issues with Dode's version than yours.