Thinking ahead: NTBB 2013

[garion]

Hmmm, sounds questionable. Guess I will have to take your word for it though.Did they actually gain any real benefit from that money, like winning the league? Because the fact they have kept their TV at 3000TV for that length of time just means they aren't losing players. If they are losing skilled players and have the money to replace them with their TV would still drop significantly. a TV 3000Tv team suggests said team had a number of legends and most players skilled. So I don't think the bank would make any difference there. They are staying at a high TV because their players aren't being killed or injured enough to warrent firing not because bank rules arent there.

MAD are the only Chaos team to have won my league in 16 seasons. They played for the first 8 seasons of OCC before moving on to the BBATTL circuit, I believe (a home for retired teams which we have). This is their TV profile (green line) from the end of their 3rd season (which is when we started using BBManager, the tool which is allowing me to display this graph) to the end of their 8th, which is the point at which their treasury reached zero:

You can see their casualty list here. The "match number" in the casualty list doesn't relate to the match number for the team: it's the number of matches the player had played. To summarize, 45 SI (i.e. not BH) or worse in ~70 games, with 7 being deaths and another 14 being stat losses or niggles. I guess they were fairly lucky with the number of non-permanent injuries (getting far more MNGs than perms), and he did keep a lot of crippled players (Gluttony was at AV6 before taking 2 MNG and eventually dying), but still, they maintained over 2500TV for 40 or so games, and over 3000TV for at least 10. I don't recall the Orc team name but will try to dig it up.

You can dump cash by hiring and firing cheerleaders/asst coaches, I believe.

So as I said before, it would appear as though the fact they had money stockpiled made little to no difference. They stayed at a high TV because they suffered very few perms and or deaths. The bank rule would make no difference hear, also this team won 1 title. That doesn't suggest to me there is an issue here at all. If they maintained their TV over the course of 3 or 4 seasons and won the league each time I would concede there was cause for concern.


So I really don't see the problem with cash as it is in CRP in this respect. Why add a rule that punishes some teams horribly but doesn't actually make any significant positive to the rules set?

also on your second point hiring and firing cheerleaders is somewhat missing the point. That is still cash dumping. There was no need for this in lrb4 and before, because you could spend your money on star players, wizards etc... without it increasing your TR. now money is almost pointless, its only use is replacing injured players and nothing more.

[dines]

also on your second point hiring and firing cheerleaders is somewhat missing the point. That is still cash dumping. There was no need for this in lrb4 and before, because you could spend your money on star players, wizards etc... without it increasing your TR. now money is almost pointless, its only use is replacing injured players and nothing more.

Yeah like the good old FUMBBL days of sneak hiring wizards and/or stars. Always a pleasant surprise

[garion]

That's nothing to do with the bank, or the lrb4 rules, that was to do with how fumbbl handled the pre and post game sequences only.

[plasmoid]

Hi Garion,
you asked about advantage from the cash. Hard to prove in any way shape or form. But we did have a chaos team that reached 275TV, and won a lot of games while climbing there.
AFAIK, certain teams have different sweet spots. If your sweet spot is early, then climbing to high TV won't help you much. But some teams have a late sweet spot - the later the better in fact. If these teams are allowed to climb a lot higher than intended when CRP was designed, then these teams do get an advantage. I'm thinking Chaos, Nurgle and elfs here. Elven teams traditionally have a hard time getting there in the CPOMB-heavy environment, but with a nerf to kill-all-mens metagame, elfs will suffer less attrition.

As for this whole dumping cash, AFAIK BB has had cash both count and not count over the years.
Back in 3rd, cash counted, but you could freeboot stuff at half price.
I remember getting a TV advantage in crucial games by firing my coaching staff and freebooting the same for half the TV.
(In fact, we eventually changed the rules to let freebooters count for full TV - kind of a precursor to Petty Cash)

But in CRP, with JJs goal of not letting the experienced team dominate too much, any advantage from treasury was intentionally eliminated.

In fact, originally treasury was going to count towards TV. Myself and a guy called Gus fought hard against this, as this would mean that TV did not really represent the team playing the game, and I'm happy to say that eventually the rules were changed... So in a way this mess is mine . Bank tries to walk the line between 'cash counts' and 'cash doesn't count'.

Cheers
Martin

[garion]

oh don't get me wrong I do understand what the bank is for and I also thought it sounded like a good idea until I played with it enough.

I also agree with your that cash should never have counted towards TV in this rules set. It just would not make sense. So for CRP I agree with you, cash certainly shouldn't have.

If these teams are allowed to climb a lot higher than intended when CRP was designed, then these teams do get an advantage. I'm thinking Chaos, Nurgle and elfs here. Elven teams traditionally have a hard time getting there in the CPOMB-heavy environment, but with a nerf to kill-all-mens metagame, elfs will suffer less attrition.

I get what you mean but as you say elves tend not to get to such a high Tv anyway because of POMB and CPOMB, but I don't think there is any real long term advantage of having money in a league. Dode has found 1 example of a team that had 'more money than they should have' that won the league, and he might have another, this is in godness knows how many seasons (at least 15 I'm guessing???) that team didn't get to and stay at such a high TV just because of money they did so because they didn't lose as many players as you would normally expect. and the bank would make no difference here. They also only won the season once, which doesnt suggest anything is wrong. I haven't found any examples of teams that have won or stayed at too high a Tv because of cash hoarding in 4 fumbbl leagues. After looking a little harder i did find a couple of nurgle teams that had more money than you might be comfortable with at a TV around 2000 but they haven't won a single tourney yet anyway.

So it all begs the question why bother with the rule when you are trying to fix something that isn't broken. Especially when it punishes teams like vampires or ogres.

I guess it is all pretty moot now anyway though as the bank is locked in now, but I just don't see the point and it leads to silly cash dumping.

Also I would be wary of lowering elf attrition too much they have never had it better than they do now and as your goal is to balance the tiers I think you really need to loo at dark elves which for my money are the best perpetual league team in the game now.


As a side, the only place I think bank rules could help is actually match maker environments. But that's not what the rules are written for.

[dode74]

Dode has found 1 example of a team that had 'more money than they should have' that won the league, and he might have another, this is in godness knows how many seasons (at least 15 I'm guessing???) that team didn't get to and stay at such a high TV just because of money they did so because they didn't lose as many players as you would normally expect.

I think this understates things a tad without the context of the league. MAD played 8 seasons, the first two were 13 games and the rest 9 iirc. In the seasons where there were 9 games (i.e. up to now) per season the structure was a single Division 1 of 10 teams, where 2 stayed up and 8 were relegated to 4 Tier 2 divisions, in which the top 2 went up and the bottom 3 went down; there were from 5-6 tiers of a similar structure to Tier 2, so around 200 coaches. As I said before, we only started using the tracking tool at the end of season 3, so I don't know how they did in those early seasons, but they were top in S4, meaning they were at least in the top 2 of their division (assuming it was T2) in S3. In S5 they were 4th in the top division (the first season Flix, a rather excellent player, reached tier 1 in OCC, and he won and kept the title for 5 seasons). In S6 MAD won their Tier 2 division undefeated (8-1-0) and in S7 they were 5th in Div 1. S8 they retired coming 5th, but were planned to retire for the whole season. MAD stayed in the top 2 tiers for the 5 seasons we have records for, the top tier itself for 3 of those. Overall they are ranked 10th in OCC in pure Elo terms, with all the other teams in the top 15 having at least 5 seasons on them.

That aside, Juriel has given you at least one example of him using cash to offset the power of an opponent's inducements, and I know I have done it with the Orcs I played for 6 seasons, taking a wizard out of cash.

Also I would be wary of lowering elf attrition too much they have never had it better than they do now and as your goal is to balance the tiers I think you really need to loo at dark elves which for my money are the best perpetual league team in the game now.

Data I have from OCC puts HE top, then Necro (although Juriel's mutant wolves helped that cause a lot), then DE, those being the only 3 teams with a win% >55, although that is based on straight win% with no margins of error applied (simply not done it yet), and DE are only just above 55%.

[Hitonagashi]

Data I have from OCC puts HE top, then Necro (although Juriel's mutant wolves helped that cause a lot), then DE, those being the only 3 teams with a win% >55, although that is based on straight win% with no margins of error applied (simply not done it yet), and DE are only just above 55%.

Is that normalized for games played? In the Box, HE are one of the most rare races, and only tend to be played by good players, whereas every man and his dog seem to have a DE team.

[dode74]

From http://bbm.islanube.com/teams.php

Bear in mind that only 9 races were available for the first few seasons.

[garion]

none of that info suggests to me for a second there is any real advantage gained by not having to dump cash.

and this is the problem with only looking at data. Dark Elves are clearly better than High Elves but the data says otherwise. They are far more TV efficient, have a better starter roster and are far easier to re-build after a bashing than high elves. don't get me wrong high elves are a top tier side and very good, but Dark Elves are better all round.

Look at how poor Skaven, wood elves and Dwarves are performing too, 3 of the best races in the game.

As Hito eludes to, the newb attraction factor is something that needs to be taken in to account really.

[dode74]

none of that info suggests to me for a second there is any real advantage gained by not having to dump cash.

My experience is otherwise, as is Juriel's. YMMV.

and this is the problem with only looking at data. Dark Elves are clearly better than High Elves but the data says otherwise. They are far more TV efficient, have a better starter roster and are far easier to re-build after a bashing than high elves. don't get me wrong high elves are a top tier side and very good, but Dark Elves are better all round.

With all due respect that's nothing more than your subjective opinion, with all the biases which come with it. Just so we're clear, I'd say the same of anyone who said "X is clearly better" with nothing but their own experience to back it up - it just means that you find DE better.
The data doesn't say that DE or HE are better - like I said, the margins of error are not considered, and a quick look at those numbers tells me that there will be an overlap without even calculating it. If there is an overlap then you can't really say which is better. Furthermore, the data is contextual: it's relevant to the Cyanide game in a league of the structure I described earlier. It may have relevance to other leagues, and probably has more relevance to leagues than it does to MM, but has the limitations which come with the Cyanide version: mainly the lack of stars, several seasons of not having most of the races, and still some races missing.
OCC does take the "noob factor" into account to an extent. There is a "prep" league where noobs get a season (9 games) to play before moving into the main leagues. After 9 games (actually 9 weeks, so likely more than 9 games due to MM) most people will at least have the basics down. That said, there is always the possibility that some races have a lot of good coaches in there - there are methods by which this could be factored or measured, but we don't have the data to hand. As I say, though, there are no complete noobs.

For a little clarity on what those numbers represent, I've added in the 3 "new" teams and I've calculated the margins of error (95CI) and charted it in the same way as stocks - the green mark is the calculated win% for that race within OCC, and the black line represents the margins of error if we assume that OCC is representative of league blood bowl on the whole (getting data from several leagues would be preferable in order to minimise individual league biases); the longer the line, the wider the margins of error due to fewer matches played by that race (Khorne have 32 games played, for example, so their data from OCC is near-useless for this purpose). If you draw a horizontal line from the Y axis, every team it crosses could actually have the same win%, because the real win% could be anywhere on the black horizontal line for that team. If nothing else it illustrates that, even with 12,000 games-worth of data, simply citing win percentages and saying that A is better than B is misleading.

Look at how poor Skaven, wood elves and Dwarves are performing too, 3 of the best races in the game.

Again, highly subjective, but hopefully the above has clarified a little on this.

[Shteve0]

Right, agreed - I've often heard it said that Orcs have a deflated record as a result of being attractive to new players (being in the box is obviously a factor too).

Then again, throwing all data out of the window and working from gut feeling doesn't strike me as a particularly holistic approach either. There must be a measure for less developed (low TV) and highly developed (even if that means TV trimming for certain teams) that gives an indication of who or what needs short term nerfs/boosts and/or long term nerfs/boosts, and what specific level of tier normalisation you're aiming for, otherwise the whole exercise is (arguably) flawed.

Edit: posted in response to garion's last message

[dode74]

Then again, throwing all data out of the window and working from gut feeling doesn't strike me as a particularly holistic approach either. There must be a measure for less developed (low TV) and highly developed (even if that means TV trimming for certain teams) that gives an indication of who or what needs short term nerfs/boosts and/or long term nerfs/boosts, and what specific level of tier normalisation you're aiming for, otherwise the whole exercise is (arguably) flawed.

The data you use will be dependent on the environment which you are trying to measure. Using tournament data to measure league performance doesn't seem massively sensible to me.

[Shteve0]

For initial league rosters it's not too bad. It's undoubtedly better than using mid-high TV records to assess rookie and early devopment track records.

Either way, I think that "it depends what you're trying to achieve" reinforces my point about explicit clarity of desired outcomes being a sensible precursor to action and testing.

[dode74]

To an extent, although there are obvious biases: people play differently in resurrection formats, skills are assigned instead of falling where they may etc.

Is the aim level win percentages across TV ranges? If so then the assessment of low/mid/high performances in the relevant environment would be useful. If not then the overall performance is what matters (since it is automatically weighted for the TV at which most games take place).

[Shteve0]

Is the aim level win percentages across TV ranges?

Very good question. My assumption had always been (a partial) yes - look at the amazon treatment, where the changes were intended to give a "short term nerf, long term buff" (even if I strongly disagree that it achieves either), with the implied recognition of relative strengths at different TV and development levels (since the existence of team sweet spots means the two are not necessarily the same thing).