Gandalf's thread of numbers and probability

[Gandalf]

Two brief thoughts I've had recently.

1) Throwing 2 dice against blocks = not that risky

I always hesitate about doing a 1 dice block without any blocking skills - but it's actually got more chance of a turnover (1/3) than a 2-dice against block with someone with the block skill (11/36). I remember Rav doing quite a few 2d against blocks in games I've had with him and thinking he was getting lucky with them, but when you realise that less than 1 in 3 end in a turnover, it's not *that* crazy a move. If you don't have block or wrestle it becomes a 5/9 chance of a turnover, so yes that is a bit crazy :p

2) Re-roll skills vs easier dice rolls

What's better to have - a 1 in 9 chance of failure or 1 in 6? In certain situations, I would argue the latter! Let me explain...
Player A has agility 3 and sure hands. To pick the ball up means rolling 3+, but he can re-roll it. 1 in 9 chance of failure.
Player B has agility 4, but no sure hands. To pick the ball up means rolling 2+. 1 in 6 chance of failure.

HOWEVER, player B can use a team re-roll if he fails, thus turning it into a 1 in 36 chance of failure. Player A does not have this option available. Given that re-rolls are often saved for important things, like picking the ball up, I would rather player B on my team than player A, quite aside from the additional benefits that agility 4 bestows.

[Raveen]

And that kids is why Ag 4 is broken

In an ideal world I think the BB would be well served by having greater granularity than the D6 system allows. Even a D8 agility system would allow for a better spectrum of probabilities, but I'd probably choose a D10 system for easy working out

When it comes to risky blocks I tend to do the following:

Leave it until the end of the turn unless A turnover on your last action is no disaster.

If you need to do it in order to achieve something else (this always happens) try to make sure that the ball is safe before you throw the block so that success is a bonus not vital to the game.

Make sure that the blocker is someone you don't mind being on the floor for your next turn - don't do an uphill block with the corner of your cage.

[DaigaroOgami]

Percy has a question for thee 'Dalf....

How many matches have been won (in RTBBL history) by the coach who has had the worst dice in a match? (Basically implying that if you get the good dice, you win)

And so, which coach has had the worst Nuffle-bumming and has still managed to win? Is this easy to find out without going through each match in BBM?

This type of 'feature' could be something to add into you're new uploader thingy - the ability to look at match stats collectively - we could work out who is really is Nuffle's favourite son...

[lawastooshort]

As long as you also build in some kind of sample size/confidence interval calculator to show that there are actually enough numbers for this to mean anything and isn't just a fanciful pile of nonsense.


(No, it probably requires going through each match in BBM.)

[Gandalf]

Yeah.... this is more for the development forum.

I did start doing some work on all the matches from a few seasons back - getting all the chance scores and match outcome - but I didn't get that far. The numbers are in a spreadsheet somewhere. There did seem to be a strong link between who had the better D6 and who won though... but you had to account for a few things, like ignoring Khemri (who rolled so few D6), and some vague tier-ing of the teams, eg my Ogres having marginally better D6 but not winning did not count as a shock.

[DaigaroOgami]

I wouldn't even get into that deep for a general gist - just look at who had the better dice in a specific game and look at who won, it doesn't need all that other mumbo jumbo because it's not doing hard science to totally prove something.

[lawastooshort]

it's not doing hard science to totally prove something.

Oh well I'm totally sold then.

[Gandalf]

la you made me laugh

Daigaro - it's not really "hard science" to exclude outliers - be they small sample sizes or teams that don't compare well with others. It's just part of producing a fair result from research... which a key tenet behind this whole thread really.

[lawastooshort]

What would give a better quality of science I think is compiling a coach's entire dice rolls 1-6 / 2-12, rather than using the BBM style "chance score".

That way you could actually look at a Khemri team's dice by compiling the lot for a season. It might still only be say 8x20d6 compared to say a Vampire's 8x100d6, but it's better than nothing. Daigaro did something similar for the season 3 stats if I recall, by looking through BBM. The variations in luck between coaches over the season were tiny.


Regarding the tiers - you can make adjustments for that. I was messing about with a spreadsheet recently where I did this, I think to make a model for league position predicting.

[DaigaroOgami]

no no, I'm not on about doing that again; not analyse a team over the course of a season, just simply look at the BBM scores for each coach; then look at who won the match. Then see if the victor matches up with who had the greater result in the BBM scores.

If you guys want to say that the BBM scores are a waste of time then we may as well never include them in match reports - I only said to use the BBM scores because that is the only current reference point of anything to do with dice rolls during a match.

[DaigaroOgami]

it's not doing hard science to totally prove something.

Oh well I'm totally sold then.

Dude, I doubt we will ever do 'Hard Science' to prove anything in BB.

[lawastooshort]

There's been millions of games played, don't be so defeatist

[Raveen]

La likes it hard?

Sorry.



I'm not sorry.

[id3nt1ty]

Where as Daigaro likes it anyway way he can get it. Even if it's just a hint of it, to make himself feel better.

[Gandalf]

Just loaded my spreadsheet for season 4. I only have 20 games in there so not the largest sample size... but the person with the better D6 chance according to BBM won 13 times, drew 4 times and lost 3 times. And one of those three was temmop's first game with us.