More blocks in tie games: why is this?

[mattgslater]

I'll be the first to admit that I'm a chronic overblocker. I love to lay the smack; the joy of knocking things around is a great part of why I play BB. One thing I've noticed is that I tend to throw more blocks in tie games than in wins or losses. In fact, on a per-turn basis (factoring out concessions and riots), this is the trend I've noticed.

Wins: I throw 1.5x as many blocks as my opponents, averaging a little over 3/turn.
Ties: I throw 1.2x as many blocks as my opponents, averaging about 4/turn.
Losses: I throw 1.3x as many blocks as my opponents, averaging about 3/turn.

Other things to note: If I throw 2x as many blocks as my opponent I usually win big. If my opponent throws more than 50 blocks, or I throw more than 70 blocks, the game is usually close.

Why is this? My opponents throw a ton more blocks than normal and I throw a ton more than that ... in ties only.

About 40% of my games are with Orcs, 25% Humans, 15% High Elves, 10% Undead, 10% other bashy teams (Amazon, Chaos, a few games each with CD and Khemri). Is this a function of my bashy balance? If so, we're talking about trends with bash teams. (Humans are a hammer in my hands: if I want finesse, I'll play the High Elves).

Other digital coaches, and those tabletop coaches out there who track blocks, do you have similar stats in ties? If not, what are your blocks/turn in tie games? Theories?

I bet part of it is that when block-heavy strategies go off right, they corral you and beat you down. But since I'm not much for PO and Claws, and most people want to play me with their bruisers, often it just turns into a corralling, leaving the pieces in play for another round, also with a boatload of blocks.

What do you think? After all, Cyanide and FUMBBL open up a new kind of statistical analysis: let's dig into it!

[Coach Grievous]

Very interesting topic and too bad I don't have stats of my own on this front collected anywhere!

Outblocking the opponent by 2x leading to big wins isn't surprising. The difference between ties and losses seems relatively small, so I'd almost lump them together.

I think the answers to your questions lie in tempo and initiative, ie. who sets the blocking pace and advantage from round to round. I think there's a whole territory of Blood Bowl theory mostly untapped in that department.

I think I'd need to see more of your playing to go much deeper, though. As questions towards that:

1) I'd be interested to see more detailed (and unrounded) numbers re: blocks thrown and received per turn.
2) Do you think your opponents - even bashy teams - keep you more at an arm's length when you don't win?
3) Are the games where the number of blocks passes the 50/70 thresholds mostly big scrums?

[mattgslater]

Crunching numbers. Will make dinner, play another round, and get some inferences. 38 games of Human data so far. Because the Human-only sample is barely large enough to make inferences in game-to-game numbers, I'll have to get stats on other races for confirmation, though I remember my orcs throwing a ton of blocks in tie games.

[Corvidius]

I suspect because Ties happen when things go wrong and you get tied up in a ruck. I would imagine it happens significantly more with Ag3 teams than Ag 4.

[RogueThirteen]

I suspect because Ties happen when things go wrong and you get tied up in a ruck. I would imagine it happens significantly more with Ag3 teams than Ag 4.

My thoughts as well. Something happens that shuts down an offense drive and the game turns into a big scrum.


As for why outblocking an opponent 2:1 usually garnishes a win is probably due to the fact that, if you're offensive or defensive bashing efforts are going well, you're knocking several players over each turn. On the opponent's turn, these players stand up (and presumably don't risk dodging away), thus letting you hit them all again. Either way, this is going to give you a numbers advantage if you're removing players or at least a positional advantage as the opponent spends more activations simply standing up than doing anything more useful.

Of course, all these stats are going to vary so much based on the team you're playing and the opponent. If I'm playing Wood Elves, yea, the opponent can go ahead and knock down more guys than I do. But I bet I'll still give them a run for the money when it comes to scoring. If I'm playing Dwarves, and the opponent is knocking down far more players than I am, well them I'm pretty much stuffed.

[Shteve0]

Is it possible to get block numbers of how different races perform against other teams?

For example, I can imagine vs high ag teams, a higher number of blocks indicates that your engagement gameplan is successful enough to bog them down, but not succesful enough to clear the pitch.

That logic applies vice versa to bash teams. High number of blocks suggests you've been bogged down, but not so much so that you're down significant numbers.

Make sense?

[MKL]

I suspect because Ties happen when things go wrong and you get tied up in a ruck. I would imagine it happens significantly more with Ag3 teams than Ag 4.

My thoughts as well. Something happens that shuts down an offense drive and the game turns into a big scrum.
(...)

Same for me.

Which sort of ties are you speaking about, Matt?
In my experience 0-0 or 1-1 games are usually bloody massacres. A stopped 2-1 grind can easily turn into WW1-style trench warfare.
On the opposite, 2-2 or the rare 3-3 games are relatively painless, with a lot of early turnovers disrupting any blocking war.

[mattgslater]

Ah, I see. This is pretty much a matter of the composition I'm facing

In the 38-game sample, I've played 4 games against agility teams (DE, WE, HE, Skaven), yielding two big wins (3-0 both) and two big losses (0-2 and a ludicrous 0-4 match; skulls/1s for EVERY ROLL in the last 5 turns, all my Guard players dead/retired in the first 3 turns), plus two games against boutique teams (Vamps, Underworld), yielding another big (0-2) loss and a small (1-0) win. The DE beat me down through sheer force of 1d pows and killer AV/inj/Cas rolls, and threw 59 blocks to my 61. Both the Vamps and HE beat me by a combination of positioning and skill/luck: the Vamp coach was awesome, the HE coach was quite good and had good d6 rolls to my Tackle guys' constant pushing. The Wood Elves were rather rookieish; the Skaven just ran into the hammer, you know? The Underworld game wasn't as close as the score. 6 games, 3-0-3.

That leaves 32 more games against heavier opposition. I've played 3 games vs Humans (1-2-0) with a combined score of 2-1. In the win and the scoreless tie, I threw more than twice as many blocks as my opponent (72-35, 78-32); in the 1-1 tie, I threw 68 to my opponent's 50. All those numbers are way above my average. I've also played 8 games against other "middleweight" teams: 5 vs. Necromantic (2-2-1), including the match with my lowest blocking ratio (a 1-1 tie where I threw 45 blocks to my opponent's 62, on account of being out-Cas'ed), 2 convincing wins vs. Norse, and one tough one against Lizardmen. I do make slightly more blocks per turn against medium-weight teams, but nothing I think is significant (4.1 vs. 3.9 for teams as a whole and 3.75 vs. lightweight teams). 11 games, 6-4-1. Subtotal 17 games, 9-4-4.

The remaining 23 games (13-4-4) are against heavies. Here's how I've fared, and how the blocks have gone.
Chaos: 3-1-2, with two of the wins being concessions. Blocks may be revealing: 63-37 and 65-40 in losses, 56-42 in full-time win, 53-65 in the tie. Good money after bad; constant concern of the manhandler coach. Concession blocking numbers were 19-9 and 31-7.
Chaos Dwarf: 1-1-0. Blocks were similar: 63-31 in the 3-0 win, 69-39 in the 1-1 tie.
Chaos Pact: 2-0-1, with the loss being a crazy bounce that should have been a W or T. 67-28, 50-29 in the wins, 64-36 in the loss.
Dwarf: 2-1-0, with the tie being me messing around trying to skill a guy up after I got control of the game (in a tourney, would have won). One of the wins is a T3 concession, the other two are 54-34 (Tie) and 54-52 (2-0 win). That T3 concession, by the way, is the only game of the 38 sampled in which I haven't scored any Casualties.
Khemri: 1-0-1, with both games being decided by block dice. 42-45, 49-46 in the win and loss, respectively.
Orcs: 1-0-0, 54-42. Smack smack smack! Only did one Cas, but roughed him up something awful, and there were turns he didn't even stand half his guys.
Undead: 3-1-0. One win was a concession. Generally speaking, I throw 3 for every 2, and a boatload of hits, mostly in the 60s and 70s. The tie was the exception, because he put the spank on me, so I only threw 49 to 44.

So yeah, that's where I am.

[Iranian Spy]

Would like to data on two types of norse teams for this blocking. Norse get to throw more blocks than most teams. With all that frenzy, block and 2 jump up, there's a good chance they will throw more blocks if they can keep most of their men on the pitch.

I'd like to see data on a team that takes jump up on their runners as a first skill and comparative data on a team that goes for a more standard dodge route.

[mattgslater]

Play FUMBBL. Track your own stats. Find other teams, track their stats. I suspect that the two factors are coaching tendency and bash/speed; anything else is minor. Norse make more double-blocks than most, but they also get hurt more, so take fewer actions.

My database skills are mediocre at best, or I'd have already crunched all kinds of advanced numbers for you.

My Humans make about the same number of blocks per turn as my Orcs, and more than my High Elves. My personal average blocks per game is 52.4; with my Humans, it's 56.6.

I picked 20 coaches, all over 160 or winning record in 100+ games, off my buddylist, and I did the math on their blocks per turn. The range is 26.5 to 54.4, with a mean of 40.1, a median of 43.1, and a standard deviation of 7.3*; I would have been 3rd on this list. There were a lot of really good coaches on both sides of that: BattleLore and Alibaba throw almost twice as many blocks as Hitonagashi and more than twice as many as Joemanji. But all four are great coaches.

*Yes, I know the sample is too small for SD to be meaningful.

[spubbbba]

You might need to factor TV range into the analysis as well.

Throwing blocks at low TV tends to be a quite risky process as most team tend to lack either the block or wrestle skill and have a distinct lack of team re-rolls. I think a lot of coaches like to throw blocks but don’t factor in how risky it is. Often they start doing that even before safe moves like standing players up or moving them.

As someone who cares little about his players when facing a block happy coach I’ve always been willing to throw my players into danger. If you give your opponent 4.5 blocks a turn they’ll fail one on average every 2nd turn. Even if they don’t fail then as long as you can keep enough players on the pitch then this will often makes holes in the cage or stop it dead.

Of course this gets gradually less viable as TV rises along with the number of opposing players with block and bashing skills. Plus if you are the weaker team then it is usually a bad idea to get bogged down in a block war unless your cas dice are rolling hot. Even then it’s easy to come unstuck if you start well with a couple of lucky armour breaks and try to press that advantage by over committing to the block war.

[mattgslater]

Funny enough, I've found that excessive blocking is more rewarding at low levels, especially in league formats. That 2 SPP for a Casualty, it's much more likely to lead to the skills that give you an early leg up. And an opponent who lacks things like Dodge, Side Step, and Strip Ball often has few recourses when blocked out of position.

When you make 4.5 blocks per turn, it's not usually 4 this turn, 5 that turn. It's 7 on one turn, then 2 on another. With TRRs and Block, you don't turn over that often. Though I find I often have to make cautious plays late in the half, as I tend to unload 2 RR in the first quarter with uncanny reliability, and often the game comes down to, "can I achieve spatial and/or numerical dominance before I run out of rerolls or fall victim to crazy dice?" It works more often than it fails, but both happen.

Early on, race really matters, as do the first couple skills. ST4 is huge when the other guy doesn't have any Guard. Some teams have 0 block, some have 2, some have 9. If I'm running Dwarfs or Norse vs. ST3, then on my T1, my aim is to lob no fewer than 6 line-blocks with 5 players on 4 defenders, and maybe 9 hits with 7 players on 6 defenders (depending on the dice). But all those hits are skilled, some are 3d, and they won't always all go off (a couple KO/Cas or a both-down, or even the wrong double-pow could reduce it to 3 blocks very easily). The chance of a turnover is actually very small.

[mattgslater]

More accurate stats:

38 games, 2 Human teams

In 22 wins: 4.21 blocks per turn (52.5 per match), 2.37 against for (29.9 per match), for a ratio of 2.06:1
In 5 concession wins (3,3,4,6,12 turns): 5.82 blocks per turn, 2.93 against, ratio 2.39:1
In 17 non-concession wins: 3.73 per turn for (58.2 per match), 2.2 (34.4) against, ratio 1.96:1
In 8 ties: 3.86 for (61.8), 2.95 against (47.1), ratio 1.42:1
In 8 losses: 3.94 for (63), 2.27 against (36), ratio 2.04:1

Hmmm... interesting. Given the limited data set, there may or may not be a very weak inverse correlation between number of blocks thrown and outcome, but there is probably a strong positive correlation between number of blocks taken and odds of a tie. At least for this team with this coach.

Notably, I've thrown under 1.25x as many blocks as my opponent 7 times, and am 2-3-2 in those games, so maybe that's an indicator. Going the other way, there's not much to note: I'm 5-0-2 when blocking more than 2.5x as many as my opponent, but if you take concessions out, I'm 3-0-2, and that only indicates a tendency to win or lose when my oppo doesn't block much, something we already know.

[mattgslater]

The casualty numbers are interesting.

Inflicted:
0 (1 game): 1-0-0 (a concession after 3 turns)
1 (9): 5-3-1 (0.722)
2 (7): 4-2-1 (0.713)
3 (11): 7-1-4 (0.625)
4 (5): 1-2-2 (0.300)
5 (3): 3-0-0 (1.000)

Received:
0 (12): 10-2-0 (0.917)
1 (14): 7-2-5 (0.571)
2 (5): 2-1-2 (0.500)
3 (4): 2-1-1 (0.625)
4 (2): 0-2-0 (0.500)

Balance:
-3 (2): 0-2-0 (0.500)
-2 (2): 0-1-1 (0.250)
-1 (0): 0-0-0
±0 (7): 5-1-1 (0.786)
+1 (7): 6-0-1 (0.857)
+2 (10): 5-2-3 (0.600)
+3 (5): 2-1-2 (0.500)
+4 (2): 1-1-0 (0.750)
+5 (2): 2-0-0 (1.000)

The main inference I can draw from this is: if I take no damage, I win. If I take any damage, it's a crapshoot.

No meaningful data on CR. I'm 2-2-1 vs. coaches over 160 and 2-0-0 vs. coaches under 145. That's no statistical sample. If you expand the high end to "over 155" I'm: 3-2-3, also meaningless.

As for TR difference, this sample is too small:
Favorite (10): 6-1-3 (0.650)
Up 5-9 (5): 4-1-0 (0.900)
Up 10-14 (4): 1-0-3 (0.250): the losses were all different inducements (WApo, Urgoth [who failed utterly], 2xBabe); the win was vs. 2 Babes.
Up 15-19 (0): 0-0-0
Up 20+ (1): 1-0-0 (1.000) vs. 2x Babe, Bribe (failed), Flint (who got hurt early)

Underdog (8): 4-2-2 (0.625)
Down 5-9 (5): 3-0-2 (0.600)
Down 10-14 (1): 1-0-0 (1.000) 1x WApo
Down 15-19 (1): 0-0-0
Down 20+ (2): 0-2-0 (0.500); 1x Babe+Wiz, 1xWApo+Helmut

[mattgslater]

FWIW, I did an unscientific scanning of my High Elf numbers (slightly smaller sample, similar outcome: 19-6-8, or .667, vs the Humans' 23-8-9, or .675). I haven't put together a body of data, but off-hand I noticed this:

Vs. bash or rounded teams, I throw somewhere between half as many blocks and an equal number. The closer to the low proportionate end, it seems, the more likely I am to tie, as my 1-1 game vs. Battlelore, where he threw 60 blocks to my 30. The closer to the high end, the more likely I am to win, and the closer to the middle, the more likely to lose.

Vs. agility teams, it's all over the place; I tend to throw slightly more than my opponents. Part of this may be that for most of my HE team's existence, they've been Guard-free.