Latest Command Zone- Another "stats" episode

[WWolfe]

So a recent episode of Command Zone, (I think it's 335ish), called Mythbusters, is basically another stats episode based on a sample size of what I believe they said was around 100 games from a few different content creators but less than that for some numbers. Below is what their information said:

How many turns does a game last? The average was 10.29 turns, with most games ending between turns 8 and 12 with 9 being the most frequent ending turn.
How often do you cast your commander? The average was 1.4 times per game. They found the less CMC the commander, the more they were cast but still less than 2 times per game on average.
How many board wipes can you expect per game? The number was 1.32 including all board wipes (Vandalblast, Wrath of God, etc.). In 23% of the games, 0 board wipes were cast. 
How many times does the average player attack per game? The average number of attacks per player per game was 2.86. They went back and looked at this stat a little more in depth and found that the player who attacked the most attacked an average of 5.25 times a game.
How many 7+CMC Spells are cast a game (hard cast, not including spells where the mana cost was cheated)? 2.38 spells of that cost FOR THE ENTIRE TABLE, not one person, for the table per game.

What do we think about these numbers? 

Link to episode: https://www.youtube.com/watch?v=Pr91Crzl7DI&t=3342s

ETA- link to video

[Morganator 2.0]

Do you know what my favorite thing about Deckstats is? The stats.

First off, thank you for putting stats in quotations. If any of you have paid attention in the past you'll know that I wasn't very fond of the last time the Command Zone did a stats episode (context: https://deckstats.net/forum/index.php/topic,45029.0.html). We'll see if that changes this time around.

I'm not going to do my own analysis right now, but right off the bat I'm disappointed. These results should have a standard deviation value attached to them. Fortunately, this is pretty simple stuff so it shouldn't take too long for me to do my own analysis. Just remember that the population that this sample was taken from is Command Zone games; they do not necessarily apply to your group. I for one, know that my group uses very few boardwipes.

[Rinzler]

I think that this all goes along fairly well with what my games generally go like, but the one thing that gets me is the game length.  I almost never play a game that is 12 turns long, and 10 seems a stretch, and my power levels aren't super high, as I like to play with budget decks or higher-level decks with minimal interaction.  Even then, my games end at 6 on the earlier end and 9 on the later end of games that I see often.  And I know that it isn't based on my playgroup since I play mostly online against many different decks and people. 

[WizardSpartan]

Do you know what my favorite thing about Deckstats is? The stats.

First off, thank you for putting stats in quotations. If any of you have paid attention in the past you'll know that I wasn't very fond of the last time the Command Zone did a stats episode (context: https://deckstats.net/forum/index.php/topic,45029.0.html). We'll see if that changes this time around.

I'm not going to do my own analysis right now, but right off the bat I'm disappointed. These results should have a standard deviation value attached to them. Fortunately, this is pretty simple stuff so it shouldn't take too long for me to do my own analysis. Just remember that the population that this sample was taken from is Command Zone games; they do not necessarily apply to your group. I for one, know that my group uses very few boardwipes.

Sic 'em, Morg.

You do make the very good point that their statistics are only applicable to playgroups similar to theirs. I have to agree that, thus far, Command Zone has done a subpar job with their stats. I did a quick look through the old thread you posted, Morg, just to help me remember exactly what the stats were about, and I'm not a fan. I'm no statistician by any means, nor does a lot of the fancy stat talk y'all say make any sense to me, but here are my takeaways from both the older thread and this one so far.

• Older thread: How they come to their conclusions is by a flawed methodology (I see the term "torturing stats" a lot) that Morg handily disproved.
• Newer thread: These stats have yet to be disproved, but they honestly seem situational and oftentimes irrelevant for many playgroups.

[robort]

I watched that episode awhile ago. It was from a small sample size and without the competitive aspect of the game. The stats were derived from casual games.

[ladof]

There are too many factors that go into these games to have an episode like this, imo.

How many times on average is a commander cast? I dunno, how many times was it removed? If I cast it once and it never left the battlefield, that might skew my numbers a bit, especially if I have a commander-centric deck.

How many 7+ cmc cards were cast? I'm sure red decks cast considerably fewer than green decks. You can build decks that abuse expensive cards or ignore them.

How many times do players attack? They kind of addressed this, it depends on what the deck wants to do. I've lost games where the winner never attacked (thanks proliferate and poison). I've lost games where the t2 'send the elf, show him you mean business' mattered.

How many board wipes per game? This made me laugh. I had just finished a game where I cast the same board wipe three times in 5 turns and they're saying less than 2. Please.

Overall, these are the kinds of episodes that make me despise pod casters. They want to act like experts or analysts but fail to recognize several key items that would skew their data. Play what you want, don't listen to the 'experts'. It's a game. Have fun.

[Morganator 2.0]

Where is the data set? The last time they did this they posted a data set. Where is it?

Great... now I have to actually watch the video.

Sample size: 109 games total, some games were excluded. For a lot of these a sample size of "around 50" was used, but no explanation for this number was given. Or maybe it was. I'm not that good at holding my attention.

Game length

The first test was to see what the average turn that games end on is. 109 games were used for this part. They found that on average, games end on turn 10.29. First, I want to see if this average is correct, and then I want to find out what the standard deviation is, so I can get an idea of what the range for a typical game length is. The problem with this, is that they gave the measure of games here based on percentages of the total. So they would say something like "Games ended on turn 9 21% of the time". Fortunately, I can back-track with this to get the actual numbers. Here's what the counts are, assuming that the Command Zone did proper rounding (safe assumption).

Turn 6: 4 games
Turn 7: 5 games
Turn 8: 15 games
Turn 9: 23 games
Turn 10: 12 games
Turn 11: 15 games
Turn 12: 11 games
Turns 13 to 17: 22 games
Turn 18: 1 game
Turn 19: 1 game

There is a gap between turns 13 and 17, which proved to be difficult to figure out. What I can conclude, is that there was a mistake made somewhere. Because even when I make the estimate of the results like this, the average ends up being 10.33.

Turn 6: 4 games
Turn 7: 5 games
Turn 8: 15 games
Turn 9: 23 games
Turn 10: 12 games
Turn 11: 15 games
Turn 12: 11 games
Turn 13: 22 games
Turn 14: 0 games
Turn 15: 0 games
Turn 16: 0 games
Turn 17: 0 games
Turn 18: 1 game
Turn 19: 1 game

It's not a big mistake, but it does mean that something went wrong somewhere. Without me being able to see the data set I don't know where the mistake lies. So I have to guess based on normal distribution. Here's what I think it actually looks like.

Turn 6: 4 games
Turn 7: 5 games
Turn 8: 15 games
Turn 9: 23 games
Turn 10: 12 games
Turn 11: 15 games
Turn 12: 11 games
Turn 13: 8 games
Turn 14: 6 games
Turn 15: 3 games
Turn 16: 3 games
Turn 17: 2 games
Turn 18: 1 game
Turn 19: 1 game

Okay, so with this rough estimate now. It gives an average of 10.6, with a standard deviation of 2.73. I should probably explain what that means. Standard deviation is a measure of error. As a general rule, 68% of the data is within the first standard deviation, and 95% of the data is within 1.96 standard deviations. So for this data, that means that 68% of games will end between turns 8 (10.6-2.73=7.87, rounded to 8) and 13 (10.6+2.73= 13.33 rounded to 13). This means you can expect most games to end between turns 8 and 13. Which is actually pretty close to what The Command Zone got. 70% of their games end between turns 8 and 12. The statement they made is correct.

Now what does this mean for you as a player? For starters, I think people are surprised by this number. One of the hosts was surprised to see that 14% of games end on turn 8 (dude on the right). I've been recording game length for my games this year, and even my casual games tend to end around turn 8, something that surprises my playgroup. Most people stop remembering which turn it is around turn 5 or 6, so I'm sure that the games do feel longer than 8 turns. Once everyone has played their mana ramp and commander, turns tend to take longer. The same could be true in your playgroup. If it is, it means that you need to be prepared to try and make game winning plays by turn 7, or at the very least a play that puts you at a really big advantage.

Quick aside: cEDH games don't end by turn 0 or 1. Dude on the right made that terribly incorrect assumption at one point.

CMC 7+ spells

Sample size here is a little under 50 games. No explanation for why.

Right, so I can't extrapolate on the results any more than what the video already stated. 2.38 CMC 7+ spells per game (all players), 0.6 per player, and 19% of games had no spells of CMC 7+ cast.

Is this a surprise to people? I typically only have 2 or so cards that I'm willing to cast for 7 mana or more. It's just easier to cheat out a high cost. So of course it's not going to happen very often.

Commander Frequency

Again, without a data set, I can't check to see if the numbers here are true, but I think it's safe to assume that they are. The Command Zone also has vested interest in these results, so there is no motivation to fudge them.

Players on average cast their commander 1.4 times. The data was also partitioned based on CMC of the commander. Commanders that cost 2 mana are cast (on average) 1.86 times per game. Commanders that cost 3 or 4 mana are cast 1.47 times per game. CMC 5 or 6 is 1.3 times, and 7+ is 0.77 times. I'm a little annoyed with the way the discussion went however. It seemed like the hosts assumed that once a commander gets removed, players are less inclined to re-cast it. I think that it's because a lot of commanders aren't worth removing. Sorry, but Rhys the Redeemed is just not worth removing. For high-cost commanders, the game could very well end before they get cast. 33% of games the CMC 7+ commander wasn't cast. Again, keep in mind that this population is for YouTubers, it might not apply to your group.

Combat & Attacks

Average number of used combat phases: 2.86

Makes sense. Combat isn't really used that much in commander. Hell, when you get into competitive commander, most decks completely skip over the combat step. Next they checked how often the player with the most combat phases used in a game. Average: 5.25.

I really with I could give standard deviations to these values. This tells me nothing except "Aggressive decks attack more". So... yay?

Board wipes

On average, 1.32 board wipes get cast per game. 0 board wipes cast in 23% of games. Seems about right to me. Nothing much to say here I don't think.

Overall

At the end of the video they mentioned that the data set will be released when they do another big stats episode. I don't have as much to complain about this time. Just that this data needs an error value. It's important. If you say that on average 1.32 boardwipes are cast, give or take 1.5 times... you can't really conclude anything.

I don't watch this content very often, so I guess I need to turn the question to everyone else; do your playgroups often resemble what you see on Commander Knights or MTG Muddstah? From what I understand they play very little combo, and mostly battle-cruiser type decks. Are your playgroups like this?

[Slyvester12]

I watched Game Knights for a while.

When they played their own decks, they usually had more powerful cards (at one point I remember them saying the infect guy had an Italian copy of Tabernacle) and synergies. A lot of people believe Shadowborn Apostle spiked in price because Josh (guy on the left) used it to combo off and win one episode.

Nowadays, they're sponsored by WotC and mostly just play showcase decks for the newest set, usually leaning into more battle-cruiser decks. They went from running $3k+ decks with OG duals to upsized Brawl decks running bounce lands, so their games naturally shifted to less combo-centric wins.

I'd say my games are usually closer to their original episodes. Most of my decks, even the less expensive ones, focus on powerful effects instead of combat. The only deck I own that regularly wins through normal combat damage accumulation is a Mayael the Anima deck that cheats out huge threats and often makes 50+ mana in a turn. That being said, commanders get cast more because people pack a lot of removal and games often go shorter or longer depending on the decks. I think my meta runs significantly more boardwipes, given that we often need to clear out terrifying board states. I know I've run Yuriko against Mayael several times, and each game I had to cast 3+ wipes myself, as well as a couple from others.

[WWolfe]

I would say my average game lasts around 8 turns. Even my combat based decks seem to go faster than their average. My combat decks also don't attack over 5 times a game on average. Usually it's one or two for chip damage then one big alpha strike. Sidisi may attack more than the others but it's built to be battle cruiser Magic (though it does have a few combos built in) because some days that's what most are playing at the LGS (typically late Friday nights when it's mostly older players).

I think their average times casting your commander is skewed because some commanders just don't get removed. Some of my commanders I cast once and that's it barring board wipes.

Board wipes I'm unsure about. Some games I see multiple, some games I see none. It all depends on who is playing what.

The 7+ mana spells I can actually see, I don't have a ton of 7+ mana spells in my decks. The one with the most is Karador with 6 but it's Shadowborn Apostles and built to cheat them (all Demons) out. Sidisi has 3 but one of them is an Eldrazi for the shuffle effect while Varina & Sliver Overlord have none. (looking at my other decks on here, I didn't realize how long it had been since I'd updated a lot of them so I'm not sure how many they currently have off the top of my head but I'd be surprised if it was more than 3 or 4 each)

Overall I think their stats are skewed based on the type of games they used to collect the data.

[robort]

My only question would be at the moment. What type of boardwipes?? Is it a specific type like artifacts, creatures, enchantments or planeswalkers? Is it a niv disk type of wipe? Is it a destroy all creatures wipe but you control an idestructible creature? So then would it be a partial boardwipe or a full board wipe?

[Aetherium Slinky]

My only question would be at the moment. What type of boardwipes?? Is it a specific type like artifacts, creatures, enchantments or planeswalkers? Is it a niv disk type of wipe? Is it a destroy all creatures wipe but you control an idestructible creature? So then would it be a partial boardwipe or a full board wipe?

I got the impression that they counted everything that resembles a board wipe even remotely. They mention Vandalblast so it's assumed stuff like Cyclonic Rift, Nev's Disk etc are board wipes too. I think they also didn't take the effect into account; only how often a spell that could be categorised as a board wipe was cast.