Once again you can view all my data here[docs.google.com]

All data is from backpack.tf (minus a few prices for killstreak kits/fabs that I pulled out of thin air) and thus means:

• Only public inventories are recorded
  
• Nothing from scm
  
• (Failed to mention in last guide) Untradable loot is not included in the counts as it doesn't show up in the stat page totals.

This time around I did automate a bit more of the process, however because stat pages only show 100 levels it still meant I had to manually go through 765 pages of premium searches which does take time and energy and means I didn't get all the data on the same day.

OS, ST, and Mecha were all retrieved May 16th
2C and GG were retrieved on May 17th
Botkillers, Aussies, and related prices were retrieved on May 20th

Considering most of the drop rates stayed within .03% of last time's I don't believe the time differences cause too much of an issue to the data's integrity. I have applied for the advanced API access tho so hopefully if I do this again I can fully automate it and get everything on the same day. (Although I probably won't).

High tours (255+) cannot be 100% counted for as premium search only shows badges as up to lvl 255 even if they are higher than that. This is not a problem for most tours as I can simply take the avg. of everyone over 255 tours on mvmlobby hall of fame, however because there are more than 500 people over 255 in 2c some estimation had to be done. My estimation formula for this was ((488*316)+lobbyavg) / 988 where 316 is the middle of 255 and 377 (the lower limit for mvmlobby) and 488 is the remaining high tours that were not on the mvmlobby hall of fame.

Finally, this time around all prices* were put into keys and then converted into USD, last time I had inconsistent pricing and that might have thrown some things off for calculating return rates. And on the price note all prices were taken day of from the avg of the lowest bot seller and highest bot buyer (a few cases where the bots prices were wildly out of reasonable range I substituted a price from scm/last price suggestion).
*except ofc the spec and pro fab prices I pulled out of thin air ;)

With australiums I would like to add something I discovered from the last spreadsheet after publishing the guide: the legendary 5 box (in 2c) is HIGHLY improbable, and the pan is in fact, almost certainly, an aussie drop. This was determined by multiplying the aussie chance by the pan chance (the odds of getting both) and then plugging it into the odds formula (1-(1-chance)^tours) with the total amount of australium tours done. The number given is the chance that ONE tour would have had an aussie + pan drop by now: 99.3%. Now that's not a 100% however it is still incredibly high so the fact we haven't seen one yet is very suspicious.

Now then...

Australium chance: still 2.11% technically rounds up to 2.12% including the pan this time.

Pan chance: down to 0.00513% from 0.00526% (-0.00013%) This translates to roughly 1 in 19500 which is ~500 more than the previous value. It is also a 0.24% chance that your aussie will be a pan.

Botkillers also stayed about the same at around the 89.5%/10.5% for bad/good ones respectively.
And instead of listing each tour's individual "rates" here's 4 nice pie charts.

ROI stands for return of/on investment. Since this is gambling and the house (usually) always wins, your ROI for any mvm tour is negative, but not all tours are created equal.

First up, the inflation in the last 6 months has treated mvm very well, every australium tour had their pan break even point almost double.

Avg. Australium price rose $10.81 to $54.17
Pan price rose a whopping $1535 to $5135
And every botkiller had a price hike of at least a few cents except the gear grinder ones and Silver Mk. II's (Tours are color coded left -> right is common -> rare and in order as tours appear in mann up queue).


Finally (for this section) here is the ROI data that I was too lazy to try and put in a graph.
The important numbers are bolded, to find your actual ROI you need to take how much a tour costs to you (ticket price * tickets needed) and then subtract the ATV (No Pan) value. For example at 99 cents/ticket 2c has an ROI of -$0.86 on avg.
The other bolded numbers are the how many tours you need to find a pan within to break even or profit (assuming avg loot). I recommend copying the sheet and playing around with values to match your personal rates or how much you would get from selling the pan (ie fees are not included).

Δ is the symbol for delta, aka change. Because I now have 2 data points we can see rates of change for a 6 month period.

For starters, $1,678,614.30 worth of tickets were used, making the grand minimum total of mvm ticket sales a whopping $18.82 million.
A total of 452,915 more tours with:

• 3,549 more oil spill tours (+2.17%)
  
• 2,578 more steel trap tours (+1.44%)
  
• 38,068 more mecha tours (+10.85%)
  
• 357,620 more two cities tours (+10.82%)
  
• 51,100 more gear grinder tours (+16.07%)

A few more interesting points

• GG was the only tour to lose level 1 badges (-1,848 | 4.58%)
  
• OS gained over twice the level 1 badges as any other tour (+8,535) but mecha had the highest % gain (+2.67%)
  
• 2C gained a whooping 84 new high tours but mecha again had the highest % gain at 18.18%
  
• Mecha badges 103 & 137 were the only badges to receive a 300% increase in ownership
  
• (Not a delta but) Badge #96 for 2C has 69 owners ( ͡° ͜ʖ ͡°)

This graph shows the % of total tours done by a certain badge group. Keep in mind that backpack doesn't record levels higher than 255 so the 255 blob is people over 254 tours. Level 1 is not included because a level 1 badge has done 0 tours. (And yes this means the saying "Despite making up only 13% of tours completed, high tours are responsible for over 50% of kick votes called." is somewhat accurate.)

This graph shows the distribution of australiums.

2C still the best biotches

Thank you for coming to my ted talk.