TF/SF Merit rewards: An observation.

Funny, you should say that:

That tells us that the median time for Virgil Tarikoss is 39 minutes. My usual time for completing that SF is c.90 minutes and has been as much as 136 minutes.

It certainly doesn't equate to 1 merit per 3 minutes going by MY measure of play. So either I am doing something seriously wrong or circumstances are such that the data is squeued by aberrant behaviour (as was the case for Katie Hannon and the Eden Trial). This was exacerbated red-side due to the comparative lack of SF's meaning that there were less options so the data became further squeued.

Had the Hero TF's been as 'well designed' as the villain SF's then I am sure the hero rewards would have been significantly less but it remains that a non-speed team is 'penalised'* more red-side than blue.


Not sure I quite managed to get my point across clearly there... oh well...


*not my word

I would just like to point out that those 3 TF's, were essentially driven into the ground PRIOR to the introduction of merits by speed teams running them over & over & over again multiple times per day since they gave exactly the same reward as every other TF in the game. My personal opinion, is that now that the activities of the speed players are more evenly distributed across a much larger number of tasks, that their overall effect on rewards is going to be negligible. The fact that there have been only minimal reductions to one or two TF's/SF's since merits were introduced backs this up.

That was the aberration in behaviour I was referring to, thank you.

however that behaviour has already affected the merit rewards.

I wonder if it would be an idea to re-calculate the merit rewards based on data since merits were released. Might that give a better balance based on how TF's and SF's are being played at this time rather than before the new rewards system?

It would probably return a different set of values, but they might not necessarily be better; certain tasks might still be sorely under-represented by that, especially the higher-level content (less players available to run) and the longer/more variable content (less time invested to accrue runs).

It's been a while since I did any work with statistics, but there's got to be some method involving standard deviations and so forth that could be used to find a more representative "middle ground" time for tasks, rather than just taking the middle number no matter what it is. If, say, the speed runs for a certain task (like Katie or Virgil) cut off around 55%, and the remaining 45% of runs are much longer, the median time is actually a very poor representative of how long the task takes to complete on average - it's almost assured that those lower times represent a smaller proportional share of the player base due to repeated runs.

Of course, it's been over a year since the last time any merit rebalance was announced. It's entirely possible, with the new reward system in place, there's already been some self-correcting a simple re-run of the numbers would show a discrepancy. Not having access to even the simplest form of the numbers, it's hard to say.

Never done a LGTF, I have no idea how long it'd take. However...

I'm a "bad" player simply because, instead of ignoring mobs and ambushes, I prefer to stay and fight them? I'm "bad" because I fight my way to objectives, instead of stealthing or simply running to them? I'm "bad" because I don't collect temp powers to steamroll over AVs that I can defeat perfectly fine, albeit a bit slower, without them?

Your definition of "bad" player excludes a lot of people, and I'm not sure I'd want to be anything but a "bad" player by those standards.

The fact that you take longer than the median time does not equate to the fact that you are doing anything wrong or that the people with times less than the median are aberrant.

Some teams, builds, and playstyles are better for some tasks than others. Some people run more TFs than others.

- If a team of people who play together all the time instead of pugging decide to do a particular TF that they can do quickly once a day, is that aberrant?
- Does it become aberrant if they all decide to play Defenders (or powersets resistant to the foes, or that the foes don't resist) and steamroll the content?
- Does it become aberrant once they decide to skip the optional encounters?
- Is it wrong to take longer on a TF because you are joking around and having fun, or visiting the tailor, or respeccing, or farming for Shivans?

I'm not saying the data isn't skewed, I'm saying that regardless of the calculation, playstyle and other factors are going to make a difference in the amount of time it takes to get those merits.

It's like building a character for performance or for RP fun. If the RP build is sub optimal, that doesn't make it wrong or bad, but isn't the lowered performance part of the choice to build and play that character?

This. Just don't whine about whichever way you chose.

There have been at least two (maybe three?) occasions since I13 where the devs have gone through and rechecked the data to recalculate merit rewards for all tasks, not just specific ones. Some haven't changed (the RSF, for example, still rewards 25 merits), some have gone down (the Nictus Insurrection arc went from 29 merits to 8 merits), but most have gone up (LGTF was originally 30, is now 37, and the ITF was 20 and is now 26). I'm going to detail the examples I just gave, because they're a good representative of why values either change or don't change:

* The RSF rewards 25 merits, compared to the STF which rewards 37 (plus another 2 for the flier, which is a bug that still hasn't been fixed for some reason). The reason for this is because the RSF's median completion time has been datamined to be faster than the STF's. It does not, however, take into account the time used to prepare for the SF, such as collecting Shivans or the Warburg nukes, nor does it take into account teams which do not complete it (i.e. give up on the last mission). The RSF is also the hardest content in the game and it's very rare to see a team slog through it and stick it out over a long period of time, which happens on some STF runs - more likely than not a team either gives up, or they steamroll it. Villain teams in general are faster and more efficient running the same type of content, which is another factor in lowering the median time. Someone brought up the "challenge" aspect of the RSF as a reason to buff merit rewards for the RSF, which Synapse acknowledged, but nothing has been done on that front as of yet.

* The Nictus Insurrection arc is an example of the original merit value simply being set far too high in the first place. That arc had been run so few times that the devs had only a handful of runs to base their completion time on, so they assigned it a 29-merit reward. Someone discovered you could run it in 25-30 minutes easily, so with more data in hand the devs were able to adjust the reward for that arc to what it should have been from the beginning.

* The ITF and LGTF had their merits buffed, and then lowered slightly, but are still well ahead of where they were when I13 launched. Considering these two pieces of content have been the most-run "speed" TFs since I13, and they're still worth more than they were when I13 launched, I'd say the argument that "speed runners cause rewards to be reduced and hurt everyone else" is bunk.

Furthermore, the "heroes have more content that gives better rewards" argument is bunk as well - I spent a while figuring out "optimal" sets of TFs and SFs to run when I13 dropped, and it turns out that *gasp* at level 50 a hero and a villain will have the opportunity to earn about the same amount of merits in the same amount of time. Of course I included co-op content in that, because it's content that's accessible to villains and it would be foolish not to take advantage of it. The argument that does hold some merit, however, is the one of heroes having access to better-rewarding content while leveling up (simply because most hero content is old, long, and clunky and has merit rewards to match the length, while villain content is more streamlined and efficient). Even on my Sonic/Sonic Corruptor, which I soloed most of the way to 35 or so, I had a few hundred merits by the time I hit that level simply by running story arcs as they became available, and hitting Ouroboros to grab the ones I missed (or to run an Ouro SF).

HERE is the spreadsheet where I figured out reward:time between heroes and villains. You'll find the initial values at I13 launch, values after the first merit adjustment, and values after the second adjustment (which I believe was I16).

@macskull, @Not Mac | XBL: macskull | Steam: macskull | Skype: macskull
"One day we all may see each other elsewhere. In Tyria, in Azeroth. We may pass each other and never know it. And that's sad. But if nothing else, we'll still have Rhode Island."

Would that data not have included the data from before I13 when activity on (some) SF's was remarkably different? I was asking if it would be an idea to re-calculate based on the data from I13 and excluding data from before that point.

If the previous data is included then it might have had an effect on the end result. I know I am over-using the example of Virgil Tarikoss here but; if it was the habit of some to run that SF repeatedly and quickly multiple times per day then they might have completed [WARNING: ARBITRARILY MADE UP NUMBER] 1000 speed runs. I13 comes in and the reward is not significant enough to carry on that behaviour so less runs are made. The up-dated data would still have a bias towards those 1000 speed runs which are no longer an accurate representation of how that SF is now played.

Or have I missed something pertinent?

This is true, the only wrench in the works is that it can and often does take a bit longer to form up a villain SF. But even pretending it take the same time.

If it takes you 10 min to form up a team and run run RSF and complete it in the 1 merit/3 min time of 75 min so the task took 85 min.
= 1 merit for 3.4 min

10 min to form up a team and run STF and complete it in the 1 merit/3 min time of 111 min so the task took 121 min.
= 1 merit for 3.27 min

It seems small, but the longer TF's usually have more corners you can cut, but team formation time is always a sunk cost. One that is often higher villain side, but not always.

The result is that the higher merit total of many hero tasks can be leveraged to earn more merits/min. Of course this disappears if your team stays together and runs multiple merit rewarding content, but that is usually the exception rather than the rule.

The difference is seemingly small, but over enough time and enough people it can really add up.

Ummm . . . Mac, your spreadsheet proves that heroes have a higher merit/minute capability and always have. Sure, it's not a big difference -- 4.44 merits per hour -- but that adds up over time. After running the content you listed five times, the heroes have enough for an extra random roll.

But I'm surprised you didn't include the first respec in your chart. In my opinion, it is very possible to get 1 merit per minute by having the same team run all 3 respecs back to back, earning 43 merits in 45 minutes or so. I think that little grouping is the most efficient merit earning program on either side.

In fact, if you substitute the first respec in for RSF that nets you 59.17 merits per hour or .98 merits per minute.

Yes, but you can't really account for form up times, gathering Shivans and the like in the formula. If nothing else, not all teams gather Shivans specifically for a task (may have already had them lying around) or even form up for a specific task ("Thanks for helping out on that mission, guys. Wanna do a SF?").

It would be tantamount to accounting for joking around, bathroom breaks, market runs, and training stops in the formula.

Similarly, it is difficult to account for 'challenge' because of the differences in builds and team composition and the like.

Besides, be careful: although everyone wants there to be more SFs, do you really want a Quaterfield redside to 'even out' task lengths? It would be one way to handle the problem without recalculating the formula and taking into account various controversial/subjective factors (like how much pre-I13 data should 'count').

Taking a post out of context only shows your ignorance. Follow the entire thread or troll elsewhere.

I apologies if quoting you makes you angry. I'll drop it.

Fixed.

median =/= to average

If it where median and a large number of speed runs put a speed time into the middle i might agree that was somewhat unfair. This is not the case though if it's based on average then the really long tf runners are balancing out for the speeder's, along with the middle ground times to get an average for the player base. To expect it to be scewed one way or the other way for your peticular playstyle is self centered. It is done for an average population. What might help if factoring in a non completion factor, i've seen quite a few teams take the RSF out to 3+ hours and can't finish and that data isn't used in factoring average time. (as Macskull pointed out)

The fact that the ITF merits where increased rather than decreased last data mine is an indication that speed runs have not lowered merit rewards as this is the one i believe is "speed run" the most (just an estimate on my own observations). However with out the actual data no one should assume anything about what manipulates the merit rewards. It's more likely that the average time is more heavily influanced on the middle ground numbers being so numerious and the larger and smaller times being so few. If that is the case you should be blaming the average population. You just don't know though without the raw data your just flapping your gums.

We don't have direct access to the data and probably never will, but we can make some reasonable assumptions. Most of these assumptions are made just make the data set mining simpler.

• The time does not include external activities (recruiting, shivans, nukes, potty breaks or forum posting)
• Team composition is ignored (number of teammates, logged out players, ATs, use of temp/vet powers, etc.)
• There is no distinction about who is running the TF (account wise). If someone chooses to run 3 ITFs with 3 different toons and the same 7 teammates, it will be counted as 3 entries within the dataset.
• Incompletes are tossed out (Inability to finish)
• Time to complete is from all runs, regardless of patches or issue (this means the times from the lowered difficulty bug allowing for easy Master of... runs are included, sticking Recluse in STF, TPing past Mold AVs in Eden etc.)

Regardless of the above list though, I believe that speed runners are skewing the average/median time heavily downwards, for two reasons.

1. Speed runners tend to specialize in TFs because they can run them fast. This means they actually like doing TFs, especially fast ones. This means they have the time to fill their day with TFs where most players don't.
2. Speed runners can do a TF in a fraction of the time it takes non speed runners to do a TF.

This means that speed runners are over represented by a heavier margin than everyone else.

As an example, we can compare a speed runner (Player S) to non-speed runner (Player N). Each plays 4 hours a day.

• Player N chooses to play 2 hours on a TF. The other time is spent playing non-TF (maybe non-merit) activities.
• Player S chooses to play 4 hours on multiple TFs. No time is spent playing non-TF activities.

Not only is Player S putting more time at TFing (which is not an issue), but they are able to finish 4-6 TFs in the same time that Player N chooses to finish one TF.

And that is if Player N does one TF a night. I know many players who do one TF a week. This can mean that Player S will be overrepresented in the samples by a factor of 4x, perhaps as high as 20x versus Player N. This will affect the average time, and will affect the median time.

This in itself is not an issue. The issue that is concerning is when the devs use the data as a benchmark for rewards. If anything, the speed runners have highlighted which TF/SFs need looking at.

When speed runners can finish a TF by a factor of 2x-3x over their non speeding counterparts, it may be assumed that the team is experienced and well designed (think ITF). However, when the factor is 5x-10x (as in Eden), I would suggest that the design of the TF needs to be looked at.

Practically speaking, there isn't a huge difference in what was going on before merits were introduced and after. The speed runners still got way more rewards than the non-speed runners, because that is the activity they have chosen to do.

I would suggest that the devs would have benchmarked a lower number of merits/TFmin if speed runner scores were reduced/eliminated anyhow; I would guess that they had a certain number of merits/day/person figured, or they would have adjusted the pricing of the recipes accordingly.

At least the non-speed runners can choose their reward now. It's just that the most broken of the TFs/SFs are putting out poor rewards.

Again, the devs should be well aware all of these issues, since I had pointed it out on more than one occasion in the forum before merits were introduced.