Can I ask if the chances of success in the Forge is actually what they state, other than the guaranteed instances?
This may be from pure bad luck but I've tried a 70% success rate forge 9 times now on 4*/5* items later levels with no success whatsoever.
The likelihood of this occuring, 9 times in a row is quite quite small (0.3^9= 0.002%) given my success chances so I was wondering whether the non-guaranteed rolls are not at face value to try convince players to roll at 100%, spending more FC per forge than one could otherwise get away with. Especially considering that at late levels, FC becomes relatively rarer and more expensive and you need exponentially more, further prompting a need to buy more/do more runs.
Or is it pure RNG bad luck, its not your day sort of thing?
Forging and the odds
My forging data are not published anywhere, but my preliminary analysis agrees with Skepticraven's. That is, my data are consistent with the stated rates.
This may be from pure bad luck but I've tried a 70% success rate forge 9 times now on 4*/5* items later levels with no success whatsoever. The likelihood of this occuring, 9 times in a row is quite quite small (0.3^9= 0.002%)...
The problem with your study is that you chose to conduct it after it happened, right? I mean, for every player who has 1/500 bad luck like you, there are about 499 players who have 499/500 luck. But we don't hear from them, because they don't post, because nothing remarkable has happened to them. I this way, the forums systematically over-emphasize extreme luck (good and bad).
I was wondering whether the non-guaranteed rolls are not at face value to try convince players to roll at 100%, spending more FC per forge than one could otherwise get away with.
For what it's worth, it seems that the statistics have little effect on player psychology. I've been trying to convince players to roll at 70% chances for years. But most players on the forums still roll at 100%, because they can't stand the disappointment of a failed forging. So the game designers don't need to make the chances worse than the stated 70%.
Cat's old thread that ran the numbers for "the number of saved radiant fire crystals" is found here: https://forums.spiralknights.com/en/node/95282
Even incorporating double levelup into the data, it is still net-cheaper. It shifts that 7.5% cheaper closer to 7% cheaper.
I had a look at the thread and it certainly was interesting.
It's true that I only pinpointed these particular instances in my first comment, but to further my case, I have had a consistently low success rate with non-guaranteed rolls. In memory, in all of my 4/5* forge attempts, only two ever succeeded. This is in contrast to 1-3* forge attempts which have had around 90% success rate roughly at 70~%.
The main problem is more the exponential cost of FC in later levels to even begin forging, which for a casual is really punishing to receive a fail. Also given the relative inconsistency of FC drops, I can't block out a certain number of runs per heat. I was also suspicious because i wouldn't be surprised that Sega implemented such a concept.
I'm happy to leave it to RNG but I just wanted to clarify on the issue.
The main problem is more the exponential cost of FC in later levels to even begin forging, which for a casual is really punishing to receive a fail.
This is the argument that always comes up. When you're paying 60 Radiants for a single forging (medium chances, heat level 7), it really stings to have the forging fail. I understand that.
What surprises me is that players seem to ignore the extra 90 - 60 = 30 Radiants cost of guaranteed chances. Basically they are saying, "I really hate to lose 60 Radiants occasionally, so I'm going to lose 30 Radiants every time."
And when you work out the probabilities, they are indeed wasting an average of 33 Radiants per 5-star item. But some people are risk-averse, and to each his own, and so forth.
The big difference is really the time investment involved. You can either lose ~33 radiants per item, but never pay more than 453 for an item, or you can save 33 radiants on average, and pay over 453 radiants for some items. For all other fire crystals, this isn't really a problem, but it is hard and very time consuming to obtain 453 radiants in the first place, so each time you fail with them it stings more than the ensured 33 more does to farm.
Gamer psych and statistics psych are two different things entirely. A statistician knows that it's better to buy any RNG item in the game, outside of fire crystals, but a gamer feels rewarded by success and downtrodden by failure if it's not something they can control. RNG to gamer psych = "I need to win", while the actual game is "I work to win". Each time you find a radiant fire crystal, it's a "win". Each time you reforge your gear successfully, it's a "win". Each time you fail and lose 90 radiant fire crystals, that's a massive loss despite statistically it evening out over time.
If you're not concerned about being "the best as fast as possible", then the 70% chance is the best way to go. If your gamer instinct is what's leading you, then the 100% is the only option for radiants.
That being said, the 100% also does rarely give forge prize boxes, so it depends how much you value the added bonuses those give.
I agree that the psychology of failing strongly influences players. And that's nothing to scoff at. Many aspects of an RPG (which this game kind-of is) are built around the delicate psychology of perceptible progress, manageable challenges, etc.
I don't agree with this statement of yours:
If you're not concerned about being "the best as fast as possible", then the 70% chance is the best way to go.
I would say exactly the opposite. If you are concerned with being the best as quickly as possible (and you have the mental discipline to handle failures), then medium (70%-80%) chances are the way to go.
But to be honest it comes down to how long you plan to play the game. My advice on these forums tends to assume long-term play. If you plan to play the game for a while, forging many items to heat level 10, then your luck will almost certainly settle down to average luck. And at average luck the fastest strategy is to use medium forging chances.
For example, if you give 1000 players the task of getting 10 fully heated 5-star items, and half of them use 100% chances and the other half use medium chances, then most of the medium players will finish before most of the 100% players.
However, you have a point for players that are just going to make a few 5-star items and then quit the game. Because they're forging so little, their luck may deviate wildly from average. Some will experience very good luck, and others will experience very bad luck (although most will still experience nearly average luck). And risk-averse players might then want to use 100% chances, to avoid the possibility of landing in the bad-luck camp.
Here is a brand-new guide, backed up by 1,000,000 simulated heatings of 5-star items.
https://wiki.spiralknights.com/User:Jdavis/Forging
The guide tries to be fair. It acknowledges that players have varying tolerances for risk, and tries to give varying advice accordingly. Enjoy and/or criticize. :)
Gona be honest, I never thought about the 70% rule till reading this thread. I've been advocating for it in my vids as of late, but so far I've heated a 5* item to level 9, and another to level 5 with the 70% with every one succeeding. I like playing the odds. Better payout then my man Punch!
Thanks Bopp
I feel the need to highlight 2 sentences that came up in Bopp's guide.
"A milder pessimist might ask, "How high would the cost be, if my luck were worse than 70% of other knights?" And the answer is around 453."
For long-term players, the 70% rule will work. A great way to think of it is that "70% of the time, they will be content with forging". If you are playing the game for progression, not progressing will hurt a lot.
We can look into how many items are being forged, and statistically how many of those items will feel good the full way through because most people aren't happy with an item until it is level 10.
Most players will need to have 4 items for the hall of heroes. This means that, using the 70% number, one of their items will suffer many failures, two will be fine, and the third will be mostly OK. However, this will likely be spread out across all 4 items, meaning that frustration will not be condensed into a single item, but instead, across every item, making upgrading not something to look forward to. Compare this to the 100% option, where each time you use it you feel excitement about leveling up your weapon and the chance at a forge prize box/other bonuses.
This is one reason why prize boxes are not as big of a disappointment - they are themed around okay or nice stuff always and a chance at a really good thing. Slime casinos are the same way - you almost always get something back, even if that something is technically at a "loss". Forging is the only place in the game where failure means failure, and it stings a lot more if you aren't prepared for that.
If you are making 10 5* items, then suddenly the fire crystal requirement at 70% will seem a lot less problematic. The more items you make, the easier you can farm for fire crystals, and the less problematic the failure rate becomes.
Basically, if you want that one item you're working on to be completed as fast as possible, the 100% will be better. If you want your entire planned arsenal to be completed as fast as possible, the 70% will be better by far.
One interesting item that Bopp didn't include in the page is that you have a 7.4% chance of spending the minimum of 302 crystals (151 fewer than 3x route).
This is an interesting comparison to the "reasonable pessimist" that Bopp notes the 5% chance to spend at or more than 597 crystals (144 more than 3x route).
You're 50% more likely to get a perfect run than hit a run where you "waste" more than what that perfect run would save you.
Also, for the record... heating 5* items is never "needed" for progression. It's a bit of a scam to entice you to upgrade them only to deal less damage until lvl 9+. Shields are literally worthless to heat (or upgrade from 4* to 5*). Armors are only worth heating to lvl 5 (for that +1hp).
One interesting item that Bopp didn't include in the page is that you have a 7.4% chance of spending the minimum of 302...
You're definitely right, because (0.80)^3 * (0.75)^3 * (0.70)^3 = 0.074. And you can see it in the huge spike on the left side of the plot (the only plot in the guide that I linked). But I hadn't thought much about it, so I hadn't realized that it's bigger than the 5% at the unlucky end of the spectrum. Thanks.
And it's always worth noting that armor and shields are low priorities for heating.
My data suggest the listed numbers are correct. Of course, I only have good enough sample sizes for 2* and 3* using 2x crystals (~60 for each level).
Sounds like unlucky chance.
A contributor to my 5* data ran into a 4-fail chain on lvl 8 5* item (30% failure rate, 0.3^4 = 0.81% chance).