constructive criticism of The Tetris Company

[exchliore]

Really? Okay let's look at some figures and see if we can't get a conservative ballpark estimate of how much Tetris is raking in.

Let's run with your ball and assume TTC charges their licensees 5% of their revenue. EA currently pays for the mobile rights.
The first quote was from 2010. EA's Tetris hit the app store in July of 2008. Let's assume sales dropped by 25% since then per your reasoning. Let's also weight the lower price bound by 3/4ths the upper bound to be conservative.

Code Select Expand


// Find total revenue, with weighted price.
(100000000*(.75*1.99+.25*4.99)) = 274000000

// Find average annual revenue for that period.
274000000/(560 days / 365) = 178589285.71

// Assume they're currently making 25% less.
178589285.71-(.25*178589285.71) = 133941964.28

// Pay 30% to the app store.
133941964.28-(.3*133941964.28) = 93759375.00

// TTC gets their 5%.
93759375*.05 = 4687968.75

Conservative estimated annual revenue from EA mobile Tetris to TTC  (app store costs already deducted): $4,687,968.75


There's 1 billion smartphones in the world and 100 million sales. Not everyone plays Tetris and of those who do, fewer would pay for it. Not only that, but the game might be region limited and it probably doesn't work on all smartphones. 10% of all smartphones is pretty good, and it's possible that most of everyone who can and would buy the game have already bought the game. I'm not saying that there's no profit, simply that it's probably not a lot. If the market is close to saturation, then revenue might drop sharply in the coming years. It's probably enough capital to back another game, but you can't exactly release a new version of Tetris every year and expect a high ROI.

DAU = Daily Active Users. Tetris Battle currently serves around 1630461 DAU, and Tetris Friends currently serves around 9700 DAU. Let's again weight the lower bound earnings by 3/4ths to the upper bound for both sites. TTC sub-licenses the brand to Tetris Online (the developer of both sites), but in reality it's just a sister company (that operates out of the same office of BPS--the licensing sub-company of TTC) whose profits, more or less, go to the same people.

Code Select Expand


// Find daily revenue for both websites with weighted earnings.
(1630461+9700)*(.01*.75+.02*.25) = 20502.01

// Convert to annual revenue.
20502.01*365 = 7483233.65

Conservative estimated annual revenue from TF and TB: $7,483,233.65


With the (very) rough estimates for mobile license and TF/TB revenues combined, TTC is conservatively making around $12 million annually. This doesn't even account for the dozens of other games currently on the market that they license out. With that included, it would be no surprise to me if their annual profits (after taxes, labor, office lease, etc.) are well within the eight figure range.

Operational expenses (servers, licenses, bandwidth) are not cheap and running a company is expensive. High quality hosting and bandwidth can run $50k/month or more and let's not get started on how much an Oracle license will run you.

There's daily revenue, but it's probably not that high. At $5 CPM (cost per 1000 views) for 1.6M daily users, you might be looking at $240,000/month ($~2.9M a year), but let's not worry about such small numbers. If we were looking at a small company or an indie studio, $12 million a year is a blessing. But realistically $12 million a year in revenue is not very much in the big world of corporations. Pay your CEO and the few guys on payroll well, but it's not a lot to work with. Also, keeping a fully fledged dev team on staff is not economical, especially when there's no work to do.

No, they really aren't.

Here is a list of features that can be tweaked with minimal code changing:

• Instant DAS
• IHS
• IRS
• Coloured garbage
• Garbage/combo tables
• Max number of players

Here is a list of features that only requires a UI and minimal code changing:

• Stats display (num tspins, tpm etc)
• Better spectator support (just UI changes)

Here is a list of features that require significant overhaul:

• Adding a game lobby/chat functions

Here is a list of lag reducing features that do not require a moderate overhaul:

• Network independent gameplay (AKA shove networking in a thread, if you don't receive messages the players gui is unaffected)

Here is a list of lag reducing features that may require a significant overhaul:

• Fps independent gameplay (AKA model/view/controller framework)
• Improving networking performance (using UDP/TCP for relevant things, optimizing centralized/player based game servers, etc etc)

Obviously due to existing code/framework these are just estimates.

The majority of changes that players ask for can be fixed in the *quick fix* section....

I've read your list, and I'm not going to go through each detail. I'd like to point out that once built, software is usually set in stone and the architecture is committed. If the program wasn't designed for these features in the first place, then it probably means there would be significant effort to make the changes.

Also, even if it is simple, it wouldn't be very professional to have one guy spend just a few hours on a whim to get the changes in. You need the changes to be code reviewed, QA'ed/tested, and supported in production. Thus, taking 1 week of estimated work and making it realistically 4+ months. This also isn't going to happen when you have 0-2 devs on payroll.

We can always go back to the old adage of moving a house a few feet to the left because it's not perfectly centered on its lot. The house is already built, you can't just pick it up and move it.

There has definitely been a paradigm shift in Tetris from the 80s to now. Primarily, it's gone from "man vs. game" (MVG) to "man vs. man" (MVM). (I don't mean "man vs. man" exclusively in regards to multiplayer, by the way.)

I agree. The game has definitely changed over time. I'd also like to add on that what we think should be a good game isn't necessarily what makes a good game. It's hard to contend with 1.6M daily active users.

[XaeL]

Also, even if it is simple, it wouldn't be very professional to have one guy spend just a few hours on a whim to get the changes in. You need the changes to be code reviewed, QA'ed/tested, and supported in production.

code review? QA'ed/testing? have u seen the last 10-20 updates to tetris battle? "Dear tetris battlers we have a new update, hope it helps the let us know" followed by the SAME bug reports.
[!--quoteo--][div class=\\\'quotetop\\\']QUOTE[/div][div class=\\\'quotemain\\\'][!--quotec--]
We can always go back to the old adage of moving a house a few feet to the left because it's not perfectly centered on its lot. The house is already built, you can't just pick it up and move it.
[/quote]
Except this is more like unbolting the windows so that they open (e.g. instant das). Sure there are SOME changes that i listed that will obviosuly require a serious rework, but the simpler features require very little effort to implement.

[Corrosive]

<moved to my own thread>

[Kitaru]

• Slow DAS and/or autorepeat (a major player-killer in NES, for example).

I'd say that slow DAS was largely a balancing measure that allowed NES to achieve something that it was ahead of its time to attempt. Due to creative or technical restrictions, features like lock delay and speeds above 1G were not considered. Tuning DAS to be a sufficient tool for survival without overpowering the top survivable fall speed allowed NES to achieve something in the spirit of challenge akin to TGM2+ Death mode or TGM3 Master mode far earlier than the release of those games. It's a bit of a quirky mechanic that further development has supplanted -- you'd probably be better off going with 20G + lock delay instead as well as give fast DAS that can't be surpassed through physical execution. So, on one hand I can understand the analogy you have leveled, but I don't think this is strictly the same thing going on here.

I believe this is the same thing currently going on with Rosti/Kitaru vs. Integration, except over Sega/Arika-lineage features. For example, I see how the ARS rotation system adds technical challenge in 20G by limiting mobility. However, like before, I don't see it as that meaningful in a "Tetrisish" / spatial reasoning / core gameplay type of way. It seems more to me about familiarizing yourself with arbitrary ARS-specific techniques and choosing placements to conform to these technical challenges in the place of choosing placements that are smart in a purer sense (i.e. do they tessellate well and make my field better for future possible pieces?).

"In the place of?" I know this goes down to the somewhat subjective nature of this thread, but this seems off-base. I find that TGM has much more in the way of those sort of 'pure' challenges than more or less any other game I can think of. You make it sound a bit like playing TGM requires some arcane or strange knowledge that makes the best stack surfaces somehow disparate from standard play. Honestly, I'm hard-pressed to think of cases where an ordered set of best placement decisions under ARS 20G is all that different from the set of best placements in absence of such restrictions. ARS, SRS, and whatever else have some weird stuff you can do that can help rectify the situation if you select for a less than optimal stack, but I believe the optimal stacks (just considering stability and tetrising here) are very similar (if not almost identical) between styles. The only times I really think "that would have been a good move in Game X but not this Game Y" is for twists and other small tricks.

Where styles differ is the ways they effect pressure against selecting sub-optimal stacks and how much leeway they afford the player in doing so anyway. In NES, one of those forces is in the form of the randomizer behavior -- since there is somewhat less certainty about the pieces that will appear and when, you want to maximize your potential placement options for any potential pieces that show up. The mobility restrictions in TGM are quite similar. Excessive castle structures are already problematic in the general case, but there is more pressure to select against them when they hinder piece mobility. When you get mobility "for free," you're allowed to do more unstable or strange things without as much fear of punishment; you can still look at it and pick out reasons to rate it as a less stable stack, but you have more flex room before it becomes a concern.

Anyway, I would hope you wouldn't go on to defend SRS twists given the line of reasoning displayed there.  With ARS, every important detail can be inferred from the visible state of the piece. There is technical and goofy stuff you can leverage, but it's hardly necessary. I don't think this is any worse than the current Guideline model with regards to multiplayer, where we sort of recognize and accept that some of the twists are a bit off the wall (heh) but add an appreciable measure of depth.

(Although, I don't agree with any spawn delay--it's a redundant feature that could be eliminated by re-balancing gravity and spawn delay.)

Well, I'll believe it when I see it. Clearly it's impossible to achieve functional equivalence, and I don't agree that making up for the missing feature by just putting the numbers elsewhere. Clearly it's possible to do better than something like Tetris Axis Master mode, but even at it's best (as in something like Tetris New Century Extra mode) it has only achieved such lofty heights as "merely clumsy."

[Wojtek]

I must say I agree with many things Integration says. Guideline sucks. Henk says he wanted to make tetris more accesible, but it may opinion he only made it more complicated. Beauty of tetris in simplicity and this is what most people liked it for. Being more complicated is not necessarily bad thing, of couse some changes add more deepness like spin recognition and reward for example. but there are also changes that remove some deepness (i would mention predictablity or randomizer). I also believe guideline failed on popularity ground, compare how many people know classic game rules and how many knows guideline. I think for avarage person, classic style clone would be more tetris-like than modern guideline game. More familiar and more enjoyable, easier to understand, more authentic, etc.

I think tgm3 ars is best tetris ruleset is best rule set yet, but of course there are many things i don't like there: lack of harddrop, useless softdrop, overcomplicated grading system, too harsh difficulty, difficulty adjustments for player performance. and of course hold and multiple preview is not something everyone will prefer but i personally like it. so maybe actually tgm2p is better. but i think tgm3 floorkick make game much more enjoyable, so probably perfection would be somewhere in between (with some extra changes, for example drops, like i mentioned before).

to sum up, i think early tgm games and their predecesors polished gameplay a lot, without compromising core tetris simplicity. tgm3 polished things even more, but overcomplicated few things and also is way too hard. guideline is poorly designed and it don't fit neither for needs of casual or hardcore players. TTC wasted lot of energy on forcing (without success) its broken rules.

[XaeL]

I think tgm3 ars is best tetris ruleset is best rule set yet,

Agree.
The only thing that makes tgm3 bad is the absurdedly high skill u need to just to get 999 on master. I think 999 should be easier to get.

[Integration]

TTC wasted lot of energy on forcing (without success) its broken rules.

I wouldn't call guideline broken per se. The problem is that surving is easy even for beginners, so the only thing you can really do is man vs man challenges. All man vs man challenges will involve speed and skill. The orignial idea behind guideline was to help beginners but it does exactly the opposite. Everyone will need A LOT of time to play faster with more previews. 6 previews is just overkill. Even 3 previews are overkill for me. Using hold slows me down massively and I forget to rehold T and I pieces. Bag randomizer allows to exclude certain piece combinations. The kick system is often not intuitive. There again, people will need time to learn this. It adds deepness to the game but I doubt this is what TTC wants to achieve with their focus on casual players. Speed and skill gap make finding a good sparing partner difficult. After 1 year practice there'll still be players who own you. Games aren't fun, if you can predict who'll win.

I think more previews is the element which adds most deepness to the game without destroying classic gameplay, but at the same time it's also that element which expands the speed gap the most.

[Wojtek]

i have not posted this, because i believed it is obvious. this is how guideline is broken:

pieces can kick up this create many problems:
  you can't have step lock delay reset
  player can avoid entering lock delay at all, so even forcing lock is problematic
  you can do crazy shit like: http://www.youtube.com/watch?v=tQbFRl5rgmU

spawning position create problem:
  game is quite unplayable without piece shadow. funny how TTC let you turn peice shadow off, but it donsn't even make sense

marathon scoring
  you get pretty much same score for doing single whole game as for doing tetris whole game (assuming you clear the game)
  high score stategy involve mostly exploiting flaws in scoring system, too the point, it is not longer played as it intended to be

indeed guideline make game very unchallenging, and only skill left in speed. this is why only thing that can make any sense is sprint and multiplayer. many people on harddrop don't even notice issues because only thing they play is sprint and multiplayer. as i said biggest problem is that they attempted to make game easier to play but in effect game became harder learn and play well, at least for speed dependent modes, because in mode like marathon any skill and strategy dependency is non-existent. there is lot of inconsistency involved, for example is speed supposed to be deciding skill, why DAS sucks so much (please note that most games do not have tunning, so you get stuck with very slow DAS)? i don't think single design choices make guideline bad, it just don't make much sense as whole.

you may be confused how i praise tgm ars changes that made game more challenging and still complain about guideline changes that make game harder to master. there is big difference. tgm has smooth learning curve and guideline has not. most people don't even try hold or T-spin or are not even aware of them. you can show tgm2 to avarage person and he undersand game instantly, and can learn some advanced movements and tricks by playing. guideline require lot of explainations to even get started.

[exchliore]

code review? QA'ed/testing? have u seen the last 10-20 updates to tetris battle? "Dear tetris battlers we have a new update, hope it helps the let us know" followed by the SAME bug reports.

Which is fine. A lot of bugs don't get fixed because it's not feasible or economical to do so. The only thing you can really do is tackle the important problems.

Except this is more like unbolting the windows so that they open (e.g. instant das). Sure there are SOME changes that i listed that will obviosuly require a serious rework, but the simpler features require very little effort to implement.

All of your listed changes would require some degree of serious work and consideration. The "why" discussion would get too technical, nor would it be interesting and it's not really on topic with the thread either. Making software is not easy. If you think otherwise, I have a bridge to sell you.

---

Keep in mind that all TGM rulesets are completely inaccessible to the average/casual player. Marathon with it's silly scoring system on the other hand, is far more accessible even though a lot of people will never make it to level 15. The heart of the marathon scoring system is more or less to track how far a player progresses through the game. Gaming the scoring system to get a higher score is nice, but most players will have enough trouble staying alive as-is.

The bag system is used to reduce the punishments a random piece generator can dish out and many people will never have to deal with repeating z's and s's, or too many squares coming out all at once or never seeing a line piece. This seems to be a bonus since it is hard to deal with such piece progressions (and somewhat unfair).

6 piece preview is good and it's extremely helpful for a lot of players. Seeing the pieces coming up relieves the mystery and stress of playing blind and not knowing what is coming in the future. It might be false hope because some players will play poorly to reach that line piece at the end, but at least players can play the game with certainty and without doubt.

Most players don't care about spawn orientation or position, nor can they play below das 7. A good ruleset would ensure that players are playing against themselves, not the game.

[caffeine]

You make it sound a bit like playing TGM requires some arcane or strange knowledge that makes the best stack surfaces somehow disparate from standard play. Honestly, I'm hard-pressed to think of cases where an ordered set of best placement decisions under ARS 20G is all that different from the set of best placements in absence of such restrictions.

I'm not sure if you're saying either that ARS 20G doesn't require different "best placements" than "non-ARS with 20G" or "non-ARS without 20G". ARS aside, there's definitely a difference between best placements in 20G and in =<1G. The playfield itself prevents pieces from getting where they might best fit.

[fumen]110@5d1llbB31lA3kbE3kbE3ibG3ibI3qbA4G[/fumen]
[fumen]110@idB3mbC3ibB3hbC3ibB3hbE3gbC3gbI3gbI3gbI3gb?I3qbSVxUAFLDmClcJSAVDEHBEooRBMoAVB[/fumen]
[fumen]110@BeA3gbRplbA3RpkbA3gbG3gbI3gbI3qbA4G[/fumen]
It may be that in a =<1G game, an optimal sequence of moves involve creating such "traps" before smoothing them out. In 20G, that sequence is no longer available. My aim here isn't to say 20G per se isn't a meaningful challenge. I'm only trying to show that there is a difference in what's available when considering best placements.

I wanted to make that clear, since I believe 20G + lock delay is a good, natural way to add challenge. It forces the player to find novel ways to make the stack work despite these natural mobility restrictions. ARS adds on top of this idea in two ways: downward-facing spawning orientations and arbitrary kick restrictions. The most obvious consequences of this are seen with I and T (likely why floor kicks were added in TGM3):

[fumen]110@4d+enb8eA3kbC3gbA3kbF3ibH3hbI3qbA4G[/fumen]
[fumen]110@4dTplbB3gbB3ibH3hbH3hbH3gbI3qbA4G[/fumen]

Now if we were to stop here, then perhaps I'd agree with you that the best placements in ARS 20G are nearly identical to those in non-ARS 20G. However, I won't stop here! It's not a fair comparison. We're not taking entry delay and IRS into account. Entry delay and IRS have become so infused with ARS, I doubt there are many TGM enthusiasts who would care to play without it.

The reason is because of ARSers forbidden, subconscious affinity for upward-facing spawn orientations! Yes, diehard ARSers are so attached to their beloved IRS-enabling upward-facing capabilities, they are actually willing to tolerate otherwise unnecessary entry delay.

Here are some examples of situations where simply flipping the pieces' spawning orientation would preclude the need for the antiquated and gameplay-retarding entry-delay mechanic:

[fumen]110@MeaikbB3YiA3gbC3gbI3gbI3qbA4G[/fumen]
[fumen]110@kdKwjbA3ibA3IwC3gbA3gbG3gbA3gbG3gbI3gbI3gb?I3gbI3qbA4G[/fumen]

Now I have heard the argument that downward spawning orientations will require less keys than upward ones, and I agree. However, using less keys is only useful in the context of saving time. When you have to add entry delay, then it defeats the purpose entirely. That means for this argument to work, you either have to stick to around =<1G or add new kicks to ARS, and I doubt ARSers find either of those options very attractive. It would change the "TGM feel" too much.

It is for these reasons that I support my claim that I do not find ARS-specific challenges to be interesting. I do not enjoy the challenge of leveraging ever-decreasing entry delay in order to IRS my way into more mobility. I'd rather skip the technical hoop-jumping and get straight to the point (without needing entry delay to do it).

Well, I'll believe it when I see it. Clearly it's impossible to achieve functional equivalence, and I don't agree that making up for the missing feature by just putting the numbers elsewhere.

For reasons I named earlier in this post (among others), lack of entry delay would probably change the TGM gameplay experience too much for those who play it to want to consider it. It doesn't seem like any TGMers would be in favor of it. However, from a broader, "Tetrisy" perspective, it makes sense to simplify the game design so that players needn't be hindered by entry delay. I find fast and responsive controls to be very important to the quality of the game.

After 20G, the way to re-balance entry delay out of the equation is to tweak the lock delay. That's where the time-challenge comes from anyway. Of course, depending on the rotation system, you'd need to put limitations on floor kicks. I've seen the argument that step-reset has advantages over move-reset, but this concern could be answered with a combination of limiting the reset to x moves+rotations along with "entry-reset" (forcing a hard drop midair after x frames).

6 piece preview is good and it's extremely helpful for a lot of players. Seeing the pieces coming up relieves the mystery and stress of playing blind and not knowing what is coming in the future. It might be false hope because some players will play poorly to reach that line piece at the end, but at least players can play the game with certainty and without doubt.

I would say that the "helpfulness" that 6 piece previews provide is almost completely overlapped by what hold and bag already do. If you were to take out hold and bag, 6 piece previews would mainly be limited to helping better-than-average players who are able to take advantage of it (with exception of the "is the I or T coming soon?" technique).

[Integration]

indeed guideline make game very unchallenging, and only skill left in speed. this is why only thing that can make any sense is sprint and multiplayer.

+1. That's exactly the kind of post I wanted to see. Just to be accurate: There's also Ultra - Mr. Rogers favourite gametype - and partly dig challenge. That doesn't change anything on the statement, that nothing goes without speed.

[jujube]

To me 20G is the core gameplay element that defines modern Tetris, with everything else geared towards achieving the desired balance of playability and challenge. Now you have the randomizer, extra previews, hold box, and generous kicks all on your side as you wage war against this giant gravity monster, who with all of his might can't seem to make the piece sit in place right away.

But for the sake of consistency these skewed elements are transported intact to slower G modes (this fact being my real beef), where you now need T-spin rewards etc. because everybody can Tetris in their sleep. I totally agree with Integration that there's a pressing need for a resurrection of classic-style play, and I'd be happy with a revamped version of anything pre-TTC. I can enjoy a modern mode like TA Death, but for different reasons because it's a different game. It still hasn't replaced Sega Tetris. How could it? (This isn't some sentimental thing; I had played Death before I knew Sega Tetris existed.) And was Tetris DS just an enhanced version of GB Tetris? Of course not. I don't think that was even the intent. They knew they were abandoning one thing for another. Why not have both.

[Kitaru]

I'm not sure if you're saying either that ARS 20G doesn't require different "best placements" than "non-ARS with 20G" or "non-ARS without 20G". ARS aside, there's definitely a difference between best placements in 20G and in =<1G. The playfield itself prevents pieces from getting where they might best fit.

What I'm trying to say is that stacks where mobility prevents the best fit are often already sub-optimal from a stability standpoint in the absence of gravity, but the restrictions to mobility provide additional incentive to select against such stacks. Some of the stacks you drew up as examples have inherently questionable properties, but zero/low gravity offers more opportunities to correct for them before they snowball into a more problematic situation. High gravity holds you to a higher standard of stacking right from the outset.

The reason is because of ARSers forbidden, subconscious affinity for upward-facing spawn orientations! Yes, diehard ARSers are so attached to their beloved IRS-enabling upward-facing capabilities, they are actually willing to tolerate otherwise unnecessary entry delay.

...

It is for these reasons that I support my claim that I do not find ARS-specific challenges to be interesting. I do not enjoy the challenge of leveraging ever-decreasing entry delay in order to IRS my way into more mobility. I'd rather skip the technical hoop-jumping and get straight to the point (without needing entry delay to do it).
For reasons I named earlier in this post (among others), lack of entry delay would probably change the TGM gameplay experience too much for those who play it to want to consider it. It doesn't seem like any TGMers would be in favor of it. However, from a broader, "Tetrisy" perspective, it makes sense to simplify the game design so that players needn't be hindered by entry delay.

I'd just like to remind you that the examples I drew from in opposition to Zero ARE + 20G were both SRS games. My stance is rotation rule agnostic; I was not considering IRS when I made that statement, though obviously it would be preferable to retain such functionality if ARS spawn orientations were used. I much was more concerned with pacing -- it would be really easy for Zero ARE + 20G to become a slippery mess or be prone to double-locking. I think the only surefire way to approach 20G without entry delay is to be very liberal with lock delay (i.e. "set it and forget it" at a fixed 30 frames or whatever and let the player set the pace), but I think that precludes delving into the kind of territory you're talking about.

I find fast and responsive controls to be very important to the quality of the game.

Again, this is where the subjective grey area comes into play, but I don't think it is correct to conflate responsiveness of controls with pacing measures.

...gameplay-retarding...

You say that like it's inherently a bad thing. I'm not convinced that we can maintain the current trend of raw unbounded speed, but that's sort of diverging into a different discussion.

Now I have heard the argument that downward spawning orientations will require less keys than upward ones, and I agree. However, using less keys is only useful in the context of saving time. When you have to add entry delay, then it defeats the purpose entirely.

I think it only defeats the purpose if you want everything to behave exactly like Sprint. At any rate, I think that argument was about ARS vs. SRS with regards to which was better able to leverage TGM's other features to minimize active time per piece. (And ultimately, that's really what TGM emphasizes -- entry delay can be leveraged to fully prepare DAS, buffer drops, etc. to shave down active time to a minimum. Real-time attack is just a reflection of being efficient within the constraints presented.)

That means for this argument to work, you either have to stick to around =<1G or add new kicks to ARS, and I doubt ARSers find either of those options very attractive. It would change the "TGM feel" too much.

Preserving the "TGM feel" doesn't matter if you're not making a TGM game. New kicks are fine, though you'd have to take care to ensure that none of the kicks lead to unexpected anti-finesse. DTET and Tetris EX have delved into that before; I have to say T-EX works pretty well, despite occasionally still having some SRSish moments when a piece warps away up a wall and instantly locks.

[zaphod77]

NES Tetris collects entropy during the course of play to allow people to get really darn lucky more often

Fixed this for you.

Seriously, collecting unfiltered entropy during game play allows you to become one with the randomizer. Anyone who's had that amazing game on NES tetris where everything went right knows the feeling. You maanged ot provide the right sort of entropy and the game just seems to gift you tetris after tetris.

Tengen tetris, on the other hand, does not collect entropy during gameplay, and uses a true memoryless randomizer.  Same with Sega Tetris.  They can both be absolutely brutal, and simply will not supply a sequence that long and friendly ever.

The reason is because of ARSers forbidden, subconscious affinity for upward-facing spawn orientations! Yes, diehard ARSers are so attached to their beloved IRS-enabling upward-facing capabilities, they are actually willing to tolerate otherwise unnecessary entry delay.

Here are some examples of situations where simply flipping the pieces' spawning orientation would preclude the need for the antiquated and gameplay-retarding entry-delay mechanic:

[fumen]110@MeaikbB3YiA3gbC3gbI3gbI3qbA4G[/fumen]
[fumen]110@kdKwjbA3ibA3IwC3gbA3gbG3gbA3gbG3gbI3gbI3gb?I3gbI3qbA4G[/fumen]

Now I have heard the argument that downward spawning orientations will require less keys than upward ones, and I agree. However, using less keys is only useful in the context of saving time. When you have to add entry delay, then it defeats the purpose entirely. That means for this argument to work, you either have to stick to around =<1G or add new kicks to ARS, and I doubt ARSers find either of those options very attractive. It would change the "TGM feel" too much.

Entry delay in TGM erves two purposes.

1) enabling IRS when not clearing lines.  You argument about flipping the pieces would seem to solve that, but there's also
2) Enabling charging of 1g DAS!

Once the lock delay lowers enough, it's vital to be able to charge the autoshift during entry delay.  Otherwise you will get double locks because you couldn't move the next piece over fast enough.

Entry Delay isn't used in SRS because SRS has MOVE RESET, and thus a simple tap or rotate will keep the piece from locking.

The only way to practically remove entry delay from TGM is to also switch to move reset.

we also have

3) giving you the choice of letting up on the soft drop or keeping it down.  With no ARE you must have double soft drop protection, and force you to let up on the stick.  Playing a game with locking soft drop with no ARE and no protection is very frustrating. but with entry delay, that protection is not needed before 20g.

[breadfish]

TTC is too interested in telling people how to play because they want money, and because they think people in general are goldfish.

It wouldn't be a problem to have different rulesets in one game. It was done well in Heboris, and it's in Nullpo as well, but Heboris is more straight forward and not so much of a settings hell. You basically select your ruleset each game and it gives you a hint as to how difficult the rules are if you never played them or don't know the mechanics (e.g. this is for beginners, this is for experts, etc.)

If just random developers can do this, a company that has a budget like TTC most certainly should be able to. But they don't care.

Edit: they also don't seem to get that people would be more willing to throw money at them if they were more interested in a broader range of players rather than trying to enforce silly guidelines and be IP vultures for the hell of it. There's people out there who would pay hundreds of dollars to import one of Arika's machines. It's not far fetched that they'd get more money if they weren't so stubborn with the rules and licenses.

Edit edit:
Also a note on 20G - The games that have it don't start with it in the normal modes. There are modes that do have it from the very start but the more 'default' modes start at a normal pace just like your typical marathon play. It doesn't hit 20g unless you play well enough to reach it, at which point it's already gone through slow/medium/super fast but still visibly falling.