[OUTDATED] Turns to win study

[Scaredgirl]

After doing my noob experiment thing, I'm more than convinced that turns to win is not the best way to calculate speed. Although it might give some idea of the overall speed of the deck, it doesn't take into consideration one huge factor: how long does it take for you to finish you turn.

Lets take this deck as an example:

Code: [Select]

7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dg 7dh 7dh 7dh 7dh 7dh 7dh 7di 7di 7di 7di 7di 7di 7dn (http://elementscommunity.org/forum/index.php/topic,11593.0.html)
With this deck, turns to win might be a bit higher than with some other, more complex decks, but the fact that you can finish your turn in less than a second by simply clicking anything that flashes, makes this deck one of the fastest, if not the fastest upgraded AI3 grinder.

Like I've said before it's the  time spent (aka clicking) that is the key here. When you have to click Immolation, unborrow of buff creatures, those extra seconds really add up.

I'd like to see some data of the top decks here where one player plays for example 1 hour non-stop with one deck. He then would count the number of wins and compare it to number of wins by other decks. So instead of turns to win, it would be wins per hour. The latter is more important for grinders, and it would be interesting to see how it would differ from turns to win.

[Gyrodiot]

Personally, the time it takes me to play this or that deck highly depends on the animations involved. For example, with my netbook which CPU is randomly fast or slow, the Nova/Supernova animation takes 3-4 seconds on bad days. If the opponent plays RoL/Hope deck, the attacks of the N Rays of Light will take 10 seconds...

In some rush decks, just clicking on the flashing cards and borders will take a very short time, but in this case, the speed of the animations can slower the play (and Low Quality hurts my eyes).

It depends too of how you know your deck. If you are testing the speed of several decks, you have to know what to do, in each situation, equally between the decks. We can count the number of clicks before the win, it could be a great indicator... but the testers will have a hard time.

[hrmmm]

i totally got your point (i get it after your first post^^).
i agree there are two different tempi of a deck:
"realtime speed": the time (minutes) you need to win a game
- interesting for grinding money/score gain per hour
"ingame speed": the amount of turns you need to win a game
- interesting for any rush deck

but again this is not a ai3 grinder study (its not in my sight). the ai3 opponent is our indicator for testing the ingame speed and consistency of a rush deck, against a opponent with ordinary cc.
this is meant to be a testing place for all the rush decks out there.

...
I decided to start taking statistics to see just how fast various rush decks are. 
...

...
I'd like to see some data of the top decks here where one player plays for example 1 hour non-stop with one deck. He then would count the number of wins and compare it to number of wins by other decks. So instead of turns to win, it would be wins per hour. The latter is more important for grinders, and it would be interesting to see how it would differ from turns to win.

dont we have The AI3 Farming Challenge (http://elementscommunity.org/forum/index.php/topic,1874.0.html) for exactly the datas you are looking for?

[Scaredgirl]

In some rush decks, just clicking on the flashing cards and borders will take a very short time, but in this case, the speed of the animations can slower the play (and Low Quality hurts my eyes).

Cards can still be played fast, but there is this new clicking delay in Elements which means that there is a delay when activating abilities. For example before when you had 4 Lava Golems on the table, you could click them all as fast as you could. Now there is a clear delay, and if you click too fast, those clicks do nothing. This is the #1 reason why "click decks" are way slower than ones that don't have any extra clicks (abilities that you have to activate).

i totally got your point (i get it after your first post^^).
i agree there are two different tempi of a deck:
"realtime speed": the time (minutes) you need to win a game
- interesting for grinding money/score gain per hour
"ingame speed": the amount of turns you need to win a game
- interesting for any rush deck

but again this is not a ai3 grinder study (its not in my sight). the ai3 opponent is our indicator for testing the ingame speed and consistency of a rush deck, against a opponent with ordinary cc.
this is meant to be a testing place for all the rush decks out there.

...
I decided to start taking statistics to see just how fast various rush decks are. 
...

Here's the problem as I see it. From the original post:

After talking with some friends on kongregate, I decided to start taking statistics to see just how fast various rush decks are.  I've seen people take a few games of statistics here and there, but no mass compilation of deck speeds has been compiled until this one.  I decided AI3 was the best location for these statistics as it has 100 hp and a more concrete deck pool than top 50.  As of now I have 50 tests with each of these decks versus AI3.  As I get time, I am going back and gettting 200 tests for each of these decks, working from fastest to slowest.  Unupped decks are located at the top, and unupped decks are located below in the first post..

It uses words like "fast" and "speed", which is kind of misleading. I would bet big bucks that any new player who reads that post, will assume that if he wants to get rich fast, these are the decks he should use because they are fastest. However that's not what this project is about. This project measures how you can win in the least amount of turns which is different from real-time speed.

Which one is the true speed of a deck? Turns to win, or how long it takes to win?

Lets say that I make a hypothetical deck that has a turns to win average of 6. Looking at the current data, it would be clearly the #1 "fastest" deck. However, lets say that this hypothetical deck requires so much clicking that it's actually twice as long to win a match. So even though it would have the best ttw, it would be the slowest deck when measuring in real-time.

Think about it. We would have the #1 "fastest" deck that would be ridiculously slow to play. Makes no sense. "No, Billy, don't use the fastest deck because it takes too long to win. Instead use the 2nd fastest deck which helps you win twice as fast."

About that "realtime speed" and "ingame speed" you talked about. You are correct, those are the two main things here. However I disagree with "ingame speed" being all that interesting. Most players want to become rich fast. Who cares if it takes 7 or 8 turns to win? What does it matter? All that matters is how fast you can gain revenue. That's the cornerstone of grinding in all games, Elements included.

Like I said, I think this study gives very good information because there is a clear correlation between "turns to win" and "time to win". For example if two decks have similar styles and clicking ratio, ttw might show which one is faster. But when decks have completely different click ratios, ttw becomes much less accurate.

To make this project more "scientific", the first post should be rewritten so that it mentions that this is not about measuring real-time speed of a deck, this only measures turns to win which is something different. Furthermore, we should have a topic titled "Time to win study" which would list wins divided by time spent. Comparing data from these two studies would be interesting.

...
I'd like to see some data of the top decks here where one player plays for example 1 hour non-stop with one deck. He then would count the number of wins and compare it to number of wins by other decks. So instead of turns to win, it would be wins per hour. The latter is more important for grinders, and it would be interesting to see how it would differ from turns to win.

dont we have The AI3 Farming Challenge (http://elementscommunity.org/forum/index.php/topic,1874.0.html) for exactly the datas you are looking for?

Yes, that would be a more accurate way to measure the true speed of a deck (how I see it). Downside is that one person would have to do all the data gathering because of computer performance and different playing styles, so it's much more difficult do.

[Clathius]

Is the title "Turns to win study" not clear enough?  Seriously?   This thread is completely off topic now. 

[Scaredgirl]

Is the title "Turns to win study" not clear enough?  Seriously?   This thread is completely off topic now.

First of all I suggest you actually read my post. I know that the title is "Turns to win study", but it also talks a lot about speed and "fastest rush deck". My point is that when people want a fast deck, they generally want something that makes them rich fast.

If you make a deck that wins on average in 1 minute, and I make a deck that wins on average in 1 minute 10 seconds, which of those would you say is a faster deck? Surely it's the time that makes decks fast or slow, right?

While ttw is interesting know and has clear correlation with the actual speed of a deck, it's not the most important piece of information for your average gamer who doesn't care how many turns it takes to win as long as it's fast. That's the overall speed that your average gamer cares about.

But like we've talked about before, ttw is good way to measure this because computer specs and playing style have no affect so multiple people can do the data collecting. If we only measured time and wins, it would be much harder, but the results would most definitely be a better indicator of the true speed of a deck.

I have nothing against ttw but I feel it should be made clear that ttw does not equal overall speed. That's an important points because some newbies don't get it.

Also I'd like to add that I suggest you keep an open mind and actually read what I post instead of getting all defensive and taking my opinions somehow personally. What I posted was 100% on-topic. Your post however is basically spam that doesn't contribute anything to this discussion, so either present arguments or don't. Playing a moderator is needless spam that is even mentioned in the forums rules.

[CB!]

It's deja vu all over again...

The new delays really do slow you down though... I don't know how many times I've not grown a lava destroyer, or played a SoG because of it...

[Nume]

Yeah I noticed those new delays too. It really hurts with rustler based decks too because it takes forever to convert all that light to life .

[jmdt]

Here's the problem as I see it. From the original post:

After talking with some friends on kongregate, I decided to start taking statistics to see just how fast various rush decks are.  I've seen people take a few games of statistics here and there, but no mass compilation of deck speeds has been compiled until this one.  I decided AI3 was the best location for these statistics as it has 100 hp and a more concrete deck pool than top 50.  As of now I have 50 tests with each of these decks versus AI3.  As I get time, I am going back and gettting 200 tests for each of these decks, working from fastest to slowest.  Unupped decks are located at the top, and unupped decks are located below in the first post..

It uses words like "fast" and "speed", which is kind of misleading. I would bet big bucks that any new player who reads that post, will assume that if he wants to get rich fast, these are the decks he should use because they are fastest. However that's not what this project is about. This project measures how you can win in the least amount of turns which is different from real-time speed.

Which one is the true speed of a deck? Turns to win, or how long it takes to win?

Lets say that I make a hypothetical deck that has a turns to win average of 6. Looking at the current data, it would be clearly the #1 "fastest" deck. However, lets say that this hypothetical deck requires so much clicking that it's actually twice as long to win a match. So even though it would have the best ttw, it would be the slowest deck when measuring in real-time.

Think about it. We would have the #1 "fastest" deck that would be ridiculously slow to play. Makes no sense. "No, Billy, don't use the fastest deck because it takes too long to win. Instead use the 2nd fastest deck which helps you win twice as fast."

About that "realtime speed" and "ingame speed" you talked about. You are correct, those are the two main things here. However I disagree with "ingame speed" being all that interesting. Most players want to become rich fast. Who cares if it takes 7 or 8 turns to win? What does it matter? All that matters is how fast you can gain revenue. That's the cornerstone of grinding in all games, Elements included.

Like I said, I think this study gives very good information because there is a clear correlation between "turns to win" and "time to win". For example if two decks have similar styles and clicking ratio, ttw might show which one is faster. But when decks have completely different click ratios, ttw becomes much less accurate.

To make this project more "scientific", the first post should be rewritten so that it mentions that this is not about measuring real-time speed of a deck, this only measures turns to win which is something different. Furthermore, we should have a topic titled "Time to win study" which would list wins divided by time spent. Comparing data from these two studies would be interesting.

Wow, lots of posting while I was asleep.

After thinking about it a bit, I will modify the first post to be more specific on the nature of the study.  I think part of the problem is that when we see 'speed' and 'faster and slower' we have a different understanding of what the terms mean to the game.  I generally associate 'speed' in terms of this study to ingame speed rater than realtime speed.  A better exampl of how this works in practice wopuld be if you took 2 rush decks, the deck with the lower ttw should win more often.

ttw and realtime speed are related, but are also completely different animals for the reasons you have described.  Of the unupped decks in the study, I usually recommend the mono death that comes in 5th precisely because it has the fewest clicks and the fewest losses; I call it an idiot-proof deck.  Its not as fast in game sas other decks, but due to the fewer number of losses and clicks it like generates equal or more electrum per time than a less stabile deck such as no land stompy.

When looking at losses, they have arguably as big or bigger of an effect on the realtime speed of as compared to number of clicks.  I've not taken a survey, but most losses seem to occur after 9-10 turns or more or quite a few turns longers than the normal kill speed of the deck.  I did once get a 5 or 6 turn loss to a TU spam of my crimson dragon, but those are quite rare.  Additionaly, losses subtract from your electrum total giving you more games to reach your electrum goal.

As to a time to win study, as I've hinted at before I will explore this in the future.  Right now I am working on phase 2 of the ttw study updating existing decks for 200 games for accuracy more than adding new decks.  Once I have most of the lower ttw decks updated so I know exactly what decks to look at, I will begin taking 50 game runs (50-90 minutes) with the decks looking at electrum, score, and games played per time.  Due to uniformity, I will have to do all of these so I want to know exactly which decks to examine before taking this step.

[jmdt]

Yeah I noticed those new delays too. It really hurts with rustler based decks too because it takes forever to convert all that light to life .

Yeah these new delays are very annoying.  When using a rustler deck now I sometimes want to throw things, lol.

[redium]

Tremendous work man.  That is an impressive labor you pulled off.

[stormfury]

i just started playing yesterday ,and i love the mono death rush deck although i modified it for a higher win ratio.
also i was able to upgrade 4 cards so far just farming AI3 with this deck only play for a total about 8hours.

after i upgrade the 4th card i thought i would try my luck vs FG ,and i almost won he only had 3hp left.

Code: [Select]

4t3 4vj 4vj 52g 52g 52g 52g 52g 52g 52g 52g 52g 52g 52g 52g 52h 52h 52h 52o 52o 52o 52o 52o 52o 52t 52t 52t 52t 710 710 710 71d

Edit: Mark is Death ofc ,but how do i add the mark to the deck img ?