EPT
i will keep you posted if we added more stuff to the stats section. for example we will add clan and nation rankings next. let me know if you have any other cool ideas to use the data except for breaking it down into the matchups aswell. Thanks!SC2 Race Stats (tournament games, no ladder) - Page 3
| Forum Index > SC2 General |
|
Tak3r
United Kingdom749 Posts
i will keep you posted if we added more stuff to the stats section. for example we will add clan and nation rankings next. let me know if you have any other cool ideas to use the data except for breaking it down into the matchups aswell. Thanks! | ||
|
EnderCN
United States499 Posts
This is really easy to see, if you add 500 Terran mirrors to your totals above and they all go 2-1 you add 1000 wins and 500 losses bringing the Terran win rate up to 1309-763 or 63%. The more mirror matches the higher the win rate just by design of your system. The more a race is played the more mirror matches you'll end up with. This has completely skewed the entire system unless I'm misunderstanding something about how you are doing it. You should probably just remove the mirror matches completely as they are irrelevant when discussing overall balance. | ||
|
TheRabidDeer
United States3806 Posts
On August 20 2010 07:04 EnderCN wrote: I have the agree with others that including the mirrors messes this up. Every time a mirror happens it results in +1 win or +2 win for that race. So since there are more Terran mirrors than Zerg mirrors it has inflated the win rates for Terran. This is really easy to see, if you add 500 Terran mirrors to your totals above and they all go 2-1 you add 1000 wins and 500 losses bringing the Terran win rate up to 1309-763 or 63%. The more mirror matches the higher the win rate just by design of your system. The more a race is played the more mirror matches you'll end up with. This has completely skewed the entire system unless I'm misunderstanding something about how you are doing it. You should probably just remove the mirror matches completely as they are irrelevant when discussing overall balance. Each mirror adds 3 wins and 3 losses (or, I suppose, 2 wins and 2 losses if there are BO3's). There is no way for the win/loss ratio for a mirror game to not be 1:1 since when one T wins the other has to lose. Unless T is so imbalanced that they can somehow skip past this rule of numbers. So your example of 500 T mirrors to the totals adds 1500 wins and 1500 losses which brings it to 1793 wins 1747 losses which is significantly closer to 50% (50.x% I think, just looking at it) | ||
|
greycubed
United States615 Posts
On August 20 2010 07:10 TheRabidDeer wrote:Each mirror adds 3 wins and 3 losses (or, I suppose, 2 wins and 2 losses if there are BO3's). There is no way for the win/loss ratio for a mirror game to not be 1:1 since when one T wins the other has to lose. Unless T is so imbalanced that they can somehow skip past this rule of numbers. Out of 3,800 views congratulations on being the only one so completely sure of a bad assumption.So your example of 500 T mirrors to the totals adds 1500 wins and 1500 losses which brings it to 1793 wins 1747 losses which is significantly closer to 50% (50.x% I think, just looking at it) | ||
|
TheRabidDeer
United States3806 Posts
On August 20 2010 07:34 wrgrbl wrote: Out of 3,800 views congratulations on being the only one so completely sure of a bad assumption. Where is the bad assumption? | ||
|
Sha[DoW]
Canada110 Posts
I might have made a mistake, it's late here, if you find one point it out and I'll fix with an edit. | ||
|
EliteAzn
United States661 Posts
Although numbers/stats like this should not be a major factor when it comes to balancing, these numbers should indeed cause all those "It's a 50/50 matchup!" people to shut up. Finally there's a good and pretty reliable reference we can copy and paste now. One problem (that I've been reading/skimmed) however is the lack of the asian server/asian tourneys, which may scew these numbers even more...it may make the numbers worse, or it may cause the numbers to be more even...If someone can clear this up, that would be great! Anyways, Good Job! Gotta love stats! | ||
|
Rabiator
Germany3948 Posts
On August 19 2010 07:25 DTown wrote: Time is on Terran's side with this "ever-lasting change" you speak of, IMO. Terran strategies have much more room to evolve and many more tools to evolve with than Z. Unfortunately, methinks the situation will only get worse here on out. = ( And what give Terrans the time to use their many early offensive strategies? Yes, you guessed it ... THE TINY BLIZZARD MAPS! The whole supposed imbalance issue is not about the units and their abilities, it is only about the maps which are small enough to make very very early cheesy rushes viable and which negate the disadvantages of an immobile army. Thats why Scrap Station is so Zerg favored ... very long walkway until you open up the path ... and for early rushes this walkway isnt open. | ||
|
Tak3r
United Kingdom749 Posts
| ||
|
EnderCN
United States499 Posts
Well he changed the wording of the post, before he changed it what you just said was an assumption that it would be 3-3. It actually said 2-1 before and it sounded like he was adding 2 wins and 1 loss for a 2-1 mirror. | ||
|
Tak3r
United Kingdom749 Posts
| ||
|
Shika
Sweden1711 Posts
| ||
|
TheRabidDeer
United States3806 Posts
On August 21 2010 02:01 EnderCN wrote: Well he changed the wording of the post, before he changed it what you just said was an assumption that it would be 3-3. It actually said 2-1 before and it sounded like he was adding 2 wins and 1 loss for a 2-1 mirror. By the time I had read the first post it was already changed to 3-3. | ||
|
Tak3r
United Kingdom749 Posts
- IEM masters - ZOTAC Cup #19 - GosuCup US Quali #2 - Craft Cup #4 | ||
|
Tak3r
United Kingdom749 Posts
| ||
|
SolHeiM
Sweden1264 Posts
| ||
|
greycubed
United States615 Posts
| ||
|
Tak3r
United Kingdom749 Posts
| ||
|
Apolo
Portugal1259 Posts
To the OP: Let's say Terran is in games played without mirror matches Wins 1250 Losses 1000 That's 55.5% win ratio. But if you for some reason want to add 500 mirror matches to that (don't know why because everyone knows they're always 50%) It would become W 1750 L 1500 Which is 53.8% So, i ask, what's the point of adding mirror matches to the statistics? Adding them will put win ratio closer to 50% changing the information that's being presented, skweing the results making it seem more balanced. Take them out and you'll see the real win ratio. | ||
|
cascades
Singapore6122 Posts
| ||
| ||
