Maybe 'games played' or 'average opponent rank' should have more influence on rank

I know this ties a bit into the whole ELO-ranking discussion, but I'm wondering if 'games played' isn't undervalued in the current ranking formula.

I've played a lot of games with my Kislev team this season, and, unsurprisingly, my opponents' average rank is by now very close to 50% as well as my own (that part basically just tied to my own continued 50 win-rate regardless of race).

The average opponent rank of all the Kislev teams ranked above me (or pretty much every other team on that ladder) is lower (and in some cases 10 % ) and they all have played far fewer games than I have.

If you look at the trends for the top teams, you notice that after a winning streak (after several team-restarts in some cases), they also start to get more and more draws and some more losses, so it's very unclear what would happen if they continued to play and if they would be able to maintain their win-rate in the face of more diverse opponents.

Basically, in this instance, playing fewer games is rewarded as opposed to playing more games, so I'm wondering if this aspect of the formula is working as intended as it uses the easy games for ranking in the same way as it uses the harder ones.

last edited by ugh
BB2 Champion Ladder Admin Team

Games played has a HUGE impact on rank, limiting your maximum ranking points severely up to about 28 games played at which point it reaches 95% of the max possible. After 42 games played it has a very small impact.

playing fewer games is rewarded as opposed to playing more games

That's simply not true. Take a look at the ranking formula.

it's very unclear what would happen if they continued to play and if they would be able to maintain their win-rate in the face of more diverse opponents.

That's not true either.

last edited by dode74

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

playing fewer games is rewarded as opposed to playing more games

I don't know what you're not getting here. No one debated that the ranking formula takes games played into account. But, it doesn't take average opponent rank into account, does it?

The more games you play, the more diverse your opponents will be, the closer their average rank will be to 50%.

It is very different if you play only a few games and meet more teams that are highly ranked compared to you meeting more teams that are ranked low (as described in the special case above) and get the same record.

So, why are both treated the same by the formula?

BB2 Champion Ladder Admin Team

I don't know what you're not getting here. No one debated that the ranking formula takes games played into account. But, it doesn't take average opponent rank into account, does it?

That's not what the bit I quoted says. It says "playing fewer games is rewarded as opposed to playing more games" - it's not. You're still capped considerably with almost no chance of qualification.

it doesn't take average opponent rank into account, does it?

I've been working on a formula which does that. Thing is, it doesn't actually produce better rankings than the current formula yet. This has been discussed elsewhere at length, so I'll not repeat it here.

It is very different if you play only a few games and meet more teams that are highly ranked compared to you meeting more teams that are ranked low (as described in the special case above) and get the same record.

So, why are both treated the same by the formula?

Well if you're a new team you are far more likely to be playing low-TVPlus teams anyway, which are themselves very likely to have a low number of rankpoints. Currently there doesn't seem to be a need to differentiate: the current rank formula produces better results.

last edited by dode74

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

That's not what the bit I quoted says. It says "playing fewer games is rewarded as opposed to playing more games" - it's not. You're still capped considerably with almost no chance of qualification.

I was specifically talking about the Kislev case. In that case, that is factually untrue (well, 'almost'). The first places at the moment have 20, 27, 18, 15, 18, 27 games played, so all are below 28 and the first, if nothing changes, 8 games beneath your cap.

The Kislev case might be special as the competition there is very low (being one of the or the least played teams).

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

Well if you're a new team you are far more likely to be playing low-TVPlus teams anyway, which are themselves very likely to have a low number of rankpoints.

Doesn't seem to be true in my specific case, but statistically probably true. Guess I'm on the wrong side of the probability curve again.

With few games, though, the variance is also higher. My first 10 games already had an average opponent rank of above 41, the first 20 above 47 while current number 1 still has 37 after his 20 games total.

While I'm sure he's better than me, I'm also sure I'd have a better rank if I had easier opponents early on, so the records up to that point are not really comparable and they become less comparable the bigger the games-played gap is, I would argue.

My take-away is, that - at least for less played races - restarting and hoping for a good starting record/worse opponents seems to be more rewarding than playing more games with one team in the long run.

A minimum of games to be played to be able to qualify would possibly ameliorate that problem.

@ugh said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

No one debated that the ranking formula takes games played into account. But, it doesn't take average opponent rank into account, does it?

No, and it shouldn't. "Rank" is ordinal in nature not scalar, and as such has no business being used in such calculations. Rankpoints, the metric on which the rank order is determined, is scalar to a sufficient degree, but its accuracy is a factor of the number of games the team in question has played. Just how legitimately scalar those points are depends on how many games the rated team has played, and how many games their opponents have played.
...
@ugh said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

I know this ties a bit into the whole ELO-ranking discussion, but...

It is literally a repeat of the ELO ranking discussion, wearing a false nose and mustache.

@ugh said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

In that case, that is factually untrue (well, 'almost').

Almost facts... which are eyeballed things you've decided are facts... seem to be your bread and butter. That's an almost fact as well.

@ugh said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

Doesn't seem to be true in my specific case, but statistically probably true.

Awesome. Issue settled, then.

@ugh said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

My take-away is, that - at least for less played races - restarting and hoping for a good starting record/worse opponents seems to be more rewarding than playing more games with one team in the long run.

But more importantly, what is your "go-away"?

As usual you are making plenty of declarations about how things work, but are providing no evidence to support your rampant suppositions. We do know that a relationship exists between number of games played and magnitude of final rankPoints... which makes sense since that's how the ranking system is designed. Starting over sends you back to zero games played. That's true because... well... it's part of the definition. That's where we're at. Go find some evidence to support the idea that its better to start over than to continue playing the same team in cases where you've lost <x> games.

@ugh said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

A minimum of games to be played to be able to qualify would possibly ameliorate that problem.

...that you haven't demonstrated exists outside of your anecdotes. I think people who solve problems should focus on the ones that conclusively exist before worrying too much about your imagined dragons.

BB2 Champion Ladder Admin Team

@ugh said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

I was specifically talking about the Kislev case. In that case, that is factually untrue (well, 'almost'). The first places at the moment have 20, 27, 18, 15, 18, 27 games played, so all are below 28 and the first, if nothing changes, 8 games beneath your cap.

The Kislev case might be special as the competition there is very low (being one of the or the least played teams).

I just looked on Goblinspy and it's 28 and 22 for the top 2 Kislev teams respectively (3rd and below are over 11 points behind with the third place team needing 5 straight wins to get into second place, and that would be with 23 games played). The top one is at 17-8-3 and the second at 15-5-2. Their relative win percentages are 75% and 79%. The less-played team has the better win percentage but the more-played team has the higher ranking. Playing more was advantageous in this case as well.

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:
Doesn't seem to be true in my specific case, but statistically probably true. Guess I'm on the wrong side of the probability curve again.

As illustrated, it is true for your specific case.

While I'm sure he's better than me, I'm also sure I'd have a better rank if I had easier opponents early on, so the records up to that point are not really comparable and they become less comparable the bigger the games-played gap is, I would argue.

And the more games played the more rankpoints you are rewarded with. You have 100% of your rankpoints, he has 95%, but he's still scoring considerably higher than you. That's nothing to do with your opponents and everything to do with you having a 47% win% and him having a 75% one. Or are you really trying to argue that you're that much lower because he's had that much easier matches? I mean, you've had 5 concessions in your favour here compared with his one, for example.

My take-away is, that - at least for less played races - restarting and hoping for a good starting record/worse opponents seems to be more rewarding than playing more games with one team in the long run.

A minimum of games to be played to be able to qualify would possibly ameliorate that problem.

Given the top team only will qualify for the less-played races, and that top team will have over 20 games, I'm not sure what you think this would achieve other than making it even more of a grind.

BB2 Champion Ladder Admin Team

So you're aware, I do a coach racial ranking (out of interest, and because it was a possible method of ranking if we were going to a Rez tournament) which uses all of the data from every match played by that coach of that race. So it includes all of Velihopea's Kislev matches, for example. Under that ranking he's still top, and still way above you.

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

I just looked on Goblinspy and it's 28 and 22 for the top 2 Kislev teams respectively (3rd and below are over 11 points behind with the third place team needing 5 straight wins to get into second place, and that would be with 23 games played). The top one is at 17-8-3 and the second at 15-5-2. Their relative win percentages are 75% and 79%. The less-played team has the better win percentage but the more-played team has the higher ranking. Playing more was advantageous in this case as well.

Yes, they switched places after I posted (which is why I wrote 'if nothing changes'), they could have easily switched places again and the lower-played one qualified.

As illustrated, it is true for your specific case.

Nothing illustrated on your part. The average opponent rank was lower for them for the same number of games.

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

Or are you really trying to argue that you're that much lower because he's had that much easier matches? I mean, you've had 5 concessions in your favour here compared with his one, for example.

That is exactly what I am getting at. My win-rate is very accurately captured because of my high number of games and my very average average opponent rating, their win-rate is a) much less accurate due to the comparably lower number of games and b) in comparison valued too-high because of the lower average opponent rank if you compare the same number of our games.

But, if someone would show, after a high number of games a better win-rate than I do (say, at accurately 65 % win-rate) and someone else with such few and easier games would overtake them, I could imagine that would not be considered completely fair, i.e. it makes it more luck-based that you qualify, dependent on the average opponent rank you meet in your first games.

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

I mean, you've had 5 concessions in your favour here compared with his one, for example.

3 of my concessions were after I outplayed the opponent (very high win-probability, 2 of them during the early games), 1 happened before the game started and 1 was a seeming rage-quit after a 1st turn fumble that could very well have lead to a considerable advantage on my part (we'll never know, of course), so it's not as if I need to be ashamed of those. Playing a higher number of games will lead to a higher number of concessions, as well, I would argue. Some qualifying teams have a much higher concession-against rate than I do.

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

Given the top team only will qualify for the less-played races, and that top team will have over 20 games, I'm not sure what you think this would achieve other than making it even more of a grind.

It's a grind no matter what. A lot of the teams on the higher part of the ladder are reboots. Wouldn't it be better if they were given the incentive to continue to play one team instead of rebooting until they get a lucky streak of low-ranked opponents?

last edited by ugh

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

So you're aware, I do a coach racial ranking (out of interest, and because it was a possible method of ranking if we were going to a Rez tournament) which uses all of the data from every match played by that coach of that race. So it includes all of Velihopea's Kislev matches, for example. Under that ranking he's still top, and still way above you.

I wouldn't doubt it for a minute. I don't claim to be better than the ranking I have. (And have repeatedly stated as much). But I would imagine the gap being closer, which is what I am getting at.

BB2 Champion Ladder Admin Team

@ugh said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

Nothing illustrated on your part. The average opponent rank was lower for them for the same number of games.

Because their actual performance was much better. With the same performance you do better if you have more games played (up to the 42-match point).

That is exactly what I am getting at. My win-rate is very accurately captured because of my high number of games and my very average average opponent rating, their win-rate is a) much less accurate due to the comparably lower number of games and b) in comparison valued too-high because of the lower average opponent rank if you compare the same number of our games.

None of which supports the contention that you are better than them. The fact is this is a season, and we are interested in performance. The inaccuracy of few games played is dealt with by giving a lower score for people with few games played. The fact that they are above you with far fewer games played simply means they have performed much better than you.

It's a grind no matter what. A lot of the teams on the higher part of the ladder are reboots. Wouldn't it be better if they were given the incentive to continue to play one team instead of rebooting until they get a lucky streak of low-ranked opponents?

It's a balance. Early on "games played" was more of a factor and it was too much of a grind. The numbers chosen (42 and 28) correlate to one match a day and 4 matches a week for the season: a reasonable play rate which doesn't exclude most people from competing.

@dode74 said in Maybe 'games played' or 'average opponent rank' should have more influence on rank:

None of which supports the contention that you are better than them.

Stop insinuating that I have claimed that.

If you look at their performance, it starts to fall with the number of games played, i.e. their win-rate for the first 10 games is higher than for the next 10 games and so forth (while the ranks of their opponents more often include high ones). Mine, on the other hand, is pretty constant all through the season which is probably due to my more average opponent pool, or could be just one big coincidence.

Extrapolating from that, it doesn't seem likely they would be able to keep their high win-rate over a longer stretch of games which would naturally include more high-ranked opponents still or why should that downward trend suddenly stop?

BB2 Champion Ladder Admin Team

Extrapolating from that, it doesn't seem likely they would be able to keep their high win-rate over a longer stretch of games which would naturally include more high-ranked opponents still or why should that downward trend suddenly stop?

You seem to be under the misapprehension that we are trying to ascertain a coach's skill level here. We're not: we're getting the best performing teams to the qualifiers. If your team isn't performing well enough to get to the qualifiers then start another team if you think you can qualify.

Looks like your connection to Focus Home Interactive - Official Forums was lost, please wait while we try to reconnect.