Armstrong Reform

zeitnot

anchor4208 wrote: »

...
Why leave out the last round? It does actually exist, it is played on a weekend, and all (or most) of the players willingly turn up and play. Leaving it out is like saying ‘obviously, we’ll leave out all matches involving Kilkenny’.

I left it out because it plays a different role within the system, and tells us nothing about what clubs and players prefer for the main routine of the season. The joint last round means that the everyone actually knows who the winners were, and the team that plays the last match doesn't have an advantage in knowing exactly what to aim for, neither of which was necessarily the case before the joint last round was brought in. It's not unreasonable even for clubs and players who generally prefer weekday night games, other things being equal, to support a single, unified, one-off finale to the league. It would make sense even if all teams went to weekdays to have a last round of the current format.

anchor4208 wrote: »

It’s not the luck of the draw. You will play on a Saturday against Bray, Kilkenny and Balbriggan whether you are home or away. ...

What I meant was that there is some randomness, as a result of the draw, in how many weekday games you actually get, if you're in a club with a nominal weekday club night. In an ideal world you would get at least your 5 home matches (no randomness), plus all away matches that happen to be against teams that have weekday home matches (so some randomness right there). As it is, you may lose some of your home weekday matches if the draw happens to put you at home to Balbriggan or Kilkenny or both. This season the 'weekday' clubs get between 5 and 7 actual weekday matches, as a result of the draw.

anchor4208 wrote: »

Taking your point and running with it though, Dun Laoghaire, Rathmines, Elm Mount, Blanchardstown and St. Benildus have all opted for weekday matches at home. As a consequence, their players will get 5/11, 6/11, 7/11, 7/11 and 5/11 matches on weekdays respectively, a total of 30/55. So, they will play 55% of their games on weekdays and 45% on weekends. These figures will flip next season if Ballinasloe get promoted.

Of the 12 teams currently in the Armstrong, 7 play their home matches on a Saturday, and 5 have chosen to stick with weekdays. So, just over 58% of the players in the Armstrong are making a conscious decision already to play 100% of their games on weekends. The remaining 42% are consciously signing up for an event where they will roughly play 55% of their games on weekdays, and 45% on weekends.

I can’t say that the majority (58%) agree with my proposal of 2 games each weekend day, and playing in a single venue. They certainly don’t agree with you though, as they have already voted with their feet.

I'm not sure I follow this. Players and clubs have already voted with their feet for a league where they (for weekday clubs) get somewhere between 45% (5/11) and 64% (7/11) (calculated your way) on weekdays. Perhaps they like that! They have voted with their feet for that, not for a all-weekend system.

anchor4208 wrote: »

Who knows what the minority (42%) think. No-one has asked them, and the debate on here involves a very small number of contributors. I’m one of that 42% though, and I’m frustrated with the number of weekends that league chess is eating into, for a single game on each occasion, and I know anecdotally that I am not alone. Maybe the next step should be a survey?

If you banned Saturday matches except where the visiting team had > 1 hour travel (or whatever), how much angst would that cause? Just to get all options on the table.

But yes, a survey might be very worthwhile.

Danville

The major change I would like to see in the Armstrong Cup is as follows;

Teams be reduced from eight to five players.

Why?

I think there is too much of a disparity in standard within teams.
Most top three boards are in the 2200-2400 ratings range.
Most bottom three boards from 1700-1900

E.g.
Board 1 playing latest theory and taking 4 hours to eek out a well deserved win against another very good player.
Board eight losing without raising a gallop, or board eight out rating his opponent by 500 points.
One all!


Some benefits of such a reduction imo
Much more competition within clubs to get on the Armstrong Cup team, which increases club activity on club nights.
Maybe more participation in in-house club championships?
Lots of the best club players don’t even attend club night as they are guaranteed
a team place.
Easier to arrange matches with less postponed games.
A knock on effect on the Heidenfeld with an increase in standard there too. Of course Heidenfeld should also be reduced to five.

A few extra leagues may be needed but that should not be an insurmountable problem.

Most of the lower leagues have five or six players

I have played Armstrong Cup myself for most of the last 20 years and make the following observation;
Almost every year only about three out of twelve team have a realistic chance of winning.
Three are favourites for relegation from the off.
Six sit comfortably in the middle, no chance of winning, little chance of relegation.

It is thus a stale league, no real excitement.

Change to five person teams with hopefully at most 400 points rating difference top to bottom. Comfort zone disappears. Every result super important.

Increased tournament participation as players try and increase their rating to get back on the team?

Thoughts please.
Are you happy with Armstrong Cup playing standards?
Your suggested improvements?

Tim Harding

Danville wrote: »

...
Most top three boards are in the 2200-2400 ratings range.

This is just factually untrue. There are very few active players rated above 2200 FIDE in Ireland (and in general ICU ratings for most players are lower than their FIDE ratings).

Currently Trinity are second in the Armstrong. We have currently no player rated above 2200, though one was briefly. Our two strongest players are in the 2100s and our next two or three in the 2000s.

Name two teams whose board three is above 2200?

Retd.LoyolaCpt

Tim Harding wrote: »

Name two teams whose board three is above 2200?

Kilkenny, Gonzaga A

Tim Harding

Retd.LoyolaCpt wrote: »

Kilkenny, Gonzaga A

AH yes I did think of Kilkenny after posting, but they have only played the top three against Gonzaga A and it's rare they all turn out together.

I think six boards might be reasonable as in most matches the captains are scratching around trying to find subs for the bottom boards. Five is too few because then one team would have three Blacks and two Whites.

cdeb

zeitnot wrote: »

Just to clarify: I didn't say that anchor4208's proposal was "poorly thought out". And I don't think that anything I wrote implied that.

anchor4208 offered a detailed proposal and rationale, and asked for a reaction. I gave a reaction, that's all.

OK, if I've taken you up wrong, my apologies.

Though I'm not saying the proposal was poorly thought out - just the arguments against. You yourself noted that the pros and cons listed were "not exactly neutral", and I agree. I probably extrapolated from there - maybe not entirely incorrectly - that you figured the "against" arguments of, essentially, "change is good" and "people are getting left behind anyway" weren't particularly good.

So my initial comment was merely to agree with what you had already said (albeit that I maybe read it somewhat, but not entirely, incorrectly), and then add my own comments. I've said I'm not entirely against the idea, but a big change like this needs more attention paid to the negatives than the positives I think - to the risks, and not the benefits, because the risks are where you suddenly find there's a serious issue.

cdeb

Tim Harding wrote: »

I think six boards might be reasonable as in most matches the captains are scratching around trying to find subs for the bottom boards. Five is too few because then one team would have three Blacks and two Whites.

While I agree with you that most teams patently don't have 2200-rated board 3s, I don't think the above statement is necessarily correct. I did a rough analysis of the relative strengths of each board in the leagues last year (to follow on from an article I did for the ICJ about ten years ago), and this was the result -

Board	        Rating 17/18
Armstrong 1	2149
Armstrong 2	2043
Armstrong 3	1944
Armstrong 4	1870
Armstrong 5	1828
Armstrong 6	1799
Armstrong 7	1735
Armstrong 8	1615
Heidenfeld 1	1855
Heidenfeld 2	1764
Heidenfeld 3	1705
Heidenfeld 4	1640
Heidenfeld 5	1601
Heidenfeld 6	1516
Heidenfeld 7	1456
Heidenfeld 8	1314
Ennis 1	        1631
Ennis 2	        1566
Ennis 3	        1497
Ennis 4	        1406
Ennis 5	        1326
Ennis 6	        1301

The Armstrong is stronger this year with Gonzaga B and Blanch replacing Curragh and St Benildus B. But even last year, board 8 on the Armstrong was mid-Heidenfeld strength, and was stronger than all bar board 1 in the Ennis. (You can also see it's not remotely true that most teams have 2200-2400 on the top three boards) I don't think it's fair to dismiss board 8 as "scratching around for subs", which to me (and to Danville, based on his post) implies a 1200/1300 making up the numbers. If every board 8 were a sub (which isn't the case, granted, but a lot would be), then you're talking some fairly strong subs - typically in the top half of the club's next team. I don't see a problem with that.

I also don't see how reducing the Armstrong from 8 to 6 would help this - you would still be "scratching around for subs" even in that instance surely?

I don't see the need to go with the sweeping changes suggested by Danville to be honest. I don't see the issue with 1700s and 2400s playing on the same team (though it's very rare, as already noted). I don't see how a reduced Armstrong size would lead to more players playing club championships. While it would be easier for a captain to arrange a 5-man team, it would be harder for a club (and the leagues controller) to arrange a 30% or so increase in the number of teams which would result.

The issues Danville notes about not all teams having a chance of winning the league are by no means unique to the Armstrong. In fact, this would be the norm in pretty much any league in any sport in any part of the world. Some teams are better than others; that's the way of things. And Gonzaga A with five players would still beat Dún Laoghaire with five players.

Danville

Thanks for the replies, in particular to cdeb who was able to furnish facts which always helps. Certainly the ratings figures I used for boards 1-3 were too high.

However in mitigation, the figures I used for the last three boards 1700-1900 were also 100 points too high!

There is a five hundred point difference between boards one and eight in both the Armstrong and Heidenfeld. I think this is too high, meaning too high a discrepancy in standards.

The issue of a five person team having three whites and two black should not be an issue as over ten rounds it evens out. I know that in the eleventh round a team in each pairing loses out but I can accept that.

I believe that our leagues are won/lost on the performance of the bottom three boards and I think the average figures for those boards are somewhat misleading.

In my previous post I said that there are usually only three teams annually really vying for the title and this is in the main because their three bottom boards are on average say almost three hundred points higher than the nine other teams.

I wonder could cdeb have a look at the average rating of the bottom three boards of the winners/contenders of the past?
I postulate that this average figure on say board eight in much higher than 1615, which would in turn mean that the average on the other teams is somewhat less than 1615 and that’s what’s leading to only three teams being in contention.

How then do we flatten the averages?, reduce to five player teams!

cdeb

Danville wrote: »

Thanks for the replies, in particular to cdeb who was able to furnish facts which always helps. Certainly the ratings figures I used for boards 1-3 were too high.

However in mitigation, the figures I used for the last three boards 1700-1900 were also 100 points too high!

The ratings you used for 1-3 were 300 points too high though.

But I still don't see why that's an issue anyway. I don't think the leagues are won/lost on the bottom 3 boards. Gonzaga last year (I only did the first four rounds, but the averages were settling down by then anyway) had an average of 2067 on board 5 and 1650 on 8 (skewed somewhat by 1180 in round 1, but never higher than 1820)

Elm Mount, by contrast, don't tend to use subs somehow, so their board 8 was the same player each time - 1882. They were also 80 points stronger, on average, than Gonzaga on 7 (1946 v 1865). So reducing to 5 boards would be to Gonzaga's advantage. Dublin are harder to measure as they scratched on board 8 (of 4 in my sample), and had an non-ICU rated player another time.

At the bottom, Curragh and St Benildus B were unusually weak for Armstrong teams and are arguably bad examples. Curragh were 1384 on 7 and 1254 on 8. St Benildus B were 1650 on 7 and 1500 on 8 - yet those were by far the boards on which we were closer to Gonzaga and Kilkenny (e.g. we were within 100 points of Kilkenny on 7/8 but more than 400 points adrift on 1/2/3, and within 220 points of Gonzaga on 7/8 and still more than 400 points adrift on 1/2/3). So 7/8 were our best chances of results against those teams (and we scored 1½/2 against Gonzaga there, while our team against Kilkenny had collapsed utterly by that stage unfortunately)

So again reducing to 5 boards would, it seems, stretch the division, not compact it.

undisputed

I dissagree with weekends matches. Games should be played only weekdays. Many players arent available on the weekends.

I agree there should be no postponements.

One other change is needed. A win for a team should be awarded with two points and a draw with 1 point. Like in the chess olympiad

cdeb

Strongly disagree with that. It makes no sense to reward a 7-1 win the same as a 4.5-3.5 win. It also means every match remains important even if your team is 4.5-0.5 behind.

The Olympiad did use game scores, but changed a few years ago I believe because it isn't really suited to Swiss tournaments. But that's obviously not an issue with the Armstrong

undisputed

cdeb wrote: »

Strongly disagree with that. It makes no sense to reward a 7-1 win the same as a 4.5-3.5 win. It also means every match remains important even if your team is 4.5-0.5 behind.

The Olympiad did use game scores, but changed a few years ago I believe because it isn't really suited to Swiss tournaments. But that's obviously not an issue with the Armstrong

It does make loads of sense. It would eliminate the problem, where one team plays well weakened opposition and gets high score. As it often happens now, some teams field very different players for different games. It would bring more balance into the league

mikhail

undisputed wrote: »

It does make loads of sense. It would eliminate the problem, where one team plays well weakened opposition and gets high score. As it often happens now, some teams field very different players for different games.

That sucks, but so would an unfortunate turnaround at 3.5 all. And so would being the last to finish when your team is 5-2 down. Literally nothing to play for. That's not a team. As it is, you can be 7-0 down (or up!) and you still watch that last game with bated breath because every point counts.

If you want fair, just award the title to the team with the highest performance rating.

cdeb

Yep, and those are far more frequent occurrences than a seriously weakened team. So why base rules on issues less likely to arise?

There are rules in place to compensate for a team gaining places by virtue of walkovers received too

Past_Pawn_99

undisputed wrote: »

I dissagree with weekends matches. Games should be played only weekdays. Many players arent available on the weekends.

I dissagree with weekdays matches. Games should be played only weekends. Many players arent available on the weekdays.