have changed their tune, and in the recently published volume of the Highland Society's Transactions, the chemical adviser of that Society has unreservedly declared his acceptance of these new doctrines. Great credit is due to Lord Rosebery for not only providing the means of carrying on this important research work, but also for throwing his home farm and experimental grounds open for the inspection and information of agriculturists.

D. YOUNG.

CRICKET IN 1899

THE VALUATION OF DRAWN MATCHES

6

THE cricket of 1899 was of unusual interest, and its devotees, an ever-growing multitude, watched its progress, either as actual spectators or through the newspaper reports, with an eagerness which often amounted to anxiety, surprising when it is remembered how entirely free the game is from the stimulus of gambling. The best Australian team that ever visited us, playing with astonishing endurance six days a week for four consecutive months on all the famous pitches of the country, and five test matches between them and the highest talent that England could produce, were attractions which, added to the lengthy programmes of fifteen first-class counties contending for the Championship, kept up a continuous current of mild sensation throughout the season. Glorious sunshine through the principal cricketing months, popular favourites in good form, and records broken every week,-what could the enthusiastic crowds or their faithful providers, the halfpenny evening papers, which fed the excitement on which they live, desire better? One o'clock scores,' 'Surrey going great guns,' Startling cricket,' 'Abel in fine form,' and similar placard announcements, were more popular lines than the duel between Kruger and Chamberlain, and even ran the Dreyfus trial close in popular favour. The trouble the Australians gave us was just the suspicion of bitterness which redeemed the sparkling draught from insipidity. But the one drawback to general enjoyment of the season's performances was the remarkably large proportion of unfinished games. Four out of the five test matches were drawn, twelve out of the thirty other matches played by the Australian team, and sixty-one out of the 150 events which constituted the tournament for the County Championship. These included many of the best and some of the most critical contests. At times, and especially when the fifth of the test matches had to be left drawn, and England had to be content to accept a verdict of inferiority based on the only one of the series which was finished, suggestions for preventing such unsatisfactory results were offered in all the papers. The familiar proposals of widening the wickets, narrowing the bats, extending the time for important matches,

deciding games by the first innings, playing matches of one innings for each side only, playing matches by relays of batsmen from each side alternately, diminishing the distance between the wickets for the benefit of the bowlers, and other ideas were ventilated. The Australians we were assured would not come again unless provision were made for playing at least the test matches to a conclusion, however long they might occupy. Perhaps the dissatisfied commentators might be excused this year for having forgotten that England has had summer seasons when steady or intermittent rain would have interfered with a test match even if a week had been set apart for it. Curiously very little attention was given, if any, to one or two timidly submitted suggestions that the drawn games should be estimated. And yet this could be done, as I hope to show, in nine out of ten games so exactly as to exhibit the result with a fairness only a little less satisfactory than that of a finished match.

The method of calculation which I should adopt would be that of comparing the actual performance of each side in every match, drawn or finished. I should ascertain the two wicket averages in every game, deduct the lower average from the higher one, and credit and debit the two sides respectively. I would only exclude from this valuation the very few matches, not more than three or four in all the Championship tournament this year, in which at least one innings on each side had not been completed, admitting, however, the few instances when one side had voluntarily declared its first innings closed. This system of valuation may be first illustrated by applying it to the Australian test matches. The record, as has been stated, shows one victory to the credit of the Australians, and four drawn games. Estimated by the method just explained, the valuation comes out as below::

==

1. (Nottingham.) England 348 for 17 wickets - 20-47 average; Australia 482 for 18 wickets = 26 77 average. Match value, 6:30 to Australia's credit and England's debit.

2. (Lords.) England 446 for 20 wickets = 22.30 average; Australia 449 for 10 wickets = 44.90 average. Match value, 22·60 to Australia's credit.

3. (Leeds.) England 239 for 9 wickets = 26 55 average; Australia 336 for 20 wickets 1980 average. Match value, 6·75, this time in England's favour.

4. (Manchester.) England 466 for 13 wickets = 35.84; Australia 542 for 17 wickets = 31.88. Match value, 3.96, again to England's credit.

5. (Oval.) England 576 for 10 wickets=57·60; Australia 606 for 15 wickets =4040. Match value, 17.20 to England's credit.

It thus appears that in this series of test matches Australia claims total match values 28.90, while England's three credits reach a sum of 27.91, leaving a balance of only 99 in Australia's favour, and showing a remarkable evenness of talent.

It is not necessary to give in detail the separate values of the other matches played by the Australians during their stay in this

country, but it may be interesting to state that, calculated in the manner here set forth, the total match values credited to them would amount to 1,195 66, while their debits would reach only 785-14. These figures include the test matches. The Australians played in all thirty-five matches, and the official record states that they won sixteen, lost three, and that sixteen were drawn. Of the drawn games eight were in their favour when judged by this method of reckoning. Seven were against them, and one (the first match with Yorkshire, which was abandoned after the county had made eightythree for three wickets) is not counted.

The application of this system to the matches which constitute the County Championship tournament is, however, its most interesting development. I am not by any means venturing to set it up in competition with the official method of computation. All I suggest is that it may be regarded as a statistical supplement to the M.C.C. table. The Marylebone Cricket Club formulated its rule for estimating the County Championship in the autumn of 1894, so that the contest has only been authoritatively recognised since the season of 1895. It provides that one point shall be reckoned for each win, one deducted for each loss: unfinished games shall not be reckoned. The county which during the season shall have in finished matches obtained the greatest proportionate number of points shall be reckoned champion county.' It should be added that the M.C.C. also nominate each year the counties which shall be eligible to compete for the Championship and fix the minimum number of matches which each shall play. In 1899 fifteen counties were included, Worcestershire coming in for the first time. In view of the Australian tour the M.C.C. allowed the minimum of inter-county matches to be twelve, six out and six at home. In ordinary years the minimum has been seven or eight out and home matches. The weather is almost entirely responsible for so many matches being left unfinished; either play is stopped by rain, or, as was the case this year, the hard dry wickets are so favourable to the batsmen that it has been found impossible to dispossess the whole of them in a well-contested game within the regulation three days. No doubt the physical conditions are sometimes supplemented by political tactics. The captain of a losing side finds a couple of stonewallers' invaluable when he has to play against time. A draw makes no difference to the position of his county; a loss does. Thus the natural desire to get a good position in the list tends directly to encourage drawn matches.

6

Nor does the official calculation take into account the extent of the wins. A bare win, a fluky win it may be, counts just the same as an overwhelming victory. Surrey beat Somerset in the last week of May by an innings and 379 runs. A week later Kent beat Somerset by 528 runs to 487. But both victories are of just the same value in the official estimate. There may be some good

cricketing reason for this inequity; but when cricketers appeal to arithmetic and reckon out their results even to the second place of decimals, ordinary statisticians may be excused for expecting something like close accuracy. Certainly, if the object of the calculation is to ascertain the exact relative merits of the various teams as judged by their actual performances, the valuation of each match in the manner here sketched will give that information much more perfectly than it can be got from the M.C.C. method. The following are specimens of the debit and credit accounts I should keep with each county. I take the two at the top of the list, namely, Surrey and Yorkshire. The matches are given in the order in which they were played.

The unfinished matches in the above lists are indicated by asterisks. Surrey played one match with Nottinghamshire, which had to be abandoned on account of rain, after Notts had played half an innings. Neither this nor the Titanic game between the two leading counties themselves, when Yorkshire scored 704, and Surrey replied with 551 for seven wickets, is included in the accounts. Two instances from the above series may be selected to show the defect of